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

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

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

At least, that was the narrative to start this week. Following an unusual presser from Warsh after the FOMC held rates, and an AI capitulation leading to Wall Street’s golden child of the AI trade, Leopold Aschenbrenner, on the receiving end of a forced liquidation (just for Citadel’s Ken Griffin to sweep up his holdings), people are asking, “Is the bottom in?”

It may be. Despite the short-term shift, though, we still prefer the question of “Where are the safe havens?” And that answer isn’t immediately obvious. Currency havens offer little respite. Treasury auctions for 5-year bonds earlier this week tailed with bidders AWOL, while the reception for the 2-year signalled no broad shortage of Treasury demand in policy-sensitive paper, where duration risk is limited. And even though equities (S&P 500) may only be 1.5% below highs set on June 2nd, equity exposure in portfolios feels anything but calm.

So, what next? August is now upon us, a seasonally low period of market liquidity. We’ve spoken already about recent Fed and ECB actions; we’ve touched on shifting portfolio exposure towards financials and our search for single-stock catalysts in global markets; and we’ve highlighted idiosyncratic risks in debt markets on our home turf.

It felt right to, instead, turn our attention specifically towards two questions:

• What factors are driving the AI trade other than a hawkish rates environment?

• Do the unusual events of this week signal short-term relief for crowded themes?

Subscribe now

Rates Are Only Part of the Story

The easiest explanation for momentum weakness is the same story that has kept markets busy since March. Higher oil raises inflation expectations, higher inflation expectations raise yields, higher yields reduce the present value of the distant cash flows on which long-duration technology valuations depend.

If this were simply a rates shock, one would expect growth stocks to fall in tandem with the broader market. Instead, the market has become increasingly bifurcated. The equal-weight S&P 500 has reached new highs while the main index has fallen and hundreds of individual stocks have advanced on days when the headline index appeared becalmed.

Rates may have lit the match. But what is now burning through the AI trade is a broader reassessment of spending and returns. We call this “The Three Cs.”

Capex. China. Credit.

A full autopsy of the AI selloff is incomplete without exploring each.

Capex - Unambiguously Bullish No More

For much of the past two years, rising capex was treated as evidence of rising conviction. A hyperscaler announcing an upward revision to its investment plans was seen as confirmation that demand remained strong. Yet, investors no longer treat spending as unambiguously bullish. A quick rundown of recent earnings/outlooks highlights this.

Alphabet (GOOGL US) was an early warning this quarter. Underlying operating results were strong, but another increase in expected capex was received as a liability rather than a sign of confidence. The market’s concern is that the amount of capital required to maintain leadership continued to grow faster than the evidence of monetisation.

Meta (META US) provided the clearest example of what the market doesn’t want to see. Advertising remained resilient, but another increase in capex arrived alongside collapsing FCF. Again, the concern here is that investment is accelerating faster than management can demonstrate where the eventual returns will come from.

Microsoft (MSFT US) offered the opposite case. Capex remains high, but accelerating Azure growth and a sharp increase in paid Copilot adoption provided tangible evidence that AI investment is already translating into revenue. The market will tolerate the spending because Microsoft showed both monetisation and flexibility, including an ability to slow infrastructure investment if demand begins to cool.

Amazon (AMZN US), however, showed that the market has not rejected higher AI spending altogether. It raised its capex outlook again, and FCF turned negative (the same story as peers), but the strongest AWS growth in more than four years gave investors a credible demand story against which to judge that investment.

Apple (AAPL US) remains the outlier. Rather than participating in the infrastructure race, the company continues to consume technology developed elsewhere, which leaves capex low and cash available for shareholder returns. That relative discipline has helped Apple remain less volatile than other names. Yet the latest selloff showed that Apple is still not immune to more conventional concerns around guidance and supply constraints, even if they are more fiscally disciplined than other tech names.

In short, if markets suspect that revenue growth can only be sustained by using an ever-larger share of FCF, capex will be the killer of share price. Tech still has to answer to markets whether the incremental revenue generated by the next dollar of investment is sufficient to justify it.

There is an uncomfortable asymmetry emerging. If hyperscalers maintain or raise spending, investors worry about FCF and returns. If they slow spending, investors worry about the earnings outlook for AI beneficiaries. More capex is no longer unambiguously bullish. And less capex would not necessarily be bullish either.

China - Changing the Calculation

The second pressure is competition, particularly from China, turning the debate away from investment payoffs and towards underlying assumptions of an AI buildout.

The release of Moonshot’s Kimi K3 model revived the questions posed by DeepSeek early last year: How much infrastructure is really required to produce a competitive model, and how durable are the economics of the incumbent AI stack if capable models can be developed or deployed at lower costs?

While much of the debate centres on national security risks to US leadership, a market-relevant question for us to answer concerns capital intensity. The bullish AI investment thesis has rested, in part, on the belief that increasingly capable models require exponentially more compute. If improvements in model architecture and open-source development allow users to achieve similar outcomes with less expensive hardware, that relationship between intelligence and capital expenditure narrows.

We can get our little AI friends to create a visualisation of this relationship.

Some argue that cheaper models do not necessarily mean less aggregate AI demand. Falling costs can broaden adoption and increase usage. But as far as the market is concerned with that assumption, there remains a timing problem. The infrastructure is financed today and built today, yet much of the demand expected to justify it remains forecast rather than realised. This creates a vulnerability whenever technological progress appears to reduce the amount of hardware required per unit of useful output.

A similar issue is appearing further down the supply chain. Progress, for example, in Chinese deep-ultraviolet lithography equipment (used for chipmaking), combined with the expansion of domestic memory producers such as CXMT, force the question of durability on the industry’s moats of today. This is another factor behind a broad semiconductor selloff. The most expensive names, the most crowded positions, the companies perceived to face the greatest risk of Chinese substitution have generally suffered the largest declines.

We also have to call into question the revenue assumptions. As users gain access to cheaper models and more efficient inference, AI adoption can be viewed as bullish, but bearish for the companies attempting to monetise each unit of compute. An index of large-language-model token expenditure has recently continued to weaken, consistent with customers migrating towards lower-cost options.

The optimist has plenty of history on their side. 1) Falling hardware costs brought computing to the household. 2) Collapsing storage prices created an almost limitless appetite for data. 3) Cheaper bandwidth enabled streaming, cloud software and the modern internet economy. In each case, lower unit costs created uses that had previously been uneconomic.

Yet the investment paths of history do not always deliver the intended results. Hard-drive capacity exploded while manufacturers consolidated and struggled to earn attractive returns. Telecom operators financed the networks over which software platforms captured much of the value. The current market assumption is that those funding the buildout reap the rewards, but history’s example is hardly dismissible right now. Cheaper intelligence could produce enormous growth in consumption without guaranteeing that the companies financing the infrastructure earn returns commensurate with the capital committed.

If the price of intelligence falls faster than usage rises, the companies funding the buildout may find themselves generating less revenue per unit of capacity at a time when depreciation and financing costs are accelerating. That shockwave would travel through the entire chain.

And on financing…

Credit - Equity Concerns Become Financing Concerns

Maybe it comes as no surprise that credit risks feature in this article.

AI investment has become an interconnected web of chip suppliers, cloud providers, start-ups, infrastructure companies, and financing vehicles. One participant invests in another, which commits to purchase capacity from a third party, which then uses those commitments to finance additional construction. “Circular financing...” Explaining such feels like the below.

Young industries frequently require creative financing, especially when demand is outpacing existing infrastructure. Circularity, however, can become dangerous when valuation depends partly on the continued spending of others. The widening of CDS spreads across parts of the AI complex, therefore, deserves more attention. Equity investors can debate whether vendor financing deserves a higher multiple or a lower one. Neither will stop credit markets from analysing the balance-sheet risk.

Big tech is widening, with clear inflexion points happening in July…

…but ORCL and CRWV need their own axis. ORCL CDS sits 90bps higher than its peers above.

Take-or-pay agreements may be described as cloud contracts, but economically they function much like credit. Providers commit billions upfront to data centres and GPUs against customers’ promises to purchase capacity over several years. The infrastructure becomes the collateral, and future revenue becomes the asset supporting today’s investment. That leads us to pose another new question:

Is AI demand growing quickly enough to justify the financial commitments already embedded in the system?

Financing can become strained even without an industry contracting. A data centre underwritten on expectations of 70% demand growth can disappoint if growth slows to 40%. Revenue may still be reaching records, the industry is certainly not contracting, but the assumptions supporting the financing change. That is the second-derivative problem.

If this sounds similar to the GFC, your connections would be correct. But this is a different market, but both rely on future growth validating commitments made in the present.

This point ties together the first two Cs we mentioned.

If more efficient models and cheaper Chinese alternatives reduce the amount of compute required per task or compress the economics of inference, it pressures the investment cycle. Ultimately, balance-sheets will determine who can absorb that adjustment.

This explains the spread that Coreweave and Oracle have over the other tech CDS. CoreWeave finances GPUs against contracted demand, and Oracle has become increasingly concentrated in AI infrastructure through debt financing and lower cash flow bases than the likes of MSFT, AMZN, and GOOGL.

So the first two Cs culminate to this. Credit will be the market’s first attempt to distinguish between the companies capable of funding the AI transition and surviving adjusting assumptions of monetisation.

A Clearing Event, Not an All-Clear

The above questions help us to navigate the AI trade in the medium- to long-term. What happened this week, however, can guide us in the short term.

Read more

The Federal Reserve’s credibility problem has rarely been about finding the right words. Central bankers have no shortage of those (although if Warsh has set a precedent, he’d like to have less in the future). The harder task is convincing markets that the words describe a reaction function rather than a pre-committed strategy.

Warsh entered his congressional testimony insisting that the Fed has “no tolerance” for persistently elevated inflation. Had the CPI report landed hot, that language would have been perceived as a threat to markets, with an appropriate hawkish extension of recent moves to follow. Instead, those words arrived alongside the weakest inflation print since 2020. Headline prices fell 0.4% month-on-month against expectations for a 0.1% decline, while core CPI was unchanged rather than rising the expected 0.2%. Annual core inflation slowed to 2.6%, undershooting the 2.8% consensus.

Warsh did not have to demonstrate his inflation credentials by delivering an immediate rate hike. The data has done some of the work for him. In that sense, the credibility question received a favourable answer. Warsh can continue to present the Fed as intolerant of inflation without forcing policy tighter into a cooling price backdrop. His rhetoric looks more like a framework that is producing the desired outcome.

The two-year Treasury yield fell about 10bps, July hike pricing fell back, and the expected timing of the next move was pushed back towards the autumn.

The most effective inflation hawk is sometimes the one who does not need to hike. If households and companies believe the central bank will respond when necessary, inflation expectations can moderate before the response itself arrives. The “Maradona theory of interest rates,” as Mervyn King once said. Tuesday’s report therefore strengthened Warsh’s hand, giving him permission to sound hawkish while remaining patient.

(For those looking for added context to this theory, Mervyn King coined it in 2005, comparing monetary policy with Diego Maradona running virtually straight through the England defence because defenders moved in anticipation of where they thought he would go. King’s point was that markets react to what they expect the central bank to do. Those expected future rate changes can influence spending and inflation even when the official policy rate moves relatively little. He described this as the monetary-policy framework “doing the work” through expectations. An interesting economic insight and incredibly well timed, given Wednesday’s game…)

Yet, one benign report cannot settle an inflation debate. Credibility is not earned by a singular win, but by how the Fed responds when the next uncomfortable number arrives. Oil is rising again, underlying measures of sticky inflation remain above target, and the economy is not obviously weak enough to rule out another inflation wave.

Warsh has passed the first test, but more difficult ones are likely still ahead. In the spirit of the ongoing World Cup… it’s Warsh 1 - 0 Inflation as the ref calls for the first hydration break.

The Disinflation Engine

The immediate temptation is to dismiss the report as an energy story… the market bear and the rates hawk might do just so. Motor fuel did much of the work in pulling headline CPI lower, reflecting the retreat in prices following the earlier easing in geopolitical tensions. That component is volatile and may reverse quickly now that oil has moved higher again.

Shelter inflation rose just 0.1% on the month. Core services excluding energy were flat. Price pressures moderated across a broad range of goods and services, with several major categories making a negative contribution. Services inflation posted its smallest contribution since the early months of the pandemic.

The report was unusually soft, which raises the more important question:

Why?

Part of the answer may be that the consumer is increasingly unwilling, or unable, to absorb further price increases. Real wage growth has been weak for several months, the labour market is losing momentum, and businesses are finding it harder to pass costs through without sacrificing volumes. Inflation is cooling partly because demand is becoming more price-sensitive.

The post-pandemic period brought excess savings, strong employment, and rising asset prices… all contributing to why households could look past repeated increases in the cost of living. That cushion is now thin. This is healthy from an inflation perspective, although less bullish for nominal growth. The economy may enter a phase in which slower demand performs some of the work that monetary tightening previously had to do.

Call that the less celebratory reading of Tuesday’s report.

Our broad view remains that inflation is not out of control. Shelter is moderating, goods disinflation has returned, and consumers are becoming more resistant to price increases. The risk of a renewed wage-price spiral has fallen meaningfully.

The path back to target will still be uneven. June benefited from a large energy reversal, while oil has since moved higher on renewed geopolitical tensions. If sustained, that will feed back into petrol, transportation and distribution costs, making the next few prints firmer. Of course, the higher oil rises, the more incentivised DJT will be to diminish tensions in the ME.

A rebound in headline inflation would not necessarily invalidate Tuesday’s message. The key distinction is between an energy-driven bounce and a broader reacceleration across shelter and labour-intensive services.

Our central case remains for headline inflation to rise from June’s unusually soft reading while core inflation trends lower more gradually. That leaves room for recurring inflation scares, but not another inflation crisis. The report changes the timing of the Fed’s next move more than the destination.

What does this mean for our cross-asset views? A softer inflation print changes the near-term policy path, but its implications extend beyond the next Fed meeting. From the shape of the Treasury curve to the outlook for gold and the durability of the equity rally, the more important question is where the market has already moved too far, and where the best asymmetries now sit.

Rates Waiting for Warsh

Read more

That is increasingly the story in our home base of the UK. For years, the equity market has been described in the language of polite dismissal: too old-economy, too low-growth, too politically messy, too easy to ignore. But neglect has a habit of creating value. British takeovers are up 250% year-on-year, with overseas buyers rediscovering that one of the most unloved corners of global equities still contains real assets, cash flows, brands, and strategic value that public markets have been reluctant to price properly.

EasyJet (EZJ LN) is the latest example from our Master Fund of that cheapness being tested in public. Castlelake, the US investment firm, made a series of approaches for the airline, initially at levels the board rejected as opportunistic and undervaluing the group. On Monday, however, easyJet agreed in principle to a sweetened 690p-per-share offer, a meaningful premium to the prevailing 603p share price.

The market had already started to close the gap. EasyJet shares have risen almost 100% over the past two months as investors have gradually come around to the value proposition. At the beginning of June, Citi argued that the airline’s fleet alone could be worth double the share price. When the aeroplanes are worth more than the equity market’s view of the company, maybe it’s not such a surprise when a financial buyer starts circling.

There is also a wider macro footnote here. The strength of inbound UK M&A may help explain one of the more surprising moves in FX this year: sterling’s resilience. The pound has had to digest geopolitics, domestic political noise, and shifting rate expectations, yet it has remained one of the better performers in G10. Deal flow will never dominate a market as deep as sterling, where daily turnover dwarfs even the largest transactions. But cross-border acquisitions do create real currency demand, hedging flows, and corporate activity.

With the bulk of the profit from our easyJet long position now locked in, the focus turns to sniffing around for other cheap UK purchases that could be ripe for a move higher.

Subscribe now

Why the UK Remains Attractive

Before moving into the more speculative part of the exercise (which UK-listed companies might be next?), it is worth stepping back and asking why the bid for UK assets has returned at all.

There are three reasons private equity firms and overseas strategics keep circling London: the valuation gap remains wide, the domestic buyer base has withered, and the market is full of cash-generative companies that public investors have deemed too unfashionable to own.

The UK is cheap for reasons that are both structural and self-reinforcing. The market has the wrong sector mix for the current cycle, too little technology, too much exposure to banks, energy, staples and healthcare, and a domestic investor base that has spent decades walking away. Although the discount has become embedded, it does not mean it is justified.

1. The Valuation Gap Remains Wide

Start with the obvious point. UK equities are still cheap. The FTSE 100 trades on roughly 16x earnings, compared with 27.7x for the S&P 500, and remains at a discount to most European and Asian markets as well. Some of that discount is deserved. The UK does not have the same technology weighting as the US, nor the same earnings momentum from AI. But the size and persistence of the discount has created a clear arbitrage for strategic buyers.

For public-market investors, a UK-listed company can remain cheap forever. For a buyer with a longer time horizon, a balance sheet and a willingness to act, that same company becomes something different: a decent global asset trading at a local-market discount.

That is the essential appeal. Buy real assets in London at UK multiples, then own them privately, restructure them, or fold them into a larger global group where the same cash flows may be valued more generously.

2. Domestic Ownership Has Collapsed

The second issue is ownership. The UK market no longer has a strong domestic buyer standing behind it. Pension fund allocation to UK equities has fallen from more than 50% around 25 years ago to roughly 4.4% today. LSEG has estimated that more than £1.9tn has exited UK equities since 2000 as pension funds and wealth managers shifted capital toward global markets, especially the US.

Valuation is not only about earnings. It is also about who is willing to take ownership of those earnings.

The answer in the UK has been “not enough people.” Persistent domestic selling has kept valuations depressed, liquidity thin in parts of the market, and boards more exposed when a credible bidder turns up offering a 30–40% premium.

3. London Is Cash-Generative But Unfashionable

Third is composition. The UK’s discount is not a mystery. The market looks the way it does because of what it owns. The FTSE 100 is roughly 28.1% financials, 15.1% consumer staples, 14.2% industrials, 12.2% healthcare, 9.8% energy, and only about 0.4% technology.

That is almost the opposite of what global investors have wanted to own over the last decade. The FTSE 100 is dominated by banks, miners, insurers, energy companies, consumer staples, and healthcare names rather than Big Tech, software platforms, or AI infrastructure winners. A market built this way can look optically cheap for a very long time.

But that is also why it keeps attracting buyers.

For public investors chasing growth and narrative, many UK-listed companies look dull. For private equity firms and overseas corporates, “dull” can be another word for “useful”. These businesses often have real assets, steady cash flows, defensible market positions, and underappreciated strategic value. They may not deserve US-style multiples, but they don’t need to. When the starting valuation is low enough, the upside can come from someone else deciding the asset is worth more than the London market is willing to pay.

Where to Look Next

Jupiter Fund Management (JUP LN)

Read more

Keir Starmer is out, which is to say Westminster has once again managed to produce a surprise everyone had spent weeks preparing for.

The exit was priced before it happened, and the market reaction since has reflected that. The Makerfield byelection gave Andy Burnham his route back to Westminster, and Donald Trump, never knowingly under-briefed on another country’s political instability, had already previewed the ending from across the Atlantic the weekend prior.

Andy Burnham looks set to replace Keir Starmer as Labour leader and prime minister, thus becoming the seventh occupant of the “top job” in just the last ten years. That said, while removal firms busy themselves once more in Downing Street, Burnham will likely face the same challenges his predecessors have faced and quickly find himself boxed in on all sides. While it is not confirmed yet, a lack of competition and polling websites suggest a near certainty that Burnham takes No. 10.

From a political perspective, there is clamour among Labour MPs to shift leftwards, to counter an insurgency in the polls from the Greens, and to provide an offering that is distinctly different from Nigel Farage’s reform. Such a shift, if it were to pan out, would likely tee up a “populist left” vs “populist right” contest at the next election, due to be held in 2029, and unlikely to be called sooner, considering the current parliamentary arithmetic.

Burnham is more interventionist in tone, but he has also tried to reassure markets. He has pledged to stay within existing fiscal rules and honour Labour’s promise not to raise taxes on working people, while arguing that privatisation and deregulation have created inefficiencies, especially in housing and utilities. The view is enough to make gilt investors nervous, but not yet to make them panic.

Gilts sold off the week after Burnham’s emphatic Makerfield victory, with the 10-year yield pushing higher and the long end once again becoming the most sensitive part of the UK political weather map. Sterling weakened too, which was the more revealing move. In normal circumstances, higher yields should support a currency. When yields rise and the currency falls, the market is not rewarding policy credibility. This has been much of the same reaction that we saw following the most recent ECB “policy mistake” (After You, Christine).

There is an echo of 2022, albeit on a far smaller scale. No Liz Truss moment has played out. No forced selling or pension-fund stress. However, markets are marking up the chance that the next prime minister might be more relaxed about borrowing, more willing to test the fiscal rules, and more tempted to treat the bond market as a political opponent rather than a funding source.

A lot was already in the price. Starmer’s authority had been fading for months. The UK was already carrying a political risk premium. The gilt market was already dealing with heavy supply, sticky inflation anxiety, weak growth, and a public-sector balance sheet with very little margin for error. Burnham did not create those constraints, but he inherits them for his reign (Gilt Trip, Starmer Drama).

In Hock, Whether He Likes It or Not

“We’ve got to get beyond this thing of being in hock to the bond markets.” — Andy Burnham, September 2025. For Andy, this may instead be the factor he has to live by. If Truss left us with anything, it’s that you have to have credibility with the bond market to survive.

That last line may be the constitution of modern British fiscal politics, a line not written down but enforced daily at the 8 a.m. open by gilt traders, liability-driven investors, overseas reserve managers, and anyone else absorbing a round of long-duration UK paper.

Burnham appears to understand this, and his tone has changed since that remark in September. In the wake of the market wobble last week, Burnham moved to contain any concerns by bringing in what we could call an A-list of fiscal personnel: former Bank of England chief economist Andy Haldane, the former Goldman Sachs economist and Conservative government minister Jim O’Neill, and the former head of the Office of Budget Responsibility Richard Hughes. A team of economic centrists, so to speak. Bringing in people with credibility around monetary policy, fiscal rules, and institutional economics sends a clear, positive signal.

Bringing in fiscal adults is in stark contrast to Starmer’s efforts (sorry, Rachel).

Markets can tolerate redistribution. They can tolerate higher taxes. They can even tolerate selective intervention if it is funded, explained, and tied to a plausible growth story… but talking about growth in a country that has become allergic to it does sound strange. What they cannot tolerate is unfunded ambition on ideological views alone.

That is where the Truss comparison is useful up to a point. Truss tried to tell markets that credibility could be asserted. Burnham’s challenge is to show that credibility can be borrowed, at least initially, from the institutions and people around him. Fiscal rules, OBR oversight, Bank of England independence and a serious chancellor are the guardrails.

There is also little to no desire among the general population for broad-based tax hikes, unsurprisingly, given that the tax burden is already on its way to a record high. And, make no mistake, the days of “just tax someone else” are now behind us. Wealth is too mobile, and the tax base far too top-heavy, for that to be a coherent strategy anymore.

Substantial additional borrowing is also off the cards, whether within or outside the fiscal rules, or whether dressed up as “borrowing to invest” or suchlike. The nub of the matter remains that higher borrowing is just that, no matter the clothing it wears, and that greater Gilt issuance would increase not only the fiscal risk premium that the UK must pay in the form of higher yields, but also the inflation risk premium that is paid.

While Burnham’s political capital will never be greater than the moment he enters No. 10 for the first time, plenty of fiscal stumbling blocks have already emerged.

The ‘Defence Investment Plan’ requires £4.7bln of funding that has not yet been found, while last year’s Budget was conditioned on departmental spending cuts and pre-election tax tweaks that always looked somewhat far-fetched. As a result, any honeymoon period that Burnham might’ve been expecting is likely to prove relatively short-lived.

Pre-Budget trial balloons will likely start flying almost as soon as Parliament returns from recess in September, beyond which Burnham’s choices will likely boil down to either growth-negative tax hikes, or politically unpopular spending cuts. With that in mind, any near-term relief experienced either by the GBP or Gilts is set to prove relatively short-lived, as after a summer of relative stability, may come yet another tumultuous autumn.

Where We Go From Here

The immediate market reaction also tells us what investors are really worried about. If we take a look at sovereign default swaps, fiscal irresponsibility is not showing up as it did in previous instances such as Covid and the Truss moment.

The concern is subtler and will show up more in yields, the curve, and currency markets.

Read more

1. Where to allocate capital on the basis of global rates either remaining elevated for longer or simply moving higher.

2. Where the better areas of the market are, from a valuation perspective, to rotate some profit from AI/tech names.

3. And defensive angles to portion some funds to help hedge or protect against an expensive market.

In our view, the common area that addresses all three points is financials. We believe bank equities have a credible case as a 12-month thematic. As with anything, it’s not simply as easy as saying “all banks benefit from higher rates,” or that “all banks trade on a low book multiple.” But there are plenty of options for finding likely winners among banks with the right balance-sheet structure (especially asset-sensitive franchises), sticky deposit bases, effective hedging, and manageable credit risk.

There are three drivers to consider: higher global rates, relative value, and defensives in an expensive market. We’ll start with the former ahead of the Fed this evening and the BOE tomorrow.

Subscribe now

1) Higher Global Rates

The current policy backdrop is directionally supportive for global bank stocks, but it is not the same story everywhere.

In the US, we pivoted our view earlier this month and now expect the Fed to hike rates this year. The move came after the strong employment report, with labour market conditions now tighter rather than deteriorating. At the same time, growth has remained resilient, and that’s before we start talking about inflation.

The economy is not overheating in a crude sense, but the policy stance is now too easy relative to the data. We think it plausible that the unemployment rate will fall gradually toward 4.0% by year-end, but see the headline CPI staying in the 3.5-4% range over the same period — a “higher-for-longer” backdrop for US banks.

In the euro area, the story is even cleaner for the theme. The ECB raised all three key rates by 25 basis points on 11 June, taking the deposit rate to 2.25%. The ECB’s own statement said euro-area headline inflation is expected to average 3.0% in 2026, and Eurostat’s May flash estimate put current euro-area inflation at 3.2%, up from 3.0% in April.

The tone of the presser was hawkish. Even though Lagarde did not pre-commit to further hikes, the updated inflation figures (and downgraded growth projections) clearly shifted the burden of proof. With inflation revised higher, core inflation above target through 2028, and adverse scenarios indicating a meaningful risk of second-round effects, the GC has effectively told markets that today’s hike won’t fix all of that and, therefore, more tightening is needed. We spoke about this factor specifically in our note last week, “After You, Christine.”

In the context of today’s equity focus, European banks have been among the main beneficiaries of post-2022 margin normalisation, and we expect this to continue into H2 2026.

The UK sits between those two cases. Headline May CPI out today held at 2.8% YoY, defying consensus expectations of a rise to 3.0% and coming in at the low end of the forecast range (2.8%–3.2%). The ONS attributed the softer reading in part to lower food costs, particularly meat and dairy products, which offset other upward pressures. Money markets trimmed BOE rate-hike bets, with swaps implying ~27bps of hikes for the year versus ~29bps the prior day.

The data reinforces expectations that the BOE will hold the benchmark rate steady at 3.75% at its meeting tomorrow, but the consensus view of one hike for this year still makes complete sense. The services inflation beat, and broader inflationary pressures already in the system, are likely to push CPI higher into year-end.

For UK banks, that makes the coming period less a pure hiking story than a higher-for-longer plus structural-hedge story.

Japan is the smallest absolute-rate market in this set, but arguably one of the most important in directional terms because it is coming off decades of ultra-loose policy. For Japanese banks, the thematic is therefore about normalisation from a very low base, which can have an outsized effect on margins and investor perception.

TLDR: The thematic is strongest in Europe and Japan, where policy direction is still moving towards tighter conditions, constructive but more selective in the UK, and more valuation-driven in the US, where the benefit comes less from fresh hikes already delivered and more from the market’s retreat from rate-cut expectations.

2) Relative Value

Cheap sectors are only interesting if they are cheap for no good enough reason. Other than that, they’re cheap for a reason.

Read more

Today, the committee, led by Christine Lagarde, raised its deposit facility rate by 25bps to 2.25%, its first rate hike since September 2023. Even though the move was widely anticipated (65 of 67 economists surveyed by Bloomberg expected this outcome), it was the updated staff projections on inflation and growth that caught our attention, as did Lagarde’s own comments at the press conference.

It’s clear to us that the ECB have swiftly pivoted to a “go big or go home” approach to tackling the inflationary impact, and we now think the deposit rate can rise multiple times this year. Lagarde clearly did not frame today’s 25bp hike as a one-off response or an insurance hike, as we thought they might, to retain optionality further out.

Yet we believe the multiple likely hikes this year will ultimately prove to be a mistake, given the likely hit to economic growth, and could see EUR come under pressure later this summer. This leads us to add a six-month EUR/USD option structure.

Inflation Projections Lifted…

The ECB revised its inflation forecasts significantly upward and trimmed growth expectations, as detailed below.

That makes it difficult for the ECB to sound dovish. In the baseline, inflation is not back at target until 2028, while core inflation remains above target through the entire projection period. The ECB also expects headline inflation to peak at 3.4% in Q3 and Q4 2026 and remain above 3% until early 2027, with the energy shock gradually feeding through to food, goods, and services.

The scenario analysis strengthens the point. In the milder scenario, oil normalises faster, inflation undershoots the 2% target in 2027 and 2028, and growth recovers more robustly. But that is essentially the “everything calms down quickly” case. The more policy-relevant risks sit in the adverse and severe scenarios. In the adverse scenario, HICP inflation runs at 3.3% in 2026, 3.0% in 2027 and 2.3% in 2028, with core inflation peaking at 2.7% in 2027. In the severe scenario, headline inflation reaches 4.0% in 2026 and 5.3% in 2027, while core inflation rises to 3.8% in 2027 as energy costs feed into domestic prices and wages.

So even though Lagarde did not pre-commit to further hikes, the updated inflation figures clearly shifted the burden of proof. With inflation revised higher, core inflation above target through 2028, and adverse scenarios indicating a meaningful risk of second-round effects, the GC has effectively told markets that today’s hike won’t fix all of that and, therefore, more tightening is needed.

It’s not even just our perception here, as the ECBspeak Index continues to climb.

The extent of any further tightening wasn’t specified, with Lagarde saying there would be no preset rate path. The neutral rate wasn’t discussed either, but consensus is that it’s around the 2.50-2.75% range, which would suggest at least two further hikes this year. Arguably, in order to materially provide an impact to dampen the inflationary pressure, we’d say it needs to be at least at the neutral range, if not higher.

…But Growth Downgraded

The problem as we see it lies in the health of the economy, which will only be put under further pressure from monetary policy tightening.

The ECB now sees eurozone GDP growth at just 0.8% this year and 1.2% next year, both lower than previously expected. That is not recessionary, but it is hardly an economy with much spare capacity to absorb additional tightening. The central bank is raising rates into an economy already being squeezed by higher energy costs, weaker consumer confidence, slower household spending and pressure on exporters.

If the inflation problem was being driven by excessive domestic demand, we’d be less concerned. But it’s not. It is being driven by an external energy shock.

There’s also concern that although higher interest rates can cool demand and lean against second-round effects, they cannot produce more oil, reopen supply routes or remove the geopolitical risk premium from energy markets. What’s to say that the base rate needs to go higher still in 2027 if we don’t get a conflict resolution in H2?

The central bank statement concluded it could no longer wait out the shock, prioritising price stability over economic momentum. That’s all well on paper, but markets aren’t stupid when assessing longer-term implications.

Turning to the Trade

Read more

On paper, issuance is still a long way from the madness of 2021. US IPO volume has edged higher in recent years, but the five-year average remains near its lowest level in four and a half decades.

A few aftermarket stats: mean first-day returns (equal-weighted) were 11.9% in 2023, 15.3% in 2024, and 29.3% in 2025.

More revealing still is the quality of the appetite. 53% of 2025 IPOs had negative trailing earnings, and the loss-making cohort enjoyed mean first-day returns of 39.8%, against 17.2% for issuers with positive EPS. That trend tracks back in five of the six years this decade. Markets react as much to narratives as they do to fundamentals.

Ahead of the big release, we wanted to run through some comments on Musk’s history of selling narratives, some equity market plumbing on float size and Nasdaq inclusion/S&P exclusion, and the read-throughs for the upcoming AI IPOs.

Subscribe now

Not Just a Space Company

In 2018, SpaceX President Gwynne Shotwell said that the company would not go public until it was flying regularly to Mars. Musk himself had echoed a similar thought over the years. For a space company, that makes sense. Having to stand up and explain your quarter to shareholders has its drawbacks. And for a research-heavy, capex-heavy company, focusing on engineering progress matters far more than quarter-to-quarter bottom lines. But that old logic has collided with a new reality. The dream has got bigger, as has the bill.

Read more

Implied yields on G7 government debt with maturities beyond 10 years had climbed above 4.6% for the first time since 2004. The US 30-year Treasury is around 5.17%, the UK 30-year gilt touched 5.82%, and the Bank of Japan is being forced to consider whether market instability might require a slower pace of balance-sheet run-off.

Each country may have individual reasons adding to its bond selloff, but the selloff itself is global.

There is one main story driving this higher, a noisy one of war, oil, and inflation expectations. Markets now have the Fed on hold through 2026, and have shifted far enough to price a possible hike by January. Beneath that headline story, there are a few dynamics we think deserve more attention.

• The first is that hyperscalers are funding an AI buildout of extraordinary scale, and have increasingly leaned on debt markets to do it.

• The second is that a continued shift higher in rates changes the equity leadership map, making financials and banks a more natural home for risk than the most duration-sensitive parts of the market.

• The third is that the UK has its own idiosyncratic fiscal and political vulnerabilities, which give long-end gilts a higher beta to the global duration selloff than US Treasuries or most other G10 markets.

Not all rising yields are equal. A move driven by collapsing credit quality is one thing. A move driven by stronger nominal growth, higher real rates, and a fatter term premium is another.

Read more

The additional domestic risk premium continues to drive UK yields higher. On Tuesday, 30-year gilt yields touched 5.81%, the highest level since 1998, while the 10-year reached 5.13%, the highest since 2008. Although we often roll our eyes at “the highest since” comparisons, we believe the above references are worthy of inclusion.

With Britain’s borrowing costs now the highest among rich economies, and having risen the most since the Iran war began, it’s being priced most aggressively at the long end, where fiscal risk conventionally lives. If the two-year trades around the policy actions from the next few BoE meetings, the thirty-year is all about whether the policy regime still looks coherent five to ten years from now.

Simply put, the current policy regime isn’t sustainable, and we believe we could be close to a serious yield melt-up that could cause significant headaches for the Government. But for macro traders, there’s potential for several attractive rate trades.

Subscribe now

Election Results Fuse Politics With Duration

The short-term driver behind the latest leg higher in the UK long end came from the local election results late last week. This matters because they converted a bad macro situation into an open political one.

By last Friday evening, the governing Labour Party had lost 1,229 council seats, while the right-wing Reform UK had gained 1,372. It was one of Labour’s worst-ever defeats in local elections, and by Tuesday, more than 80 Labour lawmakers had publicly called on PM Keir Starmer to go or set out a timetable for departure.

For gilts, the question is what the government has to promise, spend, delay or abandon in order to make survival politically viable. Investors can see two paths leading to the same fiscal destination: the current leadership loosens policy to claw back voters, or a successor arrives with a looser bias already built in.

Option 1: Starmer Remains But Loosens Purse Strings

Read more

We think the answer is more nuanced than much of the commentary suggests. The move higher in US long-end yields is not best understood as a purely domestic Treasury story. It is part of a broader global repricing of long-duration risk, with the US dragged into a move that has been far more aggressive elsewhere.

In the US, 5% 30-year yields still attracts buyers. In the UK, 30-year gilt yields at the highest level since 1998 are forcing investors to reprice political risk and diminishing fiscal credibility.

Rather than treating higher US yields as the cleanest expression of bearish duration, we think the better setup is relative, and outline a new spread trade below on the premise that US long bonds may remain volatile, but UK gilts look more vulnerable.

The Global Tail Wagging the US Dog

Let’s start our case by looking at the front-end policy logic, where recent meetings from the Bank of England (BoE) and the European Central Bank (ECB), presented a much easier stance to understand to see than for the Fed. The BoE kept the bank rate at 3.75% at the end of April, but its scenarios explicitly contemplated a “forceful tightening” if the energy shock produced material second-round inflation, and markets still price roughly two hikes this year.

The ECB also stood pat, but openly debated a hike at its April meeting and has since signalled that the chance of a move has risen, with markets now expecting the first increase by July and more coming by the end of the year. If the UK and the euro area are repricing for actual tightening, Treasuries were never going to ignore the same global energy shock altogether.

Read more

We are sharing the latest BoJ note here for AP Research readers.

The Bank of Japan (BoJ) meeting earlier today was the first major G10 central bank meeting, with many looking for clues on how the respective bankers will position policy amid reflationary pressures.

Not for the first time, the BoJ made a bit of a hash of its meeting, issuing a mismatch of comments that saw the Yen initially appreciate before giving back all gains and then some.

When we dig into the details of the meeting and subsequent press conference, the desire to continue holding our short Yen position (expressed via a long BRL/JPY) remains firmly intact.

SUMMARY

• The significant revision higher in inflation forecasts should have merited a precautionary rate hike from the BoJ today.

• The 6-3 split, and unclear comments from Ueda, only add to the worry that the central bank doesn’t have a clear vision for policy going forward.

• Clear risks of getting behind the curve or hiking aggressively and quashing domestic demand mean we continue to play the Yen from the short side.

Inflationary Pressure

Read more

Alongside that, we outline a tactical UK rates expression. The recent run of inflation and PMI data has nudged the market toward a hawkish near-term narrative ahead of next week’s Bank of England meeting, and we do not think it is priced correctly.

Before turning to those trades, however, it is worth pausing on a potential macro constraint that sits just ahead. A statutory deadline tied to President Trump’s handling of the Iran conflict could bring Congress more directly into the decision-making process, shaping the next phase of the conflict in a way markets have not yet fully engaged with.

Subscribe now

Trump’s D-Day

We could be approaching an inflexion point in US policy toward the Iran conflict, not due to commodity supply chain issues or the Hormuz Blockades-of-all-Blockades, but rather to statutory constraints imposed by the 1973 War Powers Resolution.

Under this framework, the president can conduct military operations without congressional authorisation for up to 60 days following formal notification to Congress. Given that Trump notified Congress on March 1st, the current case sets a key deadline of May 1st.

To date, congressional Republicans have largely supported the administration, blocking Democratic efforts to curtail the campaign. However, the upcoming deadline is beginning to expose fractures within the GOP, which we believe markets haven’t fully appreciated. Several lawmakers have signalled that their support may not extend beyond the 60-day window without explicit congressional approval, raising the probability of a policy pivot or heightened political tension as the deadline approaches.

What are Trump’s options? Beyond the 60-day mark, the president faces three main pivots.

1. Seek formal authorisation from Congress

2. Begin a withdrawal of US forces

3. Invoke a limited 30-day extension to facilitate a safe exit

Notably, this extension does not legally permit continued offensive operations, constraining the administration’s flexibility.

Importantly, precedent suggests the administration could challenge or sidestep these constraints, which wouldn’t surprise us with Trump. Yet, in terms of the precedent we speak about, it was actually President Obama who continued the 2011 engagement in Libya beyond the 60-day mark and argued that US operations didn’t involve sustained fighting or ground troops. This did incur bipartisan backlash, so any similar move by Trump risks increasing political pressure, particularly as some Republicans have already framed the 60-day mark as a meaningful legal and institutional boundary.

We could term the May 1st deadline as “a discrete political catalyst.” What we mean by this is that Trump is going to struggle to keep this conflict going on his own terms. Rather, as of next week, Congress is likely going to have a much larger say in deciding the path of the war. In some ways, it should help resolve the question of where the conflict goes from here, given that the range of outcomes is relatively constrained; the conflict is either extended, curtailed, or escalated through congressional involvement.

Notable Mini-Thematics

We’ve added three new standalone baskets on Plutus this week, focusing on topical areas where we believe investors can generate alpha in the coming months.

AI vs AI-At-Risk

We believe this group of 20 names is well-positioned for the current regime, as the market increasingly distinguishes between beneficiaries of AI capex and those exposed to its second-order disruption.

Read more

A ceasefire in Iran has come closest to fruition since the escalations began in late February. However, in keeping with the current administration’s preference for maximalist leverage, this isn’t so much a white flag as it is a tactical timeout defined by an aggressive “15-point plan.”

From “Obliteration” to “Optimisation”

For the markets, the shift is strong. We were staring down the barrel of a Tuesday evening deadline, the point at which the Strait of Hormuz was either to be reopened or the Iranian military was to face what President Trump termed an “obliteration” attack.

Instead of the fireworks, we got a two-week, “double-sided ceasefire.” The geopolitical risk premium in Brent is currently being re-indexed. Equities, particularly those sensitive to energy inputs, can finally take a breather. We are moving from a fear of escalation regime to a sceptical stabilisation regime.

The architecture of this truce, brokered via Pakistani mediation, is precarious. It hinges on Iran agreeing to the “complete, immediate, and safe opening” of the Strait of Hormuz.

While Tehran has signalled acceptance, the devil is in the jurisdictional details. Iran has stipulated that “safe passage” would occur in coordination with their own military. In market-speak, this is counterparty risk. If the Iranian Navy is the one coordinating the passage, the definition of “safe” remains entirely at their discretion.

The 15-Point Reality

The diplomatic math here is a hybrid of two competing visions:

1. The Iranian 10-Point Proposal, which Trump surprisingly called a “workable basis” for negotiations.

2. The US/Israel 5-Point Overlay, the “hard” conditions regarding regional proxies and nuclear enrichment.

Combined, this 15-point plan represents a messy Venn diagram. Some points overlap (opening the Strait), while others (disarming Hezbollah) remain total non-starters for Tehran.

Read more

This provided another Jenga-style piece removal of security with which the US economy is currently being propped up. The ongoing conflict in Iran is causing some to increase their probability of a recession. Although we wouldn’t position ourselves in the eternal optimists bucket, we believe there are several reasons why it’s still far too premature to start throwing around the R-word.

Things Are Different From 2022

Ok, technically, the last official US recession was in 2020. However, US GDP in the first two quarters of 2022 almost triggered a technical recession and provides us with the most recent blueprint to compare to. It’s also a relevant comparison due to surface similarities around an energy shock and geopolitical uncertainty from the start of the Russia-Ukraine conflict.

However, on reflection, the current macro backdrop differs materially from the 2022 recession template. Most importantly, the inflation dynamic is far less entrenched today. In 2022, economies experienced a shock as accelerating wage growth and tight labour markets emerged from the Covid pandemic, creating the conditions for a wage–price spiral. By contrast, wage growth is now moderating and labour market sentiment has softened significantly, reducing the risk of persistent inflation and pointing instead toward a more disinflationary or even deflationary impulse.

Policy settings also start from a fundamentally different place. Central banks entered 2022 well behind the curve, with rates deeply below neutral and forced into aggressive catch-up tightening (Powell might never live down his “transitory” comment from that time). Today, policy rates are already closer to neutral, with the Fed well into its easing cycle.

Finally, the labour and technology backdrop introduces a new asymmetry. Whereas 2022 was defined by labour scarcity, today’s environment is increasingly shaped by concerns around AI-driven labour displacement, contributing to weaker employment expectations and dampening wage pressures. Taken together, the current cycle lacks the key ingredients that drove the stagflationary shock in 2022.

Granted, this doesn’t excuse the worries about US vulnerability to growth downside, but it’s unlikely to result from a 2022 repeat of persistent inflation.

Current Indicators

There’s no exact science in determining which metric is the best when trying to predict a recession. Yet even if we take a spread of some of the most widely used or vouched-for, the probability still appears to lack any significance:

Bloomberg median forecasted probability of recession in 2026 (%):

Read more

The recent drawdown in gold prices has been, in a word, violent. After a multi-year rally that seemed to defy the gravity of high real rates, the yellow metal has hit a wall of selling pressure. But to paraphrase a market adage: do not mistake a liquidity event for a change in regime. Our assessment is that the fundamental drivers of the gold bull market remain not only intact but are actually being strengthened by the very volatility that is currently depressing prices.

Subscribe now

Sovereign Sellers

The primary question on every desk right now is why gold is falling exactly when the geopolitical risk, specifically the escalation of the Iran war, is at its zenith. That answer lies in the plumbing of global finance, not in a sudden loss of faith in gold’s value. Gold is often described as the “ultimate haven,” but in the initial stages of this crisis, it functions as the “ultimate liquidator.”

The recent selloff has been predominantly driven by the urgent liquidity needs of major sovereign actors. For several years, central banks, particularly those in emerging markets, have been the dominant source of inflows into the gold market. However, a war in the Middle East and the resulting closure of the Strait of Hormuz create a specific kind of financial stress. When oil prices spike and regional currencies come under pressure, these same sovereigns often have to de-gross their portfolios to raise immediate cash.

In the early stages of a major macro correction or a geopolitical shock, investors and sovereigns alike sell what they can, not necessarily what they want to. Because gold is a highly liquid, deep market with no counterparty risk, it is the first ATM that central banks visit when they need to defend a currency or cover margin calls on falling equity positions. These liquidity-driven moves often disregard long-term fundamentals in favour of immediate solvency. When the Strait closes, the need for dollars becomes the only game in town.

Read more

Markets tend to assume that geopolitical crises escalate until something breaks. History suggests the opposite is often true. The moment when investors begin to imagine the worst is often when the incentives to avoid it become overwhelming.

The confrontation around Iran has quickly become one of those moments. Oil has surged, tanker traffic has slowed almost to a stop, and risk assets have been forced to price a scenario that earlier this month sat firmly in the tail of the probability distribution.

Yet crises are rarely governed purely by ideology or military logic, and are more often constrained by economics, domestic politics, and the basic question of who, if anyone, actually benefits from prolonging the conflict.

Viewed through that lens, the current standoff begins to look less like the start of a long war and more like a high-stakes game of incentives (something we should be used to in Trump-office markets). And once those incentives are mapped out, several conclusions follow, including one that markets may not yet fully appreciate.

Hormuz has to reopen.

Problems in the Strait

At a fundamental level, a prolonged conflict is not in the economic or political interest of most actors involved, including Iran itself.

The Strait of Hormuz, through which roughly 20% of global oil and LNG flows transit, is a critical artery not only for Gulf exporters but also for Iranian crude shipments. Sustained disruption would therefore undermine Iran’s own ability to export oil (much of which is sold to China) and deprive the regime of one of its most important sources of foreign currency. Even amid the current crisis, tankers carrying Iranian crude have continued to transit the strait toward Asia, highlighting Tehran’s reliance on the route for revenue.

A similar logic applies to the broader region. Gulf producers, Asian importers, and Western governments all have strong incentives to prevent a durable closure of Hormuz, given the scale of global supply at risk. Tanker activity has slowed to a snail’s pace, with severe implications for global growth and energy prices if it continues for much longer.

Trump Has an Eye on Midterms

From a domestic political perspective, there are also strong incentives for the White House to avoid a prolonged conflict. President Trump has historically framed his foreign policy around decisive outcomes and rapid demonstrations of strength rather than open-ended military engagements.

A drawn-out confrontation with Iran risks undermining that narrative, particularly if it results in sustained instability in energy markets, higher fuel prices for US consumers, or the perception that Washington has been pulled into another costly Middle Eastern conflict. With the US midterm elections approaching, the administration will likely be acutely aware that geopolitical crises can quickly become domestic political liabilities.

Energy prices are particularly sensitive in this context. A persistent spike in crude would feed directly into gasoline prices, a metric closely watched by voters and historically correlated with shifts in political sentiment. For any administration heading into a midterm cycle, a sharp increase in pump prices can act as a powerful amplifier of broader concerns about inflation and the cost of living.

The optics are therefore important: a conflict that drives oil materially higher risks turning a geopolitical event abroad into an economic grievance at home.

In that sense, the political calendar matters. Midterm elections often serve as referendums on the incumbent administration’s economic management and overall competence. Allowing a geopolitical crisis to evolve into a persistent economic shock, particularly through higher energy prices or financial market volatility, would risk complicating the administration’s domestic narrative at a sensitive moment.

Peak Fear Priced

There are increasing indications that markets may already have priced in peak geopolitical fear around the Iran conflict, which could limit the scope for further large downside moves in risk assets.

Recent price action across asset classes suggests that the most aggressive phase of the risk-off adjustment may already be over (see Brent, G10 STIR, etc.).

Positioning indicators flagged levels of stress comparable to major crisis periods, with volatile markets briefly showing dynamics similar to those seen during the early stages of the pandemic. At the same time, cross-asset repricing was swift and broad, with equities, rates, and commodity markets all rapidly incorporating a higher geopolitical risk premium. This type of sharp, front-loaded adjustment is often characteristic of the point where markets begin to price worst-case scenarios rather than base cases.

From a market-structure perspective, the key point is that large directional moves become harder to sustain once investors have already repositioned defensively. In the latest selloff, the biggest drawdowns occurred in markets that had been the strongest performers earlier in the year, suggesting that investors rapidly reduced risk in crowded positions.

Historically, such episodes of intense repositioning often mark the moment when fear peaks rather than when the decline begins, as the marginal seller becomes exhausted.

Importantly, the focus for markets tends to shift from the absolute level of geopolitical risk to the change in expectations. Even if tensions remain elevated, what matters for pricing is whether conditions are getting worse or simply becoming “less bad.” Once the trajectory of events stabilises, markets often begin to recover even while the underlying geopolitical situation remains unresolved.

The likely outcome is that volatility gradually subsides and the geopolitical risk premium embedded in assets begins to compress as investors reassess probabilities rather than possibilities.

Positioning for De-Escalation

Oil Lower

• Buy a 2x1 put spread on COM6 (spot ref 102.50)

Read more

Yet what we have seen this week (and as a partial follow-up to Software-as-a-Selloff) is the rise of what we’re calling the AI disruption vigilantes. This new breed of hungry equity traders has been out in full force this week, targeting various areas of the market where a high-fee, labour-intensive operating model could prove ripe for disruption.

Three key areas have been targeted this week, but the consensus trade for each one is different, as AI nuances are already becoming apparent.

Subscribe now

Financial Services (Tuesday)

The market reaction to Altruist’s latest release underscores growing investor concern that generative AI is encroaching on bespoke advice in the sector, one of the most defensible layers of the wealth management value chain.

The firm announced that its Hazel AI platform can now construct client-specific tax strategies by ingesting 1040s, payroll data, and account-level information and produce actionable recommendations within minutes. What has historically been a manual, high-labour component of advisory work is being reframed as a data-processing problem, and markets immediately priced in the implications for operating leverage across the sector.

Two companies we noted that took a hit (which has been exacerbated as the week has gone on) are LPL Financial and Charles Schwab.

In our view, the hit to both reflects fears that AI-enabled workflow compression could erode both pricing power and the perceived scarcity value of human advice.

If a single advisor can service materially more households (Altruist CEO Jason Wenk suggested as much when he said Hazel “expands what a single advisor can handle”), then revenue per advisor may rise, but industry-wide fee pools risk dilution as capacity expands faster than demand.

Tax optimisation has long been one of the most defensible justifications for advisory fees because of its complexity, regulatory sensitivity, and the need for cross-account coordination. By automating this function, AI threatens to commoditise what was previously a high-margin, expertise-heavy offering.

To clarify, we believe the concern is less about outright disintermediation and more about margin normalisation. Simply put, advisory firms may retain clients, but with lower incremental economics per relationship.

As far as the vigilante pursuit goes, we don’t see this area of the market posing serious risks relating to a complete decimation of related stocks. But we do believe the move lower is justified based on repricing long-term earnings expectations.

Real-Estate Services (Wednesday)

The vigilantes turned to real estate services the following day, using the same argument as they had used against finance firms.

The narrative that AI-driven workflow automation is migrating from back-office experimentation into core transaction functions caused significant hits to the likes of CBRE and JLL.

Read more