The RBNZ, BoC, Fed, ECB, and BoJ land between Tuesday and Friday, with the BoE the following week. Who hikes, holds, or cuts is one thing (who are we kidding, no one is cutting), but wheter these decisions alter the narratives currently priced is where the focus is.

American exceptionalism has become synonymous with the AI trade, and stronger investment, resilient growth, and inflationary consequences of a capex boom have supported US rate differentials and, by extension, the greenback. Yet what provides a tailwind on the way up can become a headwind on the way down. Europe faces a different predicament, as not all yields are created equal. Investor demand for compensation due to Europe’s fiscal and credit risk has driven EGB yields higher, but that pressure has hindered the Euro rather than creating the usual rate advantage. Elsewhere, the UK faces a familiar test of fiscal credibility. The Autumn Budget looks set to favour restraint over ambition, but with much of the good news already priced into sterling, avoiding another gilt market scare may prove easier than delivering further upside for the pound.

G10 monetary policy is still broadly synchronised. Identifying where the prevailing economic narratives begin to diverge is where the opportunities lie. And plenty of potential for divergence remains.

Being right about the central bank and wrong about the currency can be an occupational hazard in FX. With that in mind, we look at the major currencies heading into the next fortnight, where the balance of risks may be shifting, and where we see opportunities of value.

For the rest of this note, we’ll look at how each central bank plays into the currency’s narrative, and summarise our convictions into three trades to see the year out.

USD: When Exceptionalism Becomes a Liability

US economic strength and resilient equity performance have been strong AI-driven tailwinds for the greenback for much of this year. Those same factors also present a potential source of weakness.

The AI investment boom has generated extraordinary spending, with consequences beyond stock market returns. In particular, capex spending has added to near-term inflationary pressures while also supporting economic strength. With the US disproportionately exposed to both, interest rate differentials have shifted in the dollar’s favour. This, however, works in both directions.

The dollar’s underperformance last week amid renewed AI concerns and a sharp tech sell-off closely resembles the DeepSeek sell-off in early 2025, which was also a clear negative dollar event in real time, the flipside of the US’s outsized exposure to this theme. However, dollar underperformance was more pronounced this time around despite less equity underperformance. Part of this may have stemmed from outside the US, where European markets found some relief on the back of a global duration rally and eased credit concerns.

The FOMC meeting next week is a key event, but we see little chance that the Fed’s outlook will change much. We expect a hold in October and a hike in December. It will be in that December meeting that we could get more info on the path for 2027, along with a few more inflation prints in hand for assessment.

Therefore, the key dollar driver will remain these outside factors. A significant reduction in expected AI-related demand would weigh on the greenback, given its exposure and the lofty valuations tied to this spending. On the other hand, sustained capex spending could bring renewed upside pressure, something Warsh has highlighted himself.

AI spending concerns remain an ever-present risk in many markets, but for now, we still don’t put much weight on spending slowing down. So dollar risks remain skewed to the upside.

EUR: Not All High Yields Are Equal

Read more

It was a volatile week. A hotter-than-expected inflation report fueled a risk-off move across all things AI, offsetting index gains in healthcare and consumer staples. It was the S&P’s worst day since July, ending a strong run of calm. Financing concerns returned, leading Coreweave to sell off 20% as its bonds soared to a 15% yield.

Memory stocks were also at the centre of last week’s pain after reports from China’s ChangXin Memory Technologies said it plans to introduce a denser server memory chip before the end of the year. That pushed Micron Technology down 3.9% and SanDisk down 6%, while optical-networking names Lumentum and Ciena fell 10% each. Selling spread to Asia on Tuesday, when South Korea’s Kospi sank 5.5%, and SK Hynix fell 10.1%.

Earnings season is well underway, so a few comments are in order. The big banks gave a mixed bag.  Goldman Sachs reported a 72% jump in equities revenue, but surrendered early gains. Citigroup slid 5.6% as management guided to a decline of about 20% in fourth-quarter markets revenue, and JPMorgan Chase slipped after sharply reducing its share buyback. Overall, equities trading remained solid, but most other areas did not meet the high water marks of recent quarters.

Gasoline rose 9.1% in September and accounted for more than half of the monthly increase in headline CPI. Core goods prices rose by the most since January 2025, services excluding shelter and energy climbed 0.42%, and shelter rose 0.39%. Thursday’s PPI report added to the picture of energy costs feeding through the pipeline. The data showed annual PPI accelerated to 6.3% from a revised 5.6%.

Fedspeak offered little insight, largely expected given the new “low/no guidance” regime. Warsh declined to comment on October’s meeting, while the Fed’s Hammock said policy is not yet restrictive enough to return inflation to its 2% target.

US Treasury bonds rallied at the long end even as inflation data ran hot, marking a tough week for those biased to a bull steepener. The 10-year rose to around 5.30% after Wednesday’s CPI print, then closed at 5.19%. It then fell 10bps to 5.09% on Thursday, its lowest close since September 22, as the selloff in AI shares and Asian equities drove demand for havens. Bonds finally acted like bonds. Data showed liquidations in UST futures across tenors, as short covering drove much of the rally in the latter part of the week.

The dollar rallied after Wednesday’s CPI report but couldn’t hold gains as Treasury yields, the core of recent strength, retreated. The dollar index hit its highest level since last April before fading by week’s end. Realised volatility remains relatively subdued, with price action in FX markets driven more by other assets, particularly the long end of the UST curve.

On a final note for last week’s roundup, cars now fly. Love him or hate him, Elon’s roadmap seems decades ahead of everyone else. It’s rather spectacular to watch.

The week ahead will be closely watched after last week’s action. So, let’s get into the guide to trades moving markets, where things stand, and where they may be heading.

• “Rundown of Equity Themes”

• “Markets Are Energy-Pilled”

• “Summary of House Views”

Enjoy 7 days of complimentary access to AP Research. Unlock this report, our full research archive, and our latest market insights with a free 7-day trial. Start your free trial here.

Rundown of Equity Themes

While attention has recently focused on yields, it most certainly returned to equities last week, which makes this the start of our Monday memo.

Read more

In the run-up to Brazil’s October 4 election, attention focused on Lula and whether he could survive a tightening race against Flávio Bolsonaro. The results were somewhat surprising, with Bolsonaro taking 47.0% of the vote against Lula’s 45.2%. The real rose more than 4% in the aftermath, with the Ibovespa surging and local swaps opening down more than 90bps (the exchange’s maximum initial move). It was a strong repricing for a government that doesn’t yet exist, let alone one that has shown it can deliver the reforms investors are betting on.

Brazil has spent much of this last cycle with the world’s highest interest rates, while government debt and the deficit have widened, creating a dynamic in which certain catalysts could break the equilibrium. Bolsonaro appears to be the catalyst.

Prediction markets have Bolsonaro’s implied runoff probability above 80% after the first round results. Historical precedent suggests that every first-round presidential leader since Brazil returned to direct presidential elections has gone on to win the run-off. Not that this makes October 25 predetermined, but it means an eventual Bolsonaro victory is as good as priced in by markets. With this increasingly reflected in asset prices, simply being right about the election offers no trade.

Our view on Brazil and the next chapter for Brazilian assets focuses on one question: Can Bolsonaro create a credible fiscal regime?

An Election Turned Macro Trade

First and foremost, lower rates will require investor confidence that Brazil can tackle its debt problem.

During Lula’s third term, debt as a share of gross domestic product has risen by roughly 10 percentage points, to 82.9% in August from 73.1% when he took office in January 2023. This is where Bolsonaro’s efforts come in.

Marcelo Kayath (former Credit Suisse head of investment banking in Brazil) is seen as the top candidate to take over as finance minister. After the first round, Kayath said that the local asset rally reflects hope for an economic overhaul, but should not be viewed as a blank check from investors. “It’s a change in expectations, not a definitive assessment of a government that hasn’t even begun.”

Brazil needs to break the fiscal feedback loop it’s currently in. Proposals under discussion suggest Bolsonaro’s economic team is considering a package to curb mandatory spending and deliver an initial fiscal adjustment of about R$250bn ($50bn), or 2% of Brazil’s GDP. The core measures would decouple increases in pensions and another set of continued benefits from real gains in the minimum wage, and eliminate constitutional spending floors that link health and education expenditures to federal revenue.

Under the plan, any above-inflation increase in the minimum wage for active workers would not necessarily pass through to pensions and other benefits, which could receive smaller increases or adjustments limited to the INPC inflation index. Health and education spending would instead be adjusted for inflation.

The difficulty of these spending cuts is that over 90% of primary spending is mandatory. Discretionary spending cuts alone won’t be sufficient, words from Kayath himself.

Going forward, right-wing victories in Brazil’s Congress and key state races will help Bolsonaro advance his economic agenda. The senator’s Liberal Party will hold 28 Senate seats starting next year, making it the chamber’s largest caucus. However, this strength doesn’t replace the ability to build a majority. An economic reform agenda won’t pass solely on the strength of one party.

The first-round rally priced in political change in Brazil, but not necessarily the fiscal adjustment to justify it yet. So, where do markets still offer sufficient compensation for the uncertainty around fiscal delivery? The answer differs across asset classes.

In rates, slowing growth give the BCB room to cut and ease yields at the short end, and the fiscal concerns mentioned above, if efforts succeed, would allow long-end yields to come lower. The real benefits from high carry and supportive external balances, while equities have the potential for further upside through multiple expansion as the cost of capital falls. Each area offers a different way to express the same underlying thesis, with a different degree of election risk already priced.

We start our macro memo with rates, where the fiscal premium is arguably most visible, and where the asymmetry could still be largest.

Rates: The Cleanest Expression

The BCB cut the Selic to 13.75% in September, its fifth consecutive reduction and 125bps of cumulative easing since March. On forward guidance, the bank has pledged to keep monetary policy tight as above-target inflation eases and risks from global price shocks persist. Although monetary policy has played a decisive role in the disinflation process, demand is still driving inflation, so policy mustn’t loosen too much.

Read more

Last week felt slow. Stocks ticked to new highs on Tuesday but lacked the impetus seen in earlier highs this year. As little surprise, the move was driven by AI spending. Thursday saw stocks drop after the FT reported that OpenAI is on track to generate annualised revenue of roughly $50bn based on its current performance. Yet risk-on spirits returned the following day, Friday, with the Spoos and Qs both settling the week at highs. The S&P 500 is in its longest streak of calm since July 2024, with no 1% down day in 49 straight sessions.

Per CFTC data, hedge funds have boosted long positions in Nasdaq 100 futures to their highest levels since 2017. The tech bias is also reflected in Citigroup’s analysis.

Yields remained the centre of attention, though they brought far less drama than in recent weeks. Midweek auctions for 10-year and 30-year Treasuries saw particularly strong demand, suggesting lofty yields are starting to entice some buyers. That helped spark a brief relief rally, with the 10-year falling as low as 5.22% and the 30-year touching as low as 5.59%. To us, US 10s and 30s look set to roll over further. Debt stress remains more pronounced in Europe, with OAT- and BTP-Bund spreads widening again into the week’s end, despite some initial relief from Le Pen’s fiscal plans.

The dollar is in its longest run of strength since January 2025, driven by safe-haven demand amid Middle East conflict and a selloff in European sovereign debt. But the big currency story (and overall market mover) was Brazil. Brazilian markets rallied after the October 4 vote on expectations that right-wing Senator Flavio Bolsonaro will win the October 25 runoff and deliver a much-needed fiscal adjustment to stabilise Brazil’s high and rising public debt. His chances of victory look palpable; his plans for fiscal adjustment, less so. The Brazilian real was up 5% against the yen in Monday’s session.

However, the real test this week is US CPI, which may help or hinder the Fed’s hand, and a big test for stocks arrives, with banks including JPMorgan Chase, Goldman Sachs, and Citigroup kicking off the earnings season. Any signs that higher yields are not severely weakening credit quality and borrower health will be a welcome reprieve for the KBW Bank Index, which has declined for five consecutive weeks and is down 12% from a record.

That wraps up what felt like a quieter week in markets before an important few days ahead. Let’s get into the guide to trades moving markets, where things stand, and where they may be heading.

• “CPI Game-plan”

• “Watch the Dispersion Trade”

• “Summary of House Views”

For newer readers, or those who have yet to explore the full AP Research offering, our about page⁠ is a useful place to start. It outlines what we publish, how we approach markets, and how we translate our research into actionable investment views. If you’re considering full access, we encourage you to take a look. We also offer trial access for those who would prefer to see the research first.

Subscribe now

CPI Game-plan

Read more

This marks an important step in expanding our research distribution, allowing clients to access our work directly within the Bloomberg Terminal alongside their existing market data and research workflows.

If you are an existing AP Research or APFX Research client:

Terminal access is included in your existing subscription at no additional cost.

To activate access, simply reply to this email with the corporate email address linked to your Bloomberg Terminal account1. We will share this address with Bloomberg solely to arrange your research entitlement.

Terminal access is entirely optional. Your existing access via our website and Substack remains unchanged, and you can request removal of your Terminal access at any time.

If you are not currently a paid client but are a Terminal user:

We are offering a 30-day free trial of AP Research through Bloomberg.

To request access, reply to this email with your Terminal-linked corporate email address, or navigate to ENTC <GO> and search for “AP Research” or provider code “APY”.

Thank you for your continued support as we expand the reach of AP Research and continue to develop our research offering.

AP

Bond Voyage

1.85MB ∙ PDF file

Download

Convexity looks compelling at 5%+, but the payoff still needs a catalyst.

Download

Betting Against The Budget

1.34MB ∙ PDF file

Download

A month out from the UK Budget, things aren’t looking good.

Download

What started as a continuation of the most recent trend stretched towards, dare we say, panic (at least in the headlines). Sovereigns with idiosyncratic risks, such as France, fared even worse, as the OAT-Bund spread soared toward 150bps. We had identified this risk in early September (“Qui paie ses dettes s’enrichit”), and we took exposure to the spread widening. After a move from 87bps to 120bps, we took this trade off the books, and in true market fashion, that provided the next leg higher as the spread moved to 150bps just 48 hours later. C’est la vie. One thing is certain: rising yields have found Europe’s weak spot.

Equities stayed resilient. After all, this is a growth-led pressure in bonds, which benefits equities. Rate-sensitive sectors, such as financials, bore the brunt of any selling, finishing September with the second-worst return in 15 years. Technology again provided the main support for the broader market. Semiconductor and software stocks helped the S&P 500 regain more than 60 points off its lows on Thursday and snap a three-day losing streak. Retail stayed under pressure as Nike sank to a 13-year low after poor guidance after the bell on Thursday, off 81% from November 2021 highs. Nike’s “Just Do It” mantra is looking less convincing when whatever it did clearly didn’t work.

As for economic data last week, Friday’s payrolls report showed employers added just 29k jobs, and traders have largely priced out a hike in October. Still, long-end UST yields kept climbing even as rate-hike odds fell, leading to a strong steepening in the curve by week’s end.

Regarding US-Iran and oil, a stalemate provided minimal information and mixed signals for markets. While stocks pared steeper declines on a report that President Donald Trump was willing to give Iran sanctions relief, Iranian officials expressed pessimism about reaching a deal to end hostilities. For now, oil remains off recent highs, with no clear sign that tensions are easing. Yields’ minimal reaction to last week’s drop is a fair indication that oil is by no means the driver of pressure right now.

The dollar remains well supported as rate differential remained favourable. This was evident in EUR/USD, as French budget issues kept downward pressure on the pair and ECB rate hikes were priced out. As our APFX team noted on Friday, “we believe the second-order move [of French fiscal risk] will be felt in lower ECB pricing expectations for 2027.”

That wraps up our takeaways from what was a pronounced week in markets. Let’s get into the guide to trades moving markets, where things stand, and where they may be heading.

• “Contagion Aware”

• “Getting Bulled Up”

• “Summary of House Views”

Just a quick note before we move into the rest of this memo: we joined last Wednesday to talk all things FX and macro. We thoroughly enjoyed the discussion and wanted to share a link to the audio for anyone interested.

Prometheus Research

Prometheus: In Conversation With AP Research

Listen now

23 days ago · 12 likes · Prometheus Research and AP Research

Contagion Aware

The biggest driver of markets this week will be a continuation of the yield story. Maybe the biggest worry on markets’ minds will be contagion risk, with memories of Europe’s debt crisis 15 years ago still fresh.

Read more

Investors have been unimpressed, if not outright pained, for years by long-duration exposure. With yields at current levels, locking in coupons by holding bonds makes a reasonable argument. And if yields drop with slower growth or an unexpected move in inflation, bonds’ convexity becomes more enticing: take small losses, keep income, and reap big gains if the environment normalises.

It’s an attractive setup, and many are weighing in on the conversation. At a 5.19% yield, the asymmetry on the 10-year may appeal to some: a 12.7% upside from a 100bp fall in yields versus a 1.6% downside from a 100bp rise over a one-year horizon (h/t Efficient Market Hype for the example below). Convexity is underappreciated by many, and one of the biggest reasons to own bonds here. However, the one-year payoff is not convexity alone. Carry, rolldown, and pull-to-par do much of the heavy lifting, particularly at a 5%+ starting yield.

There is also a low dollar convexity kicker here. Dollar prices of some low-coupon bonds have dipped below $50, effectively turning these instruments into quasi zero-coupon bonds with high convexity. In a substantial rally, the price return of these securities could outpace that of bonds trading closer to par.

What’s the Catalyst?

The question we’re asking is: what gets them there? An 8:1 return sounds nice on the surface until the probability attached to the upside case is considered. For duration to deliver the kind of returns we’re discussing here, something meaningful has to change.

For the rest of this note, we explore the catalysts that could drive yields lower, and where our conviction lies on receiving rates here or not. We also look at the negative ERP and why we don’t buy into the argument to abandon equities in favour of bonds.

Subscribe now

Read more

A confluence of factors soured appetite for bonds on Wednesday, as Flash PMIs showed business activity was running at its fastest pace since July 2021, and a 5-year auction drew weak demand, alongside a further move higher in global yields.

In equities, semiconductors and hyperscalers did the heavy lifting. Meta Platforms (META US) soared 11% after its Muse AI agent became the most-downloaded product on the US Apple App Store. As a result, the broader Nasdaq 100 Index closed at a record on Tuesday for the first time since June. Yields eased, and oil fell anew on Friday amid hopes the Strait of Hormuz would reopen, allowing the S&P 500 to extend its weekly advance.

The dollar approached a YTD high as yields pushed higher, with carry-sensitive FX especially volatile through the week, hinting that the market may see the USD higher for the longer term. One-week risk reversals widened to 54bps, the most since April, to capture this week’s payrolls data.

One small FYI for Substack app users, we’ll be joining a live podcast with on Wednesday, September 30, at 12:00pm EDT. We’ll be talking through FX, bonds, and some recent market events. You can set a reminder for the conversation here.

Below the paywall, we focus on one important question many clients have asked over the last week: can equities withstand yield pressure?

To read this full note and see our full research library, consider becoming a premium subscriber for a month.

Let’s get into the guide to trades moving markets, where things stand, and where they may be heading.

• “Hurdling a High Bar”

• “One Eye On Midterms”

• “Summary of House Views”

Hurdling a High Bar

Read more

Importantly, the 10-year Treasury yield moved above 5% for (say the line) “the first time since 2007.” Long-volatility demand remains elevated as a result, with large October and November Treasury straddles showing traders expect rates to remain the main source of cross-asset instability.

Reports of additional Saudi crude shipments initially eased concerns around a prolonged supply squeeze, although Aramco later told some European customers they would receive no crude next month. The resulting volatility in energy prices remains an important input for both inflation and equities, but the pullback was enough to give risk assets some breathing room into Thursday and Friday.

The dollar’s muted reaction to the Fed hike was notable. Ordinarily, the first rate increase in three years alongside a hawkish dot plot would be expected to generate a much larger move. Instead, the USD largely wavered, suggesting the September decision had already been absorbed after Jackson Hole and the recent inflation data. The bigger FX move came from Japan, where a split BOJ hike failed to support the yen as much as traders had positioned for.

Subscribe now

Two drivers will be moving markets this week: follow-through from rates and a Trump-Xi meeting. So, let’s get into the guide to trades moving markets, where things stand, and where they may be heading.

• “Hawkeyed”

• “Two Strongmen in the Playground”

• “Summary of House Views”

Hawkeyed

We’ll take each meeting as it happened, as last week was all about rates. On outcomes, each decision aligned with market expectations, but what follows next differs.

Read more

The classic case is where two criminals are arrested and given the opportunity to reduce their sentences by testifying against each other.

A current-day illustration is that some of the most powerful leaders in artificial intelligence do not trust each other enough to commit to any slowdown in regulation.

Calls for restraint have grown in the industry, with Anthropic Chief Executive Officer Dario Amodei saying Saturday that the company would introduce additional safeguards, including independent third-party evaluations, and urged the broader industry to slow the pace of development of their most advanced models. OpenAI CEO Sam Altman backed the proposal, while xAI’s Elon Musk also weighed in:

The more powerful AI agents become, the more likely they are to deceive humans and the harder they can be to monitor. On the flip side, avoiding AI advancements deprives humans of technological improvements that may be needed, most notably in healthcare. This makes the mission of Amodei, and fellow pioneers such as OpenAI’s Sam Altman, to create superintelligent systems look increasingly difficult to reconcile with any safety ethos. The dangers of trying to design such models may simply be too great.

While people have discussed safeguards in the past, what happened over the weekend raises a new concern. This is not purely about safeguards, but about pacing the capability advancement outright.

For those who may not have read Amodei’s letter, it can be summarised in the following three sentences:

“Not building the technology deprives humanity of benefits or simply places AI in the hands of authoritarian powers, while building it too fast is reckless… We must slow the pace at which we improve the capabilities of AI models… Pacing does not mean halting model training or technical progress.”

As we read Amodei’s essay and reflected on last weekend’s conversations, we wanted to explore three questions. These were the first three lines written on this draft before we underwent any further research:

1. Can AI leaders put rivalry aside for the greater good?

2. How much pressure does China exert to reverse these ethical decisions?

3. Does this actually slow the equity AI trade?

What follows is a moral and ethical approach to AI concerns, and what outcomes we see as most realistic. And for those asking “what’s the trade?” we have the answer for that, too.

Subscribe now

Can AI Leaders Actually Put Rivalry Aside?

On the surface, this sounds reasonable. Anthropic presents itself as the most safety-conscious AI company, and Amodei (once Google’s lead AI ethicist) has talked about leading a “race to the top” in setting standards of vigilance for others to follow when building powerful systems.

But this was likely an impossible ideal from the start.

The schism between OpenAI and Anthropic is deeply personal and goes back several years. Anthropic was founded in 2021 by a group of OpenAI employees who broke away from the ChatGPT maker, spurred by the belief that AI systems needed a more responsible steward than OpenAI. Ever since, Anthropic has fashioned itself as a custodian of safe AI, with a dual focus on social purpose and profit.

It has also been a productive rivalry, with OpenAI and Anthropic competing for top AI talent, public favour, technological supremacy, and commercial dominance via their products ChatGPT and Claude. The companies are on similar timelines, both eyeing IPOs, although some news reports since Amodei’s letter seem to suggest any offering is being pushed back to 2027.

We can use a recent example of rivalry and ethics to analyse what may stem from the recent spotlight on “pacing the frontier”. In February, Anthropic had opposed any use of its AI models for systems that could autonomously kill people or conduct mass surveillance, then found itself in a legal firestorm as the Trump administration cut ties. OpenAI, meanwhile, signed its own Pentagon deal just hours later, arguing that it had independently negotiated acceptable safety safeguards. Google also followed suit.

In another sign of the pressure-cooker competitive environment, Anthropic publicly pulled back on commitments in its responsible scaling policy this year, a pledge to halt or withhold AI models if it could not guarantee proper safety measures in advance. Anthropic cited competitive pressures and a lack of faith in unilateral pledges as reasons.

Anthropic then began pulling back on its responsible scaling policy commitment: its pledge to halt or withhold models when it could not guarantee proper safety measures in advance. Anthropic said this was partly due to competitor pressure, coupled with a lack of faith in unilateral pledges that could be easily disregarded in the heat of competition.

So why have the companies continued to push AI development and revenue growth at speed until now, despite what they agree are potentially catastrophic risks to human lives? One well-reasoned answer to this is ideological and moral. Companies are convinced their systems can be made benevolent but do not trust their rivals, believing them flawed. The clearest example is the relationship between Musk and Altman, who frequently exchange messages online. More prosaic, but just as important, is fear that pulling the handbrake is worse than carrying on.

The Pentagon episode and Anthropic’s retreat are also evidence of a broader coordination problem: restraint by one firm opens up opportunity for another. The toughest act of Amodei’s intentions will be making restraint possible without conceding any competitive disadvantage for exercising it.

“The most effective method of pacing is via regulation that targets all US frontier AI companies, as that covers even those who are unwilling to cooperate voluntarily.”

Amodei’s essay is slightly vague about exactly how this could be done, though he does offer one concrete recommendation: external auditors should embed with companies to check that they are upholding agreed safety practices. Whether from naivety or hubris, Amodei still wants it both ways. He believes his company can safely build and control AI systems even as they become harder to steer. Yet in the past year, both Anthropic and OpenAI have, for competitive reasons, obscured details about how their models process information, making them harder to monitor.

But there’s an added cross-border difficulty: the possibility of pacing the development of powerful AI models in democracies depends on the US continuing to stay ahead of China. Pace slowing further than that would allow unpaced projects in China to move forward. Credible restraint across borders is also harder to achieve, since the incentives to defect are bound to be greater.

Can Any Safety Agreement Survive China?

Another few snippets from Amodei’s letter are worth flagging before answering the China question.

“If we slow down by more than this amount, then (unpaced) CCP-associated projects will pull ahead, creating significant national security risk… If we greatly restrain our AI capabilities in the belief that China will do the same, and then China defects… such a defection could lead to their geopolitical dominance.”

That leaves us with an uncomfortable contradiction. The idea that slowing things down might help us avoid an AI arms race depends on assuming the US has a lead large enough to risk slowing down without being overtaken. Armodei argues the US should keep its lead over autocracies as large as possible. Tighter restrictions on exports of advanced computer chips and semiconductor manufacturing equipment, and on the theft of AI models, would do that. If those measures work, he believes they could widen America’s lead over the next three to five years, precisely when AI is likely to become most geopolitically important.

Understandably, given Trump’s focus, unilateral restraint is unlikely. Kimi K3 and other relatively open-weight Chinese models are approaching the frontier and can, unlike a Chinese “closed-weight” version, spread beyond America.

A recent analysis on Kimi K3 versus peers on the Intelligence vs. Cost per Intelligence Index Task puts Kimi at a very favourable and attractive quadrant:

Even if all the leaders genuinely wanted restraint, it’s far from clear they could trust each other not to exploit it. That same thinking applies geopolitically. Altman, Amodei, and Musk may think restraint is the right choice, but each fears one of the others will defect, turning what is already a difficult problem of corporate coordination into something even harder — international politics.

Countries’ regulation of their own domestic frontier research could cover as much as is possible. But US regulation cannot ensure that China follows. As Amodei is explicit, his argument assumes that if the US adopts a policy of restraint from the Chinese Communist Party’s perspective, China may indeed follow only to defect, precipitating an enormous shift of power. Any agreement therefore needs either ironclad verifiability or limits narrow enough that cheating would not become militarily existential. There are many good reasons to believe that would not happen. We are sceptical of a global pause, given the enormous incentives to evade it.

What Does “Slowing AI” Actually Mean?

As for markets, will slower progress on frontier AI models significantly reduce demand for next-generation infrastructure and/or ongoing, total training spending? Slower progress at the frontier only matters materially for the equity trade if it leads to less demand for the infrastructure underpinning AI. Evidence in the research is much less dramatic, indicating that the largest pretraining runs account for only around 10-15% of total training spending, while inference and agentic workloads are increasingly important sources of compute demand.

Although leading developers would be likely to slow the growth of ever-larger pretraining runs if they focus more attention on making their models safe, those massive runs account for about 10-15% of overall training spending. The latest and largest runs require roughly $400-$500mn to run, so we estimate that slowing the growth of pretraining workloads will curb the rate at which spending rises by something like 10-15% annually, but not enough to significantly reduce total spending.

We expect more spending to go toward improving models after they have been pretrained. That includes reinforcement learning, use of synthetic data, use of tools, safety testing, and extra compute for especially complex queries.

Anthropic has recently ramped up its compute partnerships with Google and Broadcom, committing to use up to 3.5 GW of capacity. Anthropic also has alliances with Nvidia and Amazon to similarly expand capacity. Long lead times should also deter customers from giving up scarce compute capacity that competitors would likely take instead.

All told, our central case is that uncertainty over how fast frontier progress can proceed without risking inherent dangers probably will not materially cut back demand for next-generation infrastructure. Instead, pace-setting efforts likely will change where the compute is used.

Does Slowing the Frontier Actually Slow the Equity Trade?

The calls by Amodei, Altman, and Musk last weekend to slow frontier artificial intelligence do not alter our constructive stance on compute. The comments won’t materially slow demand in the next 18 months, especially considering Trump’s focus on preserving the US lead over China. Long lead times should also deter customers from giving up scarce compute capacity that competitors would likely take instead. Safety guardrails may reinforce demand: evaluation, monitoring, sandboxing and redundant inference can raise compute per workload without reducing training, inference or agentic workloads.

Increased calls for a slowdown in AI development may pressure equity valuations on near-term winners such as chipmakers and supply-chain line participants. But over the longer run, such comments shouldn’t have much impact because core infrastructure spending on AI compute will likely stay strong. One interesting perspective here is that a slower AI development path is positive, giving more time to extract returns from infrastructure already built.

The three CEOs agreeing to pace things down does not really change the money being spent on chips, power and infrastructure. In fact, it extends the development timeline. If commercialisation and adoption keep growing while the pace of new capability eases off a bit, that actually helps the shift from spending money to build things towards making money from what’s already built (e.g., monetisation).

The key distinction to note is that a slower frontier is not the same as a slower investment cycle (a genuine pullback in capex on AI infrastructure). A slower frontier on its own doesn’t negate what would be a compelling asset class theme for infrastructure valuations, assuming inference demand, adoption rates, and monetisation rates continue to grow. The far more serious risk is a genuine pullback in AI capex, with fewer data centres opened, fewer accelerators deployed, less power consumed, and a shorter, less aggressive build-out period. That scenario would make a safety-led slowdown materially bearish for the broader AI equity trade.

We hear you ask:

“So, What’s the Trade?”

Don’t worry, we have that covered also. Behind the paywall, we explore what “Pacing the Frontier” looks like in reality, what areas of markets benefit from our expected changes, and where we see the best long opportunities to deploy capital. (Yes, long ideas. It’s not the end of the AI trade just yet.)

Read more

The S&P 500 fell in each of last week’s first three trading sessions, before a Friday rally pared the decline. The percentage of S&P 500 members above the 50-day moving average has declined rapidly over recent weeks. Just 39% of the 500-member gauge now trades above its average.

Brent crude and West Texas Intermediate gained 5.9% and 6.7% on Thursday, the highest closes since May 19. Energy was the only sector reliably in the green, while industrials and consumer discretionary bore the brunt of selling. But most of last week’s attention was on the Treasury market. US Treasury securities sold off across the curve, with the long end leading. The 30-year reached 5.373% on Thursday; the “highest since 2007” crowd was back as levels rose above those that prompted Bessent’s August intervention. The bear steepening reverses the Jackson Hole flattening and suggests markets are pricing in an inflation shock over a credibility restoration.

The rally did not change the rates picture, with a Federal Reserve hike very much fully expected. Goldman Sachs Prime Services sees net and gross hedge-fund leverage remaining well below this year’s highs.

The US dollar strengthened to a one-week high on Thursday, as oil-driven repricing at the front end did the work August data could not. Higher crude lifted US rate expectations faster than G-10 peers, making the dollar a passive beneficiary. That dynamic reverses quickly if the Fed hikes and signals it is done. Wednesday’s FOMC statement becomes more important for the currency than the decision itself. Precious metals took the other side, with gold selling off sharply as real yields pushed higher.

Behind the paywall, we highlight over views on global debt and state the case for our convictions in the short-term, as well as the implications for equity markets and a new rates trade following the ECB meeting.

Subscribe now

Let’s get into the guide to trades moving markets, where things stand, and where they may be heading.

• “Insufficient Funds”

• “Elsewhere… One Hike, One Pause”

• “Summary of House Views”

Insufficient Funds

“I am the house now.” — Sec. Bessent, September 9, 2026.

Of course, Bessent was talking to bears looking to short the yen against his interests, but it also symbolises his goals in the bond market. There’s a confidence, a swagger. Bessent believes he cannot lose.

Read more

This dynamic traces back to the summer of 2024, when Macron suffered a crushing defeat in the presidential election to Marine Le Pen and her far-right National Rally, prompting Him to call a snap election in the hope of clarifying French politics and restoring his standing. But instead of providing clarity, the election destroyed the parliamentary logic that was allowing normal governance to proceed.

True, the snap vote did stop National Rally from taking control of France’s already precarious finances, but that was nearly all it achieved. A left-wing coalition finished with the most seats in the National Assembly; Macron’s centrist bloc survived the new political landscape with more legislative backing than one might have expected given his personal defeat, and Le Pen emerged as the biggest single political force in the country. But no party managed to seize the 289 seats needed for an outright majority, leaving the entire country effectively split three ways.

Stopping one party had made it nearly impossible for anyone else to govern.

For the last year, this period of prolonged parliamentary improvisation was characterised by a succession of prime ministers: Michel Barnier, François Bayrou, and, most recently, Sébastien Lecornu. The first two fell after each tried to repair France’s public finances, with Barnier lasting just 3 months in office, and Bayrou slightly longer, for 9 months. That it was necessary to appoint Lecornu for a fourth time just four days after he resigned shows the pretty desperate state of affairs within France. He survived largely becuase the moderate left conceded that extracting concessions from a weak government was preferable to bringing the whole thing down.

France has rivalled its arch-rival, the UK, in the political circus stakes.

But now the Lecornu minority government and its centrist allies have been plunged into its most treacherous season yet as opposition parties dig in over deficit-cutting plans designed to calm both the bond market and most of France’s disastrous finances. This all comes at an awkward juncture with a presidential election due in April-May 2027.

Ever since the Macron government faced a hung parliament following the June 2024 vote, it has tried, and largely succeeded, in using Article 49.31 to force its proposed budgets through without debate, much less a vote. It is true that the opposition may respond with a no-confidence motion, which would force the government to either make concessions to pass its budgets or accept collapse. Yet there is little appetite to compromise at a time when opposition parties are positioning themselves for the 2027 presidential race.

The risk is that it would become impossible to agree on a full budget for the second half of 2027 as the next president dissolves parliament and calls new elections at some point while work was beginning on the 2028 budget. Such a rollover would cause unprecedented budget paralysis, freezing investment and planned defence-spending increases while welfare costs keep climbing. That would widen the deficit by at least half a percentage point, denting investor confidence further and pushing up borrowing costs.

Public debt in France stands at 117% of gross domestic product, expected to rise to 121% by 2027, to be among the largest in the euro area. Behind the US, France also runs the largest deficits, without any forecast that this is set to change soon. The International Monetary Fund (IMF) projects a 2026 deficit of 5.2% of GDP versus a budget target of 5.0%. A series of proposals for fiscal consolidation now being debated in parliament are politically dangerous at a time when candidates are fighting for dominance in the heartland of Blue France.

La Toile de Fond (The Backdrop)

Early indications are that President Macron’s final finance bill may be the most perilous of his presidency so far.

The excess that France pays to borrow money over German bonds (a key barometer of investor mood) is near its highest level since the sovereign debt crisis of 2012. Political and fiscal worries always manifest first in debt markets.

France has had no shortage of problems, but the first presidential debate was consumed by bickering over how best to deal with rising debt. In particular, there was robust dispute over a far-left proposal to cancel a portion of debt.

The on-stage questioning gave the far-right and far-left candidates a chance to stress-test unorthodox ideas. The leading far-right and far-left candidates, Marine Le Pen and Jean-Luc Mélenchon, have vexed senior business leaders in the past with disruptive propositions and have seen their repeated runs for president fall short of victory.

After three hours of hard-hitting probes, the contenders agreed only on the need to narrow the runaway budget deficit as a matter of urgency. The options ranged from clawing back welfare spending to hiking taxes, raising the retirement age, juicing the economy with more spending, or even cutting payments to the European Union.

Mélenchon doubled down on his plan to cancel debt held at the European Central Bank as a negotiable package with other euro members, telling Reuters he would be willing to go further and “just take the bonds and burn them.” We have heard similar appeals before: in Europe as the euro zone debt crisis detonated, and around the same time in the US. Debt cancellation, however, is nothing more than monetary financing and would be highly inflationary.

As these ideas increase, the case for investing in real over financial assets only strengthens. Mélenchon may have said it, but this is in no way a realistic outcome. Yet, even if low-probability, high-impact events don’t come to fruition, they materially warp the distribution of risks, skewing it with fatter tails. We should take them seriously, particularly now that the US Treasury has panicked over its long-end buyback announcement.

We expect to hear more avant-garde debt approaches, such as cancellation, in the months ahead and not just in France.

Former prime minister Philippe said the debt effect “risks generating an extremely brutal shock that will hit hardest those most exposed.” He dismissed Mélenchon’s proposal as “extremely dangerous” and risking cutting France off from financing.

Le Pen, meanwhile, claimed her savings plan of €125bn ($146bn) could be achieved by shrinking the French state, cutting immigration, and possibly reducing contributions to the EU. She proposed a constitutional “golden rule” to ensure balanced budgets, saying she would make sure France honoured its debts.

She reiterated her pension reform, allowing retirees to cash in on their schemes as soon as age 60, which she puts at around €9bn a year.

What options does France have to reduce its debt? Fiscal consolidation is too risky to drum up the votes for, and growth is already heavily driven by massive deficits (and forced down by austerity in the past). Inflation is already an issue for rising interest payments. Financial repression is coming, but too late to save the day (it’s already partly at work via enhanced Treasury buybacks). Selling off government assets such as the gold that sits in Fort Knox may make a one-off splash, but it’s unlikely to move the dial. Debt restructuring or default would cause more damage than good.

Macron seems stuck in a bind: no parliamentary majority, no budget room, and no political capital. It is not obvious what Emmanuel Macron can do to stop the bleed. Instead, the risk is that the bleed accelerates.

The 15y15y forward rate is a way of looking at borrowing costs that strips out the impact of near-term monetary policy. The spread between this metric for France and the equivalent German rate is now just short of its highest level since 2012, a sign that investors are pricing a structural deterioration in France’s financial health relative to Germany in the coming decades.

So what is the story for investors? France’s fiscal arithmetic may be ugly in aggregate, but markets will not experience the adjustment in aggregate. Any credible attempt to stabilise the debt trajectory will create winners and losers across sectors, households, and asset classes depending on where the government chooses to find the money. Tax rises, spending cuts, pension reform, changes to corporate incentives, and further pressure on household savings would each transmit through the French equity market in very different ways.

The approach, therefore, is to determine how it tries to consolidate, who ultimately pays, and where that pain is likely to appear first in markets. We start here, then turn back to the sovereign complex and lay out our convictions on OAT yields and the relative-value trades we prefer across European government bonds.

Below the paywall, we detail our favoured stock-specific long/short trade, which we believe has asymmetric opportunities, along with three rates trades, including one relative play that we have added to our portfolio.

Equity Approach

Constituents of the country’s benchmark CAC 40 Index make less than 20% of revenues at home, which limits the impact of political risk on earnings. However, domestically focused sectors such as banks, utilities, telecoms and construction do feel the stress of a wider French-German yield spread. A bigger yield gap reduces the attractiveness of French stocks and keeps a lid on companies’ ability to invest, ultimately hurting their competitiveness.

Read more

Technology led most of the week, but AI-related weakness persisted, and Broadcom shares fell despite a ninth consecutive beat and sharply higher AI semiconductor revenue; the pattern also appeared in Cisco, Samsung, and Nvidia, as fourth-quarter guidance missed expectations and weighed on shares. Energy stood out as crude oil was on course for a strong weekly gain, helped by the US and Iran resuming military exchanges in a conflict now in its seventh month.

The hawkish reaction, which saw rates rise broadly across the curve, reflected the labour market being behind part of the Fed’s mandate, with the full picture dependent on next week’s CPI figures. Fed’s Waller was framed as a proxy for the committee centre, noting a preference for holding steady if the path of disinflation resumes, and that a hike vote is likely to depend on this week’s CPI. USD edged higher on the payrolls data after retreating through Thursday when falling yields reopened a downtrend through most of August, with the move seen more about positioning than a lack of willingness to price in higher rates.

Into next week, the balance of risk leans towards further USD strength unless CPI is soft. The rates move wasn’t sustained, with 5s30s up only 3 bps over the week, suggesting the market continues to seek more concrete signals on the Fed’s policy path, and most attention shifted to next week’s inflation figures.

Let’s get into the guide to trades moving markets, where things stand, and where they may be heading.

• “September Showdown”

• “Europe Politics”

• “Summary of House Views”

Below the paywall, we look into last Friday’s US labour data and how that plays into the September FOMC decision, alongside our call on a hike or hold. We also share some insights on Europe and the political forces at play in local markets, where the stage is set for a bearish impulse on euro-area bonds to emerge ahead of this week’s ECB meeting.

A France thematic article on the 2027 elections will be sent to all paid subscribers later this week. If you are not yet a paid subscriber, you can upgrade below to receive all published research.

Subscribe now

Read more

Markets got some good news from Nvidia and some not-so-welcome news from Warsh. Nvidia did its job in validating the AI complex. Semiconductors were higher, and it gave the S&P 500 its best session since early August. Warsh then stepped up to the podium and made his inflation concerns clear, adding to the hawkish pricing in rates markets.

We’ll touch on Nvidia specifically later on, but elsewhere, software was strong and delivered some much-needed evidence that the AI opportunity is broadening, now a positive for software. Salesforce surged after stronger guidance and an expanded Anthropic partnership, while CrowdStrike rallied to a record as AI-related cybersecurity demand accelerated. It’s a continued case that “SaaSpocalypse” is company-specific. There will be winners and losers in the sector.

Breadth was weaker. Most S&P 500 sectors still finished lower. The relative attraction of equities is also becoming harder to ignore with the 10-year Treasury yield above 4.7%, leaving investors able to earn more income from government bonds than from almost every stock in the index.

Other markets responded largely to the Jackson Hole symposium speech, but we look more into this later on.

Below the paywall, we outline in detail the guide to trades moving markets, where they may be heading, and the actionable trades to consider.

• “Warsh’s Judgment Day” — cross asset implications from Jackson Hole

• “Investment Case for NVDA and META” — why Nvidia can outperform and the overhang that is now gone from Meta’s outlook

• “Summary of House Views” — recapping all macro trades on the book

Subscribe now

Warsh’s Judgment Day

Read more

UK equities trade at their widest valuation discount to developed-market peers in decades. We believe the tide is already turning, and highlight the stocks best positioned to benefit.

With the UK macro backdrop central to the thesis, and the UK being our home, we felt the team was well positioned to offer a strong and reputable voice on the topic.

At the same time, we teamed up with , a leading Canadian-based equity research firm, to leverage their equity-specific knowledge and expertise. The piece includes their latest discussion with ARM Holdings (NASDAQ: ARM), one of the UK’s largest companies and a global leader in semiconductors.

UK equities lack the AI exposure that drives valuations in the S&P 500, which has clearly contributed to the valuation gap. But there is much more to the story, including shifting investor sentiment, domestic capital reform, and a sector mix better suited to the next cycle.

Table of Contents

1. The State of the UK Market

2. Why We Are Bullish on the UK

3. Growth Driver 1: Legacy Investor Ignorance

4. Growth Driver 2: Political Capital

5. Growth Driver 3: Index Composition For The Next Cycle

   5.1 Financials

   5.2 Energy & Materials

   5.3 Healthcare

6. In Conversation: Arm Holdings

7. UK Equities: Pool of Ideas

8. Key Risks

9. Concluding Thoughts

Subscribe now

1. The State of the UK Market

Over the past decade, UK assets have significantly underperformed their global peers. The clearest example is the FTSE 100, the UK’s flagship stock index, which has gained 62% over this period, compared with a 254% gain for the S&P 500.

However, FTSE 100 performance has started to turn over the past year. With almost no tech exposure, the UK market’s gains have been driven mainly by financials, defence, and expectations of lower interest rates.

We then turn to the British pound (GBP), which was trading at 1.39 against the US dollar in August 2021 and remains at a similar level today, at 1.36. Notably, GBP/USD fell to a low of 1.035 in September 2022 amid growing political concerns in the UK.

Next, we turn to the bond market. The long end of the Gilt curve has risen significantly in recent years, with both 10-Year and 30-Year UK government bond yields reaching their highest levels in decades. Loose fiscal policy, weak economic growth, and persistent inflation have pushed yields higher as investors demand greater compensation for holding UK government debt.

The UK’s move has been partly driven by a global selloff in long-term bonds. 30-year US Treasury yields reached their highest level since 2007, while German 30-year yields reached levels last seen in 2011. However, the UK stands out for both its higher absolute yields and the large increase in the extra return required to hold longer-term UK debt. This reflects concerns over government spending, weaker demand for UK bonds, and persistent inflation, all of which have been more pronounced in the UK than in most other G7 economies.

Taken together, these three asset classes paint a consistent picture. Over the past five years, the market has demanded a higher risk premium to own UK assets.

Equities have been held back by weak earnings expectations, political uncertainty, and continued capital outflows. Sterling has struggled as weak productivity growth, a current account deficit, and inconsistent government policies have weighed on international confidence.

Meanwhile, Gilt yields have risen significantly as the market has demanded more compensation for lending to a government facing higher debt, large borrowing needs, and an economy struggling to generate sustained growth.

In our view, there is some irony here: these same headwinds have also created the foundations for an opportunity. The UK is starting from a much lower base, while some of the factors that held the market back are beginning to improve.

2. Why We Are Bullish on the UK

We split our reasoning into three growth drivers.

The first is what we call legacy investor ignorance.

Despite many of the UK’s largest companies generating the majority of their revenues overseas, they continue to trade at a sizeable valuation discount to peers in the US and other developed markets. This discount may have been justified in the years immediately following Brexit, but six years later, it increasingly looks disconnected from the fundamentals. We see this as our first growth driver because the market is already starting to recognise this valuation gap.

Takeover activity has also picked up significantly, as overseas buyers and private equity firms continue to take advantage of depressed UK valuations. UK public M&A reached £35.3bn in the first half of this year, compared with £22bn in H1 2025 and £18.5bn in H2 2025. Overseas bidders accounted for 94% of total deal value.

Inbound cross-border M&A activity tells a similar story, with foreign buyers showing a growing appetite for UK companies.

Meanwhile, the FT recently reported that takeover activity involving UK-listed companies has surged 188% year-to-date, marking the strongest pace since 2007.

A second pillar of our bullish view is the political capital being spent by new PM Andy Burnham to kick-start the UK economy. He may have been in office for only a few weeks, but the market’s initial reaction has been broadly positive.

More importantly, the government’s early priorities have focused on unlocking long-term investment rather than short-term fiscal giveaways. Alongside an ambitious programme of planning reform and housebuilding, Burnham has repeatedly emphasised the need to “mobilise British savings to back British growth.” Ministers are also exploring reforms to ISAs and pension funds that could encourage more domestic capital to flow into UK-listed companies.

Even a modest shift in domestic capital allocation could have a meaningful impact on UK equities. We explore how even a partial reversal of pension fund outflows from domestic markets could bring billions of pounds of additional demand into the FTSE 100 and FTSE 250.

Finally, at the index level, the UK market is dominated by large companies in healthcare, commodities, and financial services.

These are three areas where we are currently overweight and expect to remain so in the next cycle, with each benefiting from different tailwinds. As a growth driver, we believe the market is underestimating the potential upside in these sectors as we enter a period of higher infrastructure spending, larger defence budgets, strong commodity demand, and fundamentally higher interest rates.

3. Growth Driver 1: Legacy Investor Ignorance

The shrug of the shoulders when some overseas investors are told about UK equities tells the broad story here. Even though ignorance might be too strong a word, investor indifference and persistent under-allocation certainly help explain the valuation gap between the UK and its developed-market peers.

The scale of the discrepancy is being arbitraged away, but still remains striking. As of the end of July, the MSCI UK traded at 12.85x forward earnings, against 20.37x for MSCI USA and around 15.5x for MSCI EAFE, while offering a 3.05% dividend yield, versus just 1.13% in the US.

In other words, investors are currently paying roughly 37% less for a pound of forecast UK earnings than for the equivalent dollar of US earnings, while simultaneously receiving almost three times the dividend yield.

Of course, some discount is justified. The US market contains many of the world’s highest-growth and highest-margin technology businesses, and those companies deserve higher multiples than mature banks, miners or oil majors.

But we believe the valuation gap is still too large.

We don’t need the valuation gap to disappear entirely for the return implications to become meaningful. If UK forward valuations moved from 12.85x to 14x, for example, that alone would represent roughly 9% multiple expansion.

UK vs. US Equity Valuation Gap (2023–2026)

Signs of this process are already emerging. The FTSE 100 has lagged the S&P 500 by only around 250bps over the past year, an impressive result when considering the radically different sector compositions of the two indices (more on that later).

Arguably, the most compelling evidence that the UK discount is becoming excessive comes from ongoing corporate arbitrage. Financial and strategic buyers appear increasingly willing to exploit a valuation discrepancy that public-market investors have been slower to recognise.

By July 1, 2026, announced M&A involving UK targets had exceeded $231bn, 210% above the same point in 2025. Even more strikingly, foreign bidders represented around $197bn and a record 86% of UK-targeted deal value.

It’s not hard to sniff out that when overseas corporations and private-equity firms repeatedly conclude that UK-listed businesses are worth substantially more than their prevailing market capitalisation, there’s money to be made.

Companies themselves are reaching much the same conclusion.

UK companies had announced £34.6bn of share buybacks by April 24, 2026. Management teams are opportunistically retiring capital for shares.

This creates an interesting dynamic. Foreign buyers are removing undervalued UK companies from the market, while listed companies are simultaneously shrinking their own share counts through buybacks. Both reduce the available supply of UK equity. Then if we see a continued demand increase from other marginal buyers for UK stocks, it creates a rising tide.

Just as there is some ignorance around ignoring the UK, we also don’t want to be seen as ignorant in simply saying “UK stocks are cheap”.

Cheap markets can definitely remain cheap indefinitely.

The more compelling argument is that identifiable shifts are now starting to close this valuation gap, creating opportunities for investors to generate meaningful alpha through stock selection. Rather than simply buying the cheapest UK stocks, we favour companies where the valuation discount looks hardest to justify.

For example, international businesses still trading on UK multiples, companies with strong free cash flow and significant buyback capacity, and businesses whose strategic value appears materially higher than their public-market valuation, making them potential M&A targets.

Mid-caps are particularly interesting in this respect, as lower analyst coverage and weaker domestic flows can create more pronounced valuation gaps.

Now, let’s take a look at two ideas we find interesting.

4. Growth Driver 2: Political Capital

Andy Burnham entered Downing Street on July 20, 2026, promising a “new political model and a new economic model” after replacing Keir Starmer as Labour leader. The UK has become something of a political laughing stock in recent years, with Burnham becoming the country’s seventh PM in a decade.

Andy Burnham’s First Speech as Prime Minister

The consequences extended well beyond politics, with repeated fiscal missteps triggering serious market turmoil. The clearest example came in September 2022, when Liz Truss’s government, with Chancellor Kwasi Kwarteng, unveiled a mini-budget containing around £45bn in unfunded tax cuts.

Markets reacted badly, with gilt yields surging and severe pressure building across pension funds, particularly those using liability-driven investment strategies. To prevent the stress from turning into a broader financial crisis, the Bank of England stepped in with a temporary programme to buy long-dated gilts, initially offering up to £65bn in purchases. In the end, it bought £19.3bn of gilts before gradually unwinding the entire position once market conditions stabilised.

Even though this shows the damage that can be done, the opposite is also true. Burnham does not need to transform productivity overnight.

He needs to convince companies and investors that the direction of travel has changed, that approved projects will actually be delivered, and that domestic capital will no longer be encouraged to bypass UK assets.

In a market priced for disappointment, simply reducing policy uncertainty can be almost as valuable as an immediate improvement in earnings. Before we even get to the policies, there are already some positive signs. Labour is now polling ahead of Reform UK, the right-wing alternative in the UK’s increasingly fragmented political landscape. The recent improvement largely reflects growing confidence that Burnham can deliver the change he has promised.

What has been more constructive is the absence of a sustained fiscal-risk spiral. The 10-year Gilt yield is almost unchanged if we compare the current yield to the day Burnham took office.

Of course, yields remain high in absolute terms, but the market has so far treated Burnham’s programme as a conventional fiscal debate rather than a broader credibility issue. So, what policies could drive this change, and which sectors should we watch? One of the most interesting levers is savings reform.

UK workplace defined-contribution (DC) pensions held roughly £600bn in assets in 2023, but only 6% was invested in UK equities, compared with 70% in overseas equities. The domestic share of DC assets has also fallen from just over 50% in 2012 to just over 20% today.

The government’s own analysis estimates that a five-percentage-point increase in domestic DC allocation would add £30bn of investment into UK markets. At the same time, ISA reform will reduce the under-65 Cash ISA allowance to £12,000 from April 2027, while keeping the overall £20,000 ISA allowance unchanged. The goal is clear: encourage more savings to flow into investment products.

Given how little the current ISA and pension system supports UK equities, this growth driver does not require a dramatic change. From our calculations, a one-percentage-point increase in DC home bias, a modest shift in ISA cash savings, and a half-percentage-point increase in LGPS listed-equity allocation could generate roughly £6.7bn of incremental demand for UK-listed equities. The impact could be particularly meaningful for the much smaller FTSE 250.

Pinpointing exactly which stocks would benefit is harder. These funds would act as marginal buyers, making the impact difficult to trace directly.

However, we see two clear points.

First, money is likely to flow toward companies seen as undervalued. This supports the first growth driver we discussed above and could help close some of the valuation gap.

Second, it matters less where all the money ultimately ends up, as much of it will pass through UK asset managers, brokerage platforms, insurers, and other financial institutions. These businesses should benefit from higher AUM, greater trading activity, and increased demand for actively managed products.

There are, of course, other ways to play these reforms, such as homebuilders benefiting from potential housing reform or retail and leisure companies benefiting from measures aimed at easing the cost of living.

But if we want direct exposure to the financial reforms themselves, we believe UK financials offer one of the most attractive risk/reward opportunities.

5. Growth Driver 3: Index Composition For The Next Cycle

The lack of AI and broader technology companies in the UK public market is no secret. In theory, this should have caused the UK to lag the S&P 500 even more, given the huge contribution from tech and AI-related stocks.

Yet that has not been the case this year.

YTD, the FTSE 100 is up 8%, compared with 12% for the S&P 500.

Ironically, one factor supporting the UK market is its index composition. While the FTSE 100 has little exposure to technology, it has large weightings in sectors that are already performing well. Financials account for roughly 28% of the index, followed by healthcare at 11.6%, energy at 10.4%, and materials at 8.1%.

The contrast with the US is significant.

Information technology alone represents 36.8% of the S&P 500 and 28.9% of the MSCI World. This means the UK offers investors a very different mix of sector exposure, with greater weightings toward financials, healthcare, energy, and materials, all areas where we see strong potential in the next cycle.

So why does this setup look appealing going forward? Let’s focus on three sectors we are bullish on right now: financials, healthcare, and energy.

5.1 Financials

UK banks have performed strongly over the past year as the interest-rate backdrop has remained far more supportive than many investors initially expected. The Bank of England held its policy rate at 4% through much of 2025 before cutting by just 25bps to 3.75% in December, where it has remained.

This has allowed banks to maintain healthy net interest margins and deposit spreads. Looking forward, we see this continuing based on our macro outlook for monetary policy. The BoE held its policy rate at 3.75% at its July 2026 meeting, with a 6-3 vote and three MPC members voting for a hike.

Current market pricing indicates 25.4bps worth of hikes priced through to the end of the year. A “higher for longer” or even modestly higher rate path, driven by energy price inflation and Middle East tensions, would be incrementally positive for bank NIMs.

Importantly, the benefits of higher rates have increasingly been reinforced by fundamental hedges. Barclays, Lloyds, and NatWest have large hedge books that gradually roll onto higher yields, providing a delayed boost to net interest income even if the BoE cuts rates modestly. All three have guided towards year-on-year growth in fundamental hedge income over the medium term, making the earnings benefit more durable than simply relying on rates staying high.

When you put it all together, it’s clear that the full benefit of this cycle has yet to be reflected in earnings.

The next leg of this cycle could come from a further re-rating of the sector alongside supportive monetary policy. Investors are increasingly recognising that the improvement in profitability since Covid is more fundamental and not simply the result of higher rates. With UK banks still trading within a domestic equity market that remains heavily discounted versus other developed markets, stronger returns and better earnings visibility could give investors a good reason to close at least part of that valuation gap.

5.2 Energy & Materials

Over the past 12 months, UK materials stocks (FTSE 100 members) delivered a total return of +80%, while energy stocks returned +34%. Materials were the standout sector across the entire index, comfortably ahead of even Technology (+55%) and Financials (+26%).

The main driver has been the broad rally in commodity prices over the past year, with the Iran war acting as a key catalyst. However, it would be wrong to attribute all of these gains to the oil complex.

In fact, copper has been the strongest performer across the major commodities, rising 44% on supply deficits and recovering Chinese demand.

Next is Brent crude, up 34%, driven higher by the Iran conflict and related disruptions to maritime trade. Gold is slightly behind at +32%. While it struggled for much of Q2, it remains supported by central bank buying, a weaker US dollar, and safe-haven demand. Not all commodities have performed equally, however, with natural gas (-1%) and iron ore (-7%) standing out as notable laggards.

Looking ahead, we don’t believe this is the end of the road for materials outperformance, even with precious metals having pulled back in recent months.

Copper is a good example. Physical copper markets are likely to remain in deficit in 2026, as production growth cannot keep pace with rising consumption, supporting further price increases. Copper is forecast to average $12,800/t in 2026, with demand supported by China’s power grid, renewable energy, EVs, and machinery, while supply remains constrained.

More importantly, we expect the copper deficit to persist well beyond 2026, creating a longer-term supply problem for the market.

The fundamentals are highly compelling. A decade of underinvestment, combined with the rising capital intensity of new projects, means higher prices will likely be needed to incentivise new supply. This is already reflected in the FTSE 100, with the Anglo American–Teck merger moving towards completion and expected to create the world’s leading listed copper company.

Copper isn’t the only metal expected to face a deficit over the coming year. Aluminium is another market where the supply picture looks increasingly tight. Middle East disruptions are estimated to have caused 3.5–4.0 Mtpa of outages, with around 2M tonnes of regional supply expected to be lost in 2026. If these disruptions continue, aluminium prices could move above $4,000/t.

Regarding gold and silver, we believe the recent bounce still has further room to run higher. More broadly, UK energy and materials stocks still trade at low valuations despite their strong performance.

The median forward P/E is 10.5x, while median forward EV/EBITDA is 4.7x. If commodity prices stay high, we believe these stocks still have room to move higher as earnings grow and valuations improve.

5.3 Healthcare

The case for UK healthcare stocks is becoming increasingly attractive following a prolonged period of neglect. The sector has faced real headwinds, both in the UK and globally, including drug-pricing reforms and reimbursement pressure.

There have also been concerns that AI could reduce rather than improve margins. These factors have led investors to use healthcare as a source of funding for more popular AI and tech names, leaving positioning in the sector unusually light.

We believe this neglect is now part of the opportunity, especially given the UK’s limited exposure to healthcare. Valuations also look reasonable relative to the broader market. The FTSE 350 trades at an average forward P/E of 15.2x, while GSK trades at 10.4x and Hikma at 9.6x. On a P/B basis, versus a 3.5x average for the FTSE 350, Smith & Nephew trades at 2.4x and Hikma at just 1.9x.

Positioning also remains far from crowded, meaning even a modest rotation away from expensive tech names could provide a meaningful boost to healthcare stocks.

The fundamentals offer another reason for optimism. Healthcare companies typically benefit from strong pricing power, high margins, and limited exposure to commodity inflation. An improving earnings environment could therefore create meaningful operating leverage as margins recover. Meanwhile, healthy balance sheets and the need for large pharmaceutical companies to replenish their drug pipelines could support further M&A, particularly across biotech.

We also think the AI angle is under-appreciated.

Investors have largely focused on AI as a threat to healthcare margins, but productivity gains in areas such as drug discovery, diagnostics, and administration could eventually create meaningful upside to earnings.

6. In Conversation: Arm Holdings

Earlier this month, Aurelion Research had a really insightful discussion with ARM Holdings’ VP of Investor Relations, Ian Thornton, who has been with the company for more than 20 years. Let’s just say he knows the ARM business pretty well, and we definitely felt it during the conversation.

For context, ARM is one of the largest companies in the UK and one of the most important semiconductor companies globally.

It develops energy-efficient CPU architectures and licenses them to major technology companies around the world.

It was a great call, and they came away with much more than they typically get from larger companies. Bigger companies often stick to the same points they make on earnings calls. These conversations are about getting a better sense of growth drivers, what customers and investors are saying, as well as how competitors are moving.

First, the long-term margin opportunity is much more interesting than initially thought.

Financial Targets & A Very Different Margin Model

The IP business could become very attractive, with management targeting around a “99% gross profit margin” and potentially 60-65% operating margins.

Those numbers may sound aggressive at first, but ARM has a very different cost structure from most semiconductor companies. The reason is that “everything that goes into that chip is IP that we also license and sell.”

As management put it, “chip costs are just assembly... basically assembly costs.”

That makes the gap between gross and operating margins “a lot, lot tighter than most other semiconductor companies.” Longer term, reaching operating margins around 60% would require moving further toward customer-owned tooling, where “you own the tools that live at the test equipment factory.”

The point here is that ARM is leveraging its existing IP advantage to move further up the value chain while maintaining a relatively asset-light model.

Bringing More of the Chip Development In-House

ARM is also gradually bringing more of the chip development process in-house.

For the first chip iteration, ARM used SocioNext for the back-end processes. For the second, it moved toward a direct relationship with TSMC.

The company also recently completed its acquisition of DreamBig, giving ARM greater capabilities in high-speed interfaces. Previously, it relied on third-party tools from companies such as Synopsys.

The strategy is simple: bring more of the development stack in-house, have greater control over the product, and capture more of the economics.

This also addresses a broader issue ARM encountered when it first started approaching hyperscalers. According to management, the response was essentially: “We don’t want to design chips. We are software and service companies. Don’t make us become silicon designers.”

That is important because ARM is not necessarily asking customers to become chip designers. It can provide much more of the finished solution itself.

And there is a meaningful economic opportunity here.

Management pointed out that if “90% of the IP that goes into that chip is ours,” there is a question of why ARM should let someone else capture a 30% cut for assembling what is largely its own architecture into a finished product.

Enterprise Could Be a Much Bigger Opportunity

One of the more interesting parts of the conversation was around enterprise demand. Companies don’t want to design its own chips. Instead, the question they were asking was essentially: “Why is there no one to sell me this chip?”

That is where ARM saw an opportunity. If “90% of the IP that goes into that chip is ours,” there is “very little opportunity” for competitors to differentiate.

That creates a pretty unique position for ARM. Customers get a chip tailored to their needs, while ARM can capture a significant portion of the economics because it already owns most of the underlying IP.

We think this could become an increasingly important part of the story.

The hyperscalers have already shown that custom silicon can make sense, but there is a much larger group of enterprise customers that simply does not have the resources or desire to build an internal chip team.

Competitive Moat & Future Workloads

When it comes to competition, management was pretty direct:

“The only boulders on the runway are Intel and AMD with their x86 chips.”

ARM’s core advantage remains performance per watt. As they put it, “the primary performance metric customers care about today is performance per watt.”

That becomes increasingly important as AI infrastructure puts more pressure on power availability. The computing industry can keep adding GPUs, but eventually you run into a simple problem: there is only so much power available.

This is also where agentic AI becomes particularly interesting.

Management believes “the rise of agentic workloads” could be an important opportunity because “the agentic is pretty much 100% a CPU task.”

GPUs handle the heavy matrix computation, but the orchestration, API calls and sequential logic around those workloads are largely CPU-driven. If agentic workloads become a meaningful part of future computing demand, ARM could benefit from another structural shift toward more efficient CPU architectures.

Robotics & Physical AI

The other area we found particularly interesting was robotics and physical AI.

These applications require much more than raw computing power. They need “redundancy, safety islands, parallel processing.” As management explained, you need “safety critical systems to go into robotics and cars” because a computing failure in these environments can have very serious consequences.

That plays directly into ARM’s strengths. As more computing moves into environments where power efficiency, reliability and safety are just as important as raw performance, ARM’s architecture should become increasingly relevant.

Overall, the biggest takeaway for us was how broad ARM’s opportunity could become. AI infrastructure is only one part of the story. The company is gaining exposure to data centers, enterprise computing, agentic AI, robotics and physical AI, all of which are pushing demand toward more efficient computing. After this conversation, we came away with a much clearer view of why management is so confident in ARM’s long-term growth potential.

7. UK Equities: Pool of Ideas

We built a Pool of UK Ideas classified by the three growth drivers we identified earlier, with exposure to financials, healthcare, commodities, and other areas where we see the strongest potential for a re-rating.

This basket is live and investable for eligible accounts via Plutus and can be viewed here.

8. Key Risks

The simplest way to explain the main risk from our primer is that UK equities are not necessarily expensive, but the catalysts needed to unlock their value may fail to materialise.

1. The Valuation Gap Could Persist

The UK discount, particularly in equities, could prove to be more fundamental than we expect. Although we believe it is cyclical and already starting to change, our thesis assumes that the valuation gap continues to close as international investors rediscover the market. Cheap, however, does not automatically mean mispriced, and investors could continue to look elsewhere for years.

2. Reforms Could Disappoint

Another risk is that Burnham fails to turn political momentum into meaningful economic reform. The early days of any government are not always a good indicator of future investment performance. A key part of our thesis is that greater political stability and policies aimed at mobilising British savings can improve domestic investment. The risk is that implementation disappoints, reforms are watered down, or political opposition prevents meaningful change.

3. Fiscal Risk

Finally, there is fiscal credibility, an area where the UK has struggled in recent years. Burnham has so far avoided a major fiscal-risk event, but the October Budget will be an important test.

Higher borrowing, unfunded spending commitments, or overly optimistic growth assumptions could push gilt yields higher and put pressure on sterling and UK equity valuations. The 2022 experience under PM Truss showed how quickly UK assets can reprice when confidence in fiscal policy is lost.

9. Concluding Thoughts

Here we fly the flag for the UK.

Expectations remain relatively low, valuations remain attractive, and for the first time in a while there are identifiable catalysts capable of changing the narrative.

Importantly, our bullish case doesn’t require Britain suddenly to become the world’s fastest-growing economy. That’s just not realistic. Rather, we need the direction of travel to improve. The signs are there that this can be the case, from credible fiscal policy and a modest redirection of Britain’s enormous savings pool.

AP and Aurelion Research

Disclosures & Methodology

All forecasts and estimates are based on publicly available information, management commentary, and our own modelling assumptions. This report reflects independent analysis by Aurelion Research and/or AP Research. All opinions and estimates expressed herein reflect the author’s judgment as of the date of publication and are subject to change without notice. This publication is for informational and educational purposes only and does not constitute investment advice, a solicitation, or a recommendation to buy or sell any security. Aurelion Research, AP Research, and its analysts may or may not hold positions in securities discussed. Use only where legally permitted.

Alibaba’s (BABA US, 9988 HK) $10.2 billion (HK$80 billion) share sale, Hong Kong’s biggest follow-on offering on record, is one of them.

Markets have immediately focused on dilution. Shares fell around 8% after the announcement, and criticism followed. Even Michael Burry, a now-native writer on Substack, announced that he had moved his entire Alibaba position in favour of JD.com, and said he would not be interested in Alibaba unless the share price were half what it is now. While Burry stated that he “cannot bless share issuance”, we contrastingly see this action as positive for Alibaba’s pursuit in the AI race and believe the investment, after some short-term pain, will be rewarded.

We are not here to rebut the opinion shared, but instead to highlight what the market is overlooking. Focus is too heavy on the shares issued and too light on what those shares buy Alibaba. The raise provides fuel for the AI buildout while preserving cash, limiting additional leverage, and retaining balance sheet flexibility for what could become an even more capital-intensive arms race.

We don’t often drill down into the micro details and publish a report on a single stock, but today readers can enjoy a different approach. A top-down overview of our bull case for Alibaba in the AI world, and a bottom-up analysis from contributor , an independent equity investor and researcher.

Preserving Debt Capacity

The major concern with share issuance is diluted earnings. That in turn hurts stock performance in the near term. It is a valid concern for an investor, but it does not, by itself, determine a company’s prospects. More has to be considered here.

Alibaba’s fundraising will help drive AI-led growth, and the impact on earnings will be limited if the company can improve AI payoffs. We see this raise as a bigger contributor to the AI “arms race” than the dilution itself.

Let’s start with the mechanics; after all, a few specifics never hurt anyone. The company issued 710 million new shares at HK$112.70, equivalent to roughly 3.6-3.7% of its enlarged share count, priced at an 8.4% discount to Friday’s HK close, with the proceeds earmarked for its full-stack AI capabilities. This creates a mathematical headwind for existing shareholders and was likely a contributing factor in Burry’s decision. Corresponding EPS dilution will be somewhere around 3.5%, but focusing purely on dilution risks missing another consideration:

Why equity? And why now?

Alibaba is hardly a distressed company scrambling for repairs. The balance sheet is strong, liquidity is ample, and in recent years the company has spent cash buying back stock, reducing share count by more than 11%. Switching tack towards issuing shortly after comes across as a slightly awkward capital-allocation approach. Retire equity, and then come to the market asking for more?

On the surface, it seems strange, but there is some logic to it. We would distinguish between needing capital and choosing how to finance growth. Alibaba sits in the latter camp. Financing options are abundant, and Alibaba is issuing equity so that they remain so.

As with many big tech firms, there is little certainty about what the eventual tab will be for the AI buildout. That concern is as present with Alibaba as it is with the others. Estimates rise year on year. The management team’s original three-year infrastructure commitment was (¥)RMB380bn ($56.5bn), but it is already indicating expenditure could exceed this figure. In the last quarter, capital spending jumped 75% YoY to ¥68bn ($10.1bn), against ¥23bn of operating cash flow. Alibaba is a highly cash-generative company, but not sufficiently so to finance the AI buildout at the speed the management team currently wants.

Cash burn also appears to be partly elective. A foothold in AI development means cash must go out before revenues come in, and considerable amounts at that. Servers, chips, and data centres are financed today and rewarded tomorrow. Per Alibaba’s management, servers currently repay their cost within roughly three years, with a potential for that figure to fall towards two and a half, while carrying a depreciation life of five years. At current margins and payback periods, Alibaba could bring the AI-cloud and computing business back to positive cash flow if revenue growth were restrained below roughly 33%. Instead, the company is growing at 45%, allowing it to ease pressure off the accelerator, but choosing not to.

This lets Alibaba continue growing AI infrastructure at 40%-plus rates without forcing the balance sheet to absorb the entire gap between today’s capex and tomorrow’s cash generation.

Dilution at a Fair Price

The alternative funding method would have been tapping debt markets, as has been the predominant case with US Tech. This avoids dilution, but doesn’t avoid cost. 1-year dollar funding options would cost the company about 5.4%, while Dim Sum debt could be issued at a lower rate of 2.5%. However, this option was ruled out because of shallow market depth, an RMB-USD mismatch, and limited global investor reach.

Debt also introduces an obligation. Whether the investment pays off or not, the coupon payment remains, a concern we raised recently in our exploration of Nvidia’s new SPV and the risks to AI financing. Preserving debt capacity deserves more attention than a 3.5% EPS dilution. To us, the headlines and sentiment incorrectly emphasise the risks of this raise.

Alibaba’s balance sheet strength stands out compared with China and global peers. Equivalents of $70bn in cash and investments against $39bn of debt leave the company with considerable financial flexibility. Using equity maintains that balance sheet buffer, which we view as a much preferable approach than converting aggressive but discretionary AI investment into fixed claims on future cash flows.

Preserving this strength and borrowing capacity rather than stretching themselves in the opening stages of the race may prove fruitful. We have already mentioned that Alibaba has increased spending expectations, and these expectations will continue to rise as competitors do their best to keep moving forward. Tencent and Baidu face exactly the same incentives: as one participant raises the stakes, the others must decide whether to follow.

Of course, if Alibaba spends tens, maybe even hundreds, of billions building capacity that customers do not sufficiently value, issuing shares rather than debt makes the economic outcome no different. But if you take the side that AI investment pays off, then Alibaba has limited any risks relating to a financing problem as well. For this, we value the recent decisions and see this as an opportunity to increase exposure.

On the economics, evidence suggests the underlying business is already better than the headline. The company disclosed ¥5.6bn of adjusted EBITDA from cloud computing in the latest quarter, at a margin close to 12%, while revenue grew by 45% in this period. With the server economics repaying costs in a two-and-a-half- to three-year period, as stated previously, against a five-year useful life, this implies a project-level internal rate of return of around 20-30%. Cloud revenue has also accelerated for nine consecutive quarters and just recorded its fastest growth rate in 22 quarters. If these numbers hold, a 3.5% dilution looks fairly minimal.

In an AI race where the distance is yet to be disclosed, keeping some fuel in reserve seems prudent.

Bottom-Up Analysis

Underneath the hood, Alibaba is executing well, but by no means fantastic. To draw that conclusion, we first have to understand what Alibaba is all about. Most investors still know Alibaba mainly as an e-commerce giant. While that is still true for most of its earnings, it transcended that basic premise long ago.

The best way to look at the company is to divide it into three segments. First, there were: China e-commerce, International commerce, and Cloud Intelligence.

But as of this quarter, Alibaba has started reporting its business in a different way, one that makes more sense looking ahead and where they want to end up in a decade’s time.

The new segments are:

• Alibaba E-commerce Group

• AI Cloud and Compute Services

• AI Labs and Applications

Alibaba Health, Amap, Lingxi Games, and the media assets are categorised in the “all others” segment. ​

One thing that stands out is that this “reorganisation” doesn’t clearly improve investors’ view of the company. They actually separated the model labs and consumer AI apps into their own segment.

And by doing so, they exposed the losses clearly to everyone. Instead of hiding it, management showed transparency and brought the Qwen app spending into the open.

As you can see, the E-commerce section is still responsible for the vast majority of revenue. 71.7%, so almost three-quarters of revenue still comes from E-commerce. But that is not the growth story.

Alibaba E-commerce Group

We think the best way to view this segment is as a “cash cow,” funding growth in the other segments. Revenue grew only 4%, and adjusted EBITA fell 0.6%. These numbers are underwhelming, and they are not why one would invest in Alibaba right now.

​

Today, this is by far the most profitable segment, with group-wide adjusted EBITA of RMB27.3bn. E-commerce produced RMB39.7bn. Cloud added RMB5.6bn, but the other segments were unprofitable overall.

China E-commerce, which is responsible for over 50% of this segment, fell 8%. Management claimed this was due to weak transaction activity and a soft customer base, a headwind they’ve flagged multiple times in the past quarters. It’s clear that the China E-commerce engine is facing weaker transaction activity and a soft macro environment.

Meanwhile, the China Quick Commerce segment grew 45% YoY. In this segment, you can find Taobao Instant Commerce, Freshippo, and on-demand grocery delivery. Losses in this segment narrowed this quarter, which is actually quite impressive given how fast it scales right now.

International e-commerce was roughly flat YoY, down about 1%. This is mostly because of geopolitical headwinds and the tariffs. A big milestone to be celebrated here, with AliExpress finally being profitable after years of draining cash to fund growth.

Global wholesale, the smallest segment, rose 7% YoY. In this segment is Accio Work, an AI agent built specifically for cross-border merchants. Accio reached over 50k paying merchants shortly after launch. ​

Looking ahead, management believes quick commerce should become profitable in FY29, and in the long term, they think it will contribute 30% of platform GMV.

A simple conclusion can be drawn: the legacy marketplace is now a cash machine, under pressure from a weak Chinese consumer, but this is where funding for the other business lines starts.

AI Cloud and Compute Services

If there’s one thing to get excited about with Alibaba, it’s the AI cloud and Compute segment. External revenue growth accelerated for the last nine quarters, and in Q2, YoY growth was 45%. Alibaba Cloud margins improved to 11.6%. And if this cloud thing is new to you, growth and margin expansion are very unusual in this line of business.

The key driver here is AI product revenue. It’s unclear what this exactly entails, but Wu recently said it’s mostly AI compute, MaaS (Model-as-a-Service through Bailian / Model Studio), and AI applications and software. Qwen is NOT in here; more on that later.

AI product revenue saw triple-digit growth for the twelfth consecutive quarter. As of the latest quarter, that makes up about 35% of external cloud revenue, up from 30% last quarter. So it’s starting to continue significantly more and more to headline numbers

This also explains why margins were up, as these products carry much higher margins than the rest of the cloud portfolio.

Everyone knows compute is scarce right now, and consensus is that this will remain true until at least 2030. This pushes GPU pricing toward the most profitable monetisation model.

Alibaba claimed that it actually renewed its contracts at a better price than when they initially signed them. So for them, compute is no longer just a cost factor; it’s now more of a revenue-linked asset. CFO Toby Xu said that Alibaba runs servers for five years, and that AI servers cover their cost in three, so years four and five are free cash flow. Utilisation rates for their older GPUs also seem very high. An A100 bought in 2020, or a V100 bought in 2018, is still running at full capacity today. This basically tells us that the useful life runs well past the accounting depreciation schedule.

Another important addition to this segment is T-head, a chip designer. They have already shipped over 500k previous-generation chips.

Their latest new Zhenwu M890 went live on Alibaba Cloud as a SuperNode instance in August. It runs inference for models above two trillion parameters.

As you are probably aware, this is quite a big deal. Chips and storage are the most important parts in an AI data centre, and commercial GPU’s are very expensive right now. So every little percentage point they can squeeze out of their own chips flows straight to their gross margin.

Alibaba’s ambition here isn’t small; they want to do over $100B in external cloud revenue by 2030, preferably with gross margins of 20%. Management stated that they have “good visibility” on both. This means that growth should accelerate further in the next quarters, with EBITA margins expanding on top of it. But doing $100b in 2030 implies a 45% CAGR, which seems extremely aggressive to us.

Alibaba seems well positioned to capture a large share of the cloud market in China, and it’s safe to conclude that this part of the business is likely the most important growth lever to pull in the next few years.

Any meaningful slowdown here will certainly put the eyes back on the e-commerce section, and that’s not what will get investors anywhere close to excited.

Last but not least, the AI Labs and Applications section.

AI Labs and Applications

This is where things start to hurt, as this is the opposite of a cash cow. AI Labs and Applications show an adjusted EBITA loss of RMB13.9bn, against a RMB3.2bn loss a year ago. When you look at the headline numbers, this segment delivers much of the pain.

So what is causing the pain exactly? AI Model Labs, the Qwen app itself, and QwenWork, which is their new enterprise agent. Like other AI model builders, a large chunk of the losses come from training and inference costs. This also applies to the Qwen app. Luckily, the loss narrowed significantly this quarter because marketing spend came down. Management believes this downward trend should continue as marketing normalises and training efficiency improves.

An important distinction to make here is that the money that’s being spent in this segment shows up as revenue somewhere else. Think of the MaaS revenue, API calls, inference, etc. They are all booked into the cloud segment.

So why the focus on open-source models rather than closed models like OpenAI? Wu’s argument is that a growing, open ecosystem drives more demand for compute. And would you know, Alibaba sells compute. He also said gross margins on hosting third-party open models are roughly the same as hosting Qwen. More open models on Bailian is good for Alibaba either way.

Now we are full circle. This is where the connection to the E-commerce platform starts coming through. Over 250M users have used Qwen’s agent to do shopping.

While the share offering rightfully spooked investors, insiders opened their wallets. Alibaba Group co-founder Jack Ma has joined a pair of senior executives in picking up over $100M in shares in support of Alibaba’s AI ambitions. Jack Ma bought more than HK$600 million (S$97 million) worth of shares. This is on top of the US$20 million (S$25.4 million) worth of stock that senior execs Joe Tsai and Eddie Wu had together acquired in recent days.

And we all know the saying: “Insiders only buy for one reason.”

Our Take

Betting on Alibaba is not simply a bet on its e-commerce business. It’s the belief that the cloud and AI segments will drive growth over the next few years.

But to facilitate that growth, management has to continue spending a lot of cash. Capex jumped from RMB38.7bn to RMB67.7bn, taking cumulative spend to RMB190bn of the RMB380bn three-year plan.

While the ROI on these investments remains unclear for now, this seems to be the way forward, especially given where Alibaba’s core business stands today. This is where it will build its growth engine for the next decade.

Wu received the simple question in an earnings call: where does value end up in the AI stack?

He answered that right now, nobody knows; the short-term is concentrated in chips and infrastructure. But Alibaba built all four layers precisely so the answer doesn’t have to matter. They are spreading their bets, ready to make the move when the time is right.

The large increase in capex and the latest $10.2bn share sales are what it costs to place the bet.

The bet on the future.

AP, TH

Disclaimer: both AP Research and TacticzHazel have a position in Alibaba (BABA US).

This article was written in collaboration with TacticzHazel. TacticzHazel is an Independent equity investor and researcher. He focuses on high-quality companies with moats and strong balance sheets. If you enjoy sector deep dives, company research and other investment ideas, you will enjoy his work.

We had a great time writing this piece together as our styles and research approach overlap a lot. We’d highly recommend subscribing to his newsletter via the link below.

TacticzHazel's Substack

The move had been building since early July as structural concerns such as the deficit rising, inflation staying above 3%, and a wave of AI-related corporate issuance competing for the same duration buyers compounded. The pain point was hit, and Bessent showed his hand, which we’ll speak more on later.

Across equity sectors, traditional defensive areas like healthcare and materials benefited, while energy continued to get support from elevated oil prices. Healthcare was especially lifted after Moderna’s extraordinary rally following positive melanoma vaccine data.

Tech stocks bore the brunt of the weakness on the sector breakdown (as well as the most $ net sold last week, per Goldman). Walmart delivered its slowest US comparable-sales growth in more than six years and guided below expectations, sending the stock down more than 9%. That view is largely comparable with the macro data, as households increasingly make trade-offs. Target had a better quarter and raised its outlook, but the broader picture remains one of uneven consumer spending.

The dollar was used as Bessent’s sacrifice, with the dollar index (DXY) falling about 1%. Bitcoin had an unlikely winning week, benefitting from the buyback announcement, lower yields, renewed liquidity expectations, and optimism following Trump’s meeting with crypto industry leaders. Gold also continued higher, with the “debasement” thesis getting a fresh tailwind.

Instead of doing several short memos on different assets, as is our usual style on a Monday, we’re combining all our thoughts into one continuous analysis this week, all focused on the knock-on effects of Bessent’s buybacks. Yields, dollar, gold.

Let’s get into the guide to trades moving markets, where things stand, and where they may be heading.

• “Not QE, Not YCC”

• “Summary of House Views”

Not QE, Not YCC

This title felt apt to correct any misinformation out there. This was not quantitative easing or yield curve control. The Treasury is altering the maturity and liquidity composition of the debt held by the market. More Operation Twist than QE.

Bessent’s actions in the Treasury market are likely to drive most themes in the coming days, so this week we focus on this core area while exploring knock-on effects across multiple assets. We’ll start first and foremost with Treasuries themselves.

The initial reaction saw the US Treasury’s expanded long-end buybacks deliver a bull flattener on a one-day lookback, but the more consequential market signal is that Washington has now shown its hand. On a 10-day lookback, the 2s10s curve remains in a bear steepener.

Doubling long-end liquidity support is small relative to supply, but the 10- and 30-year still had sizable reactions. Coupled with the administration’s willingness to intervene elsewhere, the message from Wednesday’s move was difficult to miss, and the size of the intervention may be beside the point. Washington is increasingly uncomfortable with the signal coming from the long end, and markets now know that discomfort exists. Importantly, pain points in markets have a habit of being revisited.

It’s an effort to lean against price discovery and compress long-term borrowing costs. That can matter tactically, particularly when positioning is stretched, but it does little to alter the forces that pushed term premium higher in the first place.

It’s all very poetic when you consider the side of a currency trade that Bessent was on in 1992. He should be fully aware of the consequences of intervention signal failure.

Running growth hot, welcoming easier financial conditions, and resisting higher policy rates is difficult to reconcile with structurally lower long yields. A higher long end is an accurate reflection of Washington’s long-term neglect to run a balanced budget. National debt now tops $40tn. As our friend likes to say, “nothing stops this train.”

We still view steepener trades as attractive, and are in our 2s30s trade from a good entry point, but of course there are increasing risks here, that being th Treasury eliciting coordination from the Fed to provide sustained relief at the long end of the yield curve. That raises more concerns about Fed independance, but would offer enough of a showing to the market to maybe not test pain points… So far, the consequences of Bessent’s actions have been short-lived.

The other market that felt this shockwave is the dollar: Bessent’s sacrificial lamb in all of this. Rather than tackle the country’s structural fiscal imbalances, Bessent has clearly shown he’ll take the easier path and manage yields, shifting pressure from bonds to the currency.

The knee-jerk reaction from the news was to add dollar bearish positions, and the sterling and euro attracted the strongest demand. We’ll have to wait for Tuesday to see the positioning change in CFTC data, but prior to the announcement, positioning was still largely positive.

Bessent would welcome these FX movements, as the Trump administration has been praising the benefits of a weaker dollar as a way to increase US competitiveness and reduce trade imbalances.

As for under-the-hood measures, markets are paying for protection against, or exposure to, a higher EURUSD/weaker-dollar outcome. Implied volatility on the pair is still low, but 1m 25-delta risk reversals have moved meaningfully higher and are now positive (orange line below). That means risks have become increasingly one-sided toward USD weakness.

With higher risk reversals but still low implied volatility, we’re in a managed depreciation regime. If, however, implied volatility follows skew higher, that would indicate that the market is starting to view policymakers as having lost control of the move rather than a more orderly dollar sacrifice.

Gold is closely linked to the dollar, and much has been said about fresh tailwinds for the “gold higher” camp and the debasement traders.

It’s been a good couple of weeks for bullion. The precious metal was already on course to log its best month since February before shooting higher on Wednesday. The rebound from a year-to-date low in June has coincided with a pullback in US two-year yields from their recent peak as traders pared bets on Federal Reserve interest-rate hikes.

That inverse relationship will continue to favour gold. US economic data keeps disappointing, leaving the door open to further dovish shifts in policy expectations. Options traders are already hedging the risk that the Fed pivots to cutting rates in 2027.

Meanwhile, long-end US yields remain elevated even after Wednesday’s pullback, but their relationship with gold is more nuanced. The negative correlation strengthened during the Iran war but reversed in June, around the same time the US term premium began to rise. That suggests higher yields pose less of a headwind to gold when investors demand greater compensation for long-duration risks such as inflation and fiscal uncertainty.

As for our call butterfly spread, we remain holding this position, which is up about 7x, with a maximum payout of 1x at $4750. It’s been a great trade so far.

Elsewhere In the Week

The week ahead includes the July US PCE index and the Kansas City Fed’s annual economic symposium in Jackson Hole. Nvidia earnings will also be released, with investors focused on whether AI demand can sustain its valuation.

We spoke about these two topics in more detail last week. In short, Warsh has an opportunity to become better understood by markets, and markets will focus on Nvidia’s continued role as the Bank of AI, looking for fresh details on AI financing and circularity risks.

Top Trade Ideas and Views

Curves

• US 2s10s steepener

• UK 10s30s flattener (here)

• JP 2s10s steepener

Rates

• Fed and BOE on hold for the remainder of the year

• BOJ hike in September, and two hikes in 2027

• ECB October hike

FX

• Bearish USD

• EURGBP six-month put spread, 0.855-0.845

Equities

• SPX call condor (here)

Commodities

• XAU 3-month call butterfly spread (here)

That’s all to start this week. See you soon,

AP

The past few years have been bogged down by AI thematics, leaving little room for markets to explore opportunities elsewhere. That’s not to say they haven’t existed, but one had to search past the noise to find these opportunities.

Of course, markets never turn up on an opportunity for durable growth, and we see healthcare as an under-appreciated market theme. It offers a rare combination of durable growth, Technology-like profitability, attractive valuation and diversification away from concentrated AI and AI-adjacent themes.

For much of the past three years, capital has been singularly focused on the momentous shift in the technology landscape, the biggest change since the start of the millennium. Whether it be building the models, semiconductors and other parts of the supply chain, or data centre buildouts, markets have judged and rated themes based on their proximity to this overarching narrative. Healthcare, conspicuously distant from the centre of the enthusiasm, became a popular funding source for the market’s most crowded positions.

Performance has underscored this dynamic. If we mark the release of ChatGPT as the unofficial start of the tech trade, US healthcare stocks have underperformed the broader market in every calendar year. On five occasions, healthcare has experienced drawdowns of more than 10% even as the market moved higher. Sector weighting in the S&P 500 index has fallen from 16% to 9% (see Figure 2), moving from the second-largest constituent to the fourth-largest (see Figure 1).

A wholesale demotion.

There have been reasons, beyond a hot theme appearing, for healthcare to underperform. Drug pricing reform, reimbursement pressure, Medicaid cuts, and policy uncertainty (initially DOGE1, then tariffs, alongside ACA2 exchange policy) have all weighed on sector sentiment. Investors have also begun to question margin compression, viewing AI as a compressor rather than an improver.

That has led to a deeply negative 3-month correlation (see Figure 3) between Healthcare and Tech, and one of the largest positioning resets in the sector’s history.

Neglect eventually creates opportunities. Healthcare stocks trade at a discount to the broader market, near the 26th percentile of its historical range. Many of the concerns that weighed on sentiment have either moderated or have been reflected in valuations. Earnings growth (inflecting up from -1% in 2026 to +22% in 2027) is expected to be among the largest contributors, outside of technology, to S&P 500 growth next year. A potential gridlock outcome in the coming midterms could alleviate concerns around policy uncertainty, while the bar remains low for potential upside surprise from AI productivity improvements.

Following this period of underperformance, we believe the balance of risks has shifted in favour towards US Healthcare. Not all of its problems have disappeared, but the dynamics have shifted enough to expect the next phase of the recovery to be increasingly driven by stock selection.

With sector weights, active ownership, and relative performance still near cyclical lows, Healthcare appears better characterised as a recovery opportunity, particularly as investors seek to broaden exposure beyond AI-driven market leadership.

Subscribe now

Under Owned

The positioning backdrop is fundamental to our exploration of Healthcare as an overweight allocation. As the sector lived through a prolonged period of underperformance, Healthcare has become a main funding short against AI exposures (see Figure 4). Per S3 data, short interest remains near five-year highs and has steadily increased over that timeframe as markets crowded into AI names. This has also led to an increasingly negative correlation between Healthcare and Technology.

Even with recent relative outperformance, Healthcare remains one of the least crowded areas of the market, while Tech and AI remain at elevated levels by historical standards. Within Healthcare, Equipment & Services remains deeply under-owned, with net exposure in the 3rd percentile of the past 12 months and just the 1st percentile versus the full history since 2018.

We expect capital flows to be a tailwind for this trade. Healthcare ETF flows appear to be turning in favour of investors (see Figure 5), something that had previously been a significant headwind to sector performance. ETFs have received $6.4bn of net inflows over the past 12 months, representing 6.3% of ETF AUM. This is the first time that flows have been significantly positive since early 2023. The largest three US ETFs are shown below.

This combination of funding shorts, low crowding, and structural inflows points to a positive backdrop for Healthcare outperformance and an increasingly attractive allocation away from crowded themes.

Inexpensive

Alongside a constructive positioning backdrop lies an inexpensive valuation. We often argue that expensive valuations are no reason to under-own stocks, so the same should be applied to the opposite. Cheap valuations are no reason, on their own, to hold an asset. We view valuations as a breeze. A high valuation means you are unvesting with the wind in your face. Healthcare, at its current rate to history, is a gentle breeze to our backs. Always welcomed.

Like many sectors in the S&P 500, Healthcare is expensive relative to its own history, but remains fair relative to the broader market (see Figure 6).

These valuations not only attract public market investors, but may also create a more constructive M&A backdrop. Healthy balance sheets, accessible capital markets, ongoing innovation (not even including the potential for AI to enhance this innovation), and a need for large companies to replenish product pipelines should support acquisitions, particularly in Biotech.

Combined, fair valuations and potential catalytic drivers add to Healthcare’s high-quality growth, an area overlooked amid the AI trade.

Poised to Accelerate

Looking towards 2027, the sector is expected to deliver steady revenue growth, but earnings growth is poised to accelerate significantly as margins recover. Consensus estimates see revenue growth of 5%, while net income is expected to shift from -1% currently to +22% in the upcoming year.

Healthcare is expected to have the largest contribution to earnings growth in the S&P 500, second only to the Tech sector. Due to pricing power, the sector enjoys tech-like margins and remains much higher than defensive peers. Most Healthcare industries earn margins north of 20% given patent protection and scale, while Biotech (25.4%) and Pharma (26.2%) generate among the strongest margins of all S&P 500 industries.

As for inflation factors, Healthcare margins remain insulated from rising costs, especially components like agriculture and energy, due to the industry’s limited exposure, whereas other defensive peers, namely Staples, are some of the largest users of commodities.

Demand Drivers

Healthcare has three fundamental demand drivers that support growth in the sector: an ageing population, rising life expectancy, and continued expansion of the middle class.

On the latter point, within the US, Healthcare expenditures as a portion of consumer spending have risen from 5% in the 1950s to 21% as of today (see Figure 7). In contrast, food and beverage spending has decreased from 21% to 7% over the same period. This sustained shift in spending has driven exceptional growth in Healthcare revenue. Since the GFC, Healthcare has generated the second-strongest revenue growth of all sectors, with a 9.2% CAGR vs 5.3% for the S&P 500.

Of course, rising energy costs and inflation concerns are a risk to this measure. These concerns could weigh on discretionary spending and potentially crowd out healthcare spending in the short term, but dynamics could offset this. A shift in spending demographics, along with strong household savings and rising wages, should keep discretionary spending resilient despite macro concerns.

These structural demand advantages have led to consensus estimates of mid-single-digit revenue growth over the next two years. This reinforces Healthcare’s position as an attractive source of durable growth, especially one outside of AI risks.

Midterms

One overarching question in any thematic report in the coming months is connected to the US Midterms. What effect do the Midterms have on the market in general, and specifically towards Healthcare?

While Midterms are often the weakest-performing year in the four-year cycle, performance has been dispersed by sector. Healthcare (10.7%) and Energy (8.9%) have historically held up better, while Industrials (0.6%) and Financials (-0.6%) have tended to drag the market down in midterm years.

Election outcomes, of course, matter. Scenarios where one party lost control of the political trifecta (holding the presidency, and majorities in the House and Senate) saw material underperformance in the six months after midterms (10.4%) compared to when control was gained, or Congress remained divided (16.1%). Thus, the market has still rallied, but it was likely hampered in part by the market re-pricing the expected ability of the government to pass legislation. This, in particular, is key to our thematic.

While healthcare affordability and OBBBA-linked3 Medicaid changes are likely to be included in the political discourse heading into the midterms, we do not see a material change in Washington towards the health policy backdrop. Why? Given limited incentives to compromise between the Administration and Democratic-controlled chambers of Congress, we believe other economic and affordability concerns are more likely to be more central and more impactful in voter behaviour. Therefore, we see minimal policy risks for the Healthcare sector in the coming months.

Subscribe now

Sub Industries

US Pharma remains healthy (no pun intended there). 2026 has been a solid year with upside to estimates. For the upcoming calendar (18 months), Pharma has a rich pipeline and minimal policy risk after the late-2025 MFN4 agreements. The sub-sector also trades at a 10% discount to the broader market, with potential for multiple expansion on improving sentiment.

As for Biotech, performance in 2026 has been mixed. However, sentiment has turned positive, with tailwinds (M&A, clinical catalysts) balanced against headwinds (regulations, rates). As for our recent calls within Biotech, we have now closed out of AtaiBeckley (ATAI US) following its acquisition by Eli Lilly (LLY US) (see Figure 9), marking a 59% return in just over one month.

Life Science Tools has underperformed YTD, with the end-market lagging in any recovery and minimal generalist appetite against AI, semis, or memory. However, Agilent’s (A US) move higher (see Figure 10) in May following its earnings call highlighted a change within this sub-sector.

After a three-year losing streak to the broader market, MedTech is either broken or undervalued, depending on who you ask. Decelerating growth, negative revisions, and a thin calendar have weighed on it this year. The potential as an AI hedge has done little to draw investor interest so far, but that could change if it becomes part of a recovery trade; support would need to come via positive 2027 estimates and new catalysts.

Managed Care may have hit a turning point. We see most MCOs troughing on earnings in 2026 (or already in 2025), setting up upward revisions, with Medicare Advantage a likely margin-upside driver and early signs of positivity in tougher products. Distributor volumes support growth at or above targets. Across facilities, conditions remain stable despite ACA/Medicaid headwinds.

Finally, the area of Healthcare most prone to AI risk: Vertical SaaS and HealthTech. This remains the most cautious sub-sector while the market still finalises its decision and quantifies ROI. The dynamic is driving investors to differentiate between “AI that sells” (revenue expansion) and “AI that saves” (measurable efficiency, lower cost-to-serve).

Now to the part of the article that puts the research and the thesis into action: our Second Opinion Thematic Basket.

Second Opinion Basket, “The Cure for Crowding”

Read more

The AI trade also chugged along. CoreWeave rallied after quarterly results showing revenue more than doubling and backlogged demand growing. Intel upsized its equity offering to $20bn. It marks two weeks of Nasdaq outperformance relative to the S&P 500, its first notable stretch since the tides changed in late June. Energy sectors led as the Iran conflict remains unresolved, while consumer-facing areas lagged after the retail sales miss.

Rates followed the inflation data lower. September hike odds fell, while the opposite was true for the BOJ after reports from Japan.

That’s the short recap of the week. No need for much more to be said.

Let’s get into the guide to trades moving markets, where things stand, and where they may be heading.

• “Right Tail Risks”

• “Largely On Hold”

• “Summary of House Views”

Subscribe now

Macro Monitor
Chartbooks

Right Tail Risks

Despite the lingering question among investors of “what could go wrong?”, equities continue higher. As we look at the equity backdrop, the remainder of August looks constructive, and the risks may instead be on the right tail. We see few macro hurdles until the end of the month, while several supporting fundamentals have returned to the market, so we look towards a new index trade for the coming weeks.

Read more

News broke last night of Nvidia’s new initiative (in partnership with US investment giants Apollo, Blackstone, BlackRock, Brookfield, Goldman Sachs, and KKR) to source financing for artificial intelligence, mobilising $500bn of third-party capital for AI infrastructure.

To break it down to its simplest explanation, (1) create a special-purpose financial vehicle that can issue debt or bonds, (2) use the proceeds of this issuance to buy GPUs and compute infrastructure, (3) then lease these assets and capacity to customers.

So far, the AI buildout has required huge upfront capital to build assets whose economic value is derived from renting out compute. But some good old-fashioned financial engineering never stood in the way of closing a gap before, and is set to do the same this time around. Now, the owner of the GPUs no longer needs to be the financer of it.

CoreWeave has already done something similar. An $8.5bn non-recourse facility is secured against AI infrastructure and customer contracts, while Apollo has also financed a vehicle to acquire $5.4bn of Nvidia’s equipment and lease it to xAI, Musk’s AI house.

However, financial innovation can change the question that needs to be asked. Peter Drucker once said, “The most serious mistakes are not being made as a result of wrong answers. The true dangerous thing is asking the wrong question.”

There is no point questioning demand for Nvidia GPUs. Quarterly calls from the company answer that question without doubt. However, one question hangs over us: Would the demand be equivalent if the financing available to these buyers were any different? Attractive financing can bring demand forward, but it also blurs the lines as to whether that timeline change is demand-led or financing-led.

Per Stanford’s research:

“The cost of querying an AI model that scores the equivalent of GPT-3.5 (64.8) on MMLU (a popular benchmark for assessing language model performance) dropped from $20.00 per million tokens in November 2022 to just $0.07 by October 2024 (Gemini-1.5-Flash-8B), a more than 280-fold reduction in approximately 18 months.”

Applying this to our economics, a lender may currently be financing multi-year debt against assets whose economic usefulness can be repriced significantly over a relatively short period (18 months, as the example above illustrates). This is the point that takes us to the juice of this thought exercise.

The bear case for AI investment need not be that AI falls short of current expectations and brings no ROI, that revenues are fictitious, or that GPUs sit unused. History tells us that technological shifts produce real assets and real cash flows, while still generating financial excess along the way.

Take a look at how the chain plays out.

Cheap financing means more deployed compute. More deployments create more chip sales and demand backlogs. That validates higher valuations. And higher valuations mean more equity and credit available. Nothing in this circularity needs to be fraudulent to bring it down. Reflexivity can do enough on its own to cause a crash (the core of our stress test to end this article).

The circular nature of such transactions has raised concerns about concentrated risks in the sector. We raised this initially in our autopsy of the recent AI selloff, citing the three Cs as factors driving the AI trade, be it higher or lower. Capex (or we could call this C “circularity”) deserves a note of its own.

Subscribe now

C For “Circularity”

Circular financing is a business arrangement where a supplier or investor funds a customer, and that customer uses the money to buy products or services from that same supplier. The cash moves in a loop: money goes out as an investment and comes back as reported revenue.

Several high-profile examples illustrate the depth of these interconnections.

• Microsoft and OpenAI: Microsoft has invested $13.8bn in OpenAI and now holds 27% of its for-profit entity. In return, OpenAI has agreed to purchase up to $250bn of Azure cloud capacity through 2032.

• Amazon, Anthropic, Microsoft, and Nvidia: Amazon has now committed up to $8bn to Anthropic and remains its primary cloud and training partner via Amazon Web Services (AWS). More recently, Microsoft and Nvidia have agreed to invest up to $15bn ($5bn from Microsoft and $10bn from Nvidia), while Anthropic commits to purchase $30bn of Azure compute over the coming years, running on Nvidia systems.

• xAI and Nvidia: Elon Musk’s xAI is structuring a $20bn special-purpose vehicle that will buy Nvidia GPUs and lease them back to xAI. Nvidia is reportedly contributing $2bn of equity for the same vehicle, while also supplying the hardware, effectively financing a portion of its own future sales.

This new deal takes circularity one step further, although Jensen Huang might disagree.

Much of Nvidia’s ecosystem has been direct as described above. Relationships stay between Nvidia and their customers. Whereas this new arrangement allows Wall Street to take a seat between the two. Rather than Nvidia taking exposure directly to the operator, and, indirectly, to its own GPUs, this SPV owns a broader layer of AI infrastructure and can raise different layers of capital against it. A financing stack, so to speak.

Equity sits at the bottom, absorbing the losses first but gaining much of any upside. One layer up sits private credit or mezzanine debt, gaining a higher return for more risk. Senior debt sits at the top level, gets paid first, and therefore remains the most protected.

It’s all very similar to how Wall Street has performed for its existence. Take risky cash flow and divide the losses among investors willing to bear the most risk. Nothing is inherently troubling about this structure, as, for one, it’s common practice on the Street. Uncertainty and unknowns are the risk.

A larger pooled SPV can diversify both the counterparty and the assets, so the risk attached to each dollar of financing falls. It also means that the total number of financed dollars can rise. Yes, the SPV does what it was intended to do, but increased capital efficiency during an investment boom can extend the investment boom. The buildout has transitioned from Big Tech’s balance sheets (strong after years of high FCF) to debt, private credit, new SPVs, and a need for supplier support (circularity).

The bullish takeaway is that the opportunity has become too large for traditional financing channels. Maybe the bearish spin is that the capital required to maintain the current rate of change has become too large for traditional financing channels. The balance between the two is left to the eye of the (risk) beholder.

The development of this new financing is a natural evolution of the industry, but changes the nature of the investment cycle.

Collateral GPUs

One question to ask relates to the collateral. In the intro of this article, we mentioned the use of collateral such as railways and aircraft. There are reasons as to why these were attractive. Economic value persists for decades once they are built. The operator can fail while the asset is still viable, which is where the use of GPUs differs from traditional infrastructure.

There is no question about their value today, but it can be difficult to calculate their useful or economic life when that may differ significantly from their physical life. A GPU may still be functional even as the value of the computing power it creates falls sharply. Nvidia is effectively paid to exacerbate this problem. Blackwell chips are on offer while Jensen boasts of how substantial Rubin is in comparison. It doesn’t take a genius to figure out that Rubin, in time, will also be superseded. The relentless improvement is what equity investors want from the company, and why they have been rewarded handsomely so far. But it remains somewhat of a puzzle what a lender wants from that same asset that supports a multi-year loan.

Therefore, a creditor financing a GPU is underwriting much more than whether a chip remains operational. Important questions remain as to the future earnings power.

• What utilisation rate will hardware achieve in the years to come?

• How quickly will inference costs fall?

• Could workloads migrate towards Google TPUs, custom silicon, or future technology that does not currently exist?

• And what would today’s GPUs fetch in the secondary market if the borrower fails during the life of the loan?

For example, CoreWeave’s GPU-backed financing extends to 2032. The asset securing that borrowing needs to retain sufficient economic value through several generations of hardware development. Nvidia’s new SPV approach does improve this problem marginally, as the collateral is not made up entirely of GPUs. Other areas of the infrastructure stack (cooling, networking, power) will also be used. Much like the railway comparison, a power connection could be released to another operator as the useful life will remain well past the length of the loan.

There is also a limit on what good diversification does here. An SPV with exposure to ten different operators is safer than one with exposure to one, of course. One management mistake or one customer failure causes significant damage without that diversification. However, many plausible problems can arise that are not common to one entity but are instead sector-specific. These risks are prevalent and well documented regarding AI, and lie in the basket of “What Ifs?”

What if AI applications take longer to generate revenues than currently assumed? Utilisation and rental pricing would come under pressure across the sector. The diversification protection disappears as every borrower owns the same exposure, the value of scarce compute. Ten names do not provide protection if all ten face the same stress.

The same issue applies to the claim that GPUs are liquid collateral. Under normal conditions, they likely are. If one borrower needs to sell a fleet of accelerators while demand for capacity remains high, there is plenty of demand for that supply. Yet liquidity is always easiest to demonstrate when nobody needs it. If several heavily financed operators simultaneously discover that they own more compute than they can profitably rent, the secondary market would have to absorb that supply as expectations for the earning power of that hardware deteriorate.

The risk of GPUs as collateral is that the recoverable value is much more cyclical than the current price surge suggests.

The Stress Test

Let’s wrap this all up in a stress test, not an AI doomsday one, just a scenario with some adjusted assumptions.

Read more