Recent feedback from investment managers suggests that many clients are currently somewhat bemused by the continued rise in the value of balanced portfolios. Such is the nature of financial markets – often confounding expectations and dancing to a different tune to the one playing in media headlines. Admittedly, this sort of disconnect tends to happen at a turning point at the bottom of an economic cycle, as investors anticipate better times ahead. What makes this period different is that May saw global equity indices achieve new all-time highs. 

The sustainability and extension of such gains will depend on a number of factors. One is the outcome of the ongoing conflict in the Middle East. Another less-reported factor is the sheer strength of corporate earnings growth, especially in the US and emerging markets (EM). Much of this momentum is being driven by the incredible growth of capital expenditure (capex) related to artificial intelligence (AI). Shades of the dot.com boom. We discuss this later – see ‘More AI’ section. Domestic politics, in both the UK and the US, also look set to play a role. And June will bring the biggest initial public offering in history (in terms of funds raised), an event that is set to test demand. 
 

Iran: discounting de-escalation

When the first strikes on Iran were made at the end of February, not many, if any, pundits envisaged markets being at their current levels by now – with the US’s S&P 500 stock index reaching new record highs. That’s especially the case as the Strait of Hormuz – safe passage through which is critical to the global economy – remains officially closed to traffic. In last month’s commentary we outlined some of the negative second-order effects of the lack of supply of oil, gas, and other derivative products from the Gulf region, but the impact has so far been limited. 

Alternative pipeline access to Gulf crude oil has been a key element in maintaining some of the flows, while flexibility has been seen elsewhere in the supply chain. For example, some sanctioned Russian oil has been allowed to return to the market. Refiners in Europe have increased jet fuel production to take advantage of wider spreads – the profit margin between the cost of crude oil and the selling price of the jet fuel refined from it. Jet fuel might now be more expensive, but at least it’s still available. There’s also evidence of lower overall demand for crude as consumers switch to alternatives. 

It also now seems there was far more inventory immediately available than first thought, either in national governments’ strategic reserves (notably in China) or stored in tankers at sea. This might well be why the US administration has been willing to string out the negotiating process for so long. Even so, this cushion won’t last forever and some sort of peace plan needs agreement soon. On Polymarket (a website where people trade on the likelihood of future events), the probability for Strait of Hormuz traffic to return to normal by the end of June is just 20%, vs 50% at the beginning of May. It’s now 40% for the end of July.

Our playbook for this conflict has remained consistent since the early days. At the most basic level, we concluded that it would suit neither Iran nor the US to conduct a prolonged campaign. The former has limited economic resources and a longer war would risk societal and regime collapse. In the case of the US, the mid-term elections loom ever larger and the American electorate has limited appetite for such overseas interventions, especially ones that contribute to higher domestic fuel costs. This playbook has dissuaded us from taking a more defensive stance, even in the face of apocalyptic headlines and warnings of the beginning of World War Three

Of course, a positive outcome is not guaranteed. And a re-escalation of hostilities, especially anything that results in the destruction of more physical infrastructure, would be bad for markets. Even so, while constantly monitoring developments, we currently judge it right to remain fully invested in equities, relative to benchmarks. 
 

US stocks: growing into the valuation

One concern about US equities, which continue to dominate global indices, is that they look expensive. But this has been a concern for some time, and it has not hindered their progress. We’ve long believed that a simple price/earnings ratio (PE) valuation – a measure comparing a company’s share price with its earnings – does not fully capture the profitability of the leading companies. Their ability to earn returns well above their cost of capital and reinvest those profits into new, equally profitable growth opportunities is far more important. 

This thesis was tested recently when the share prices of many software companies fell sharply due to the threat of disruption from AI-based competition. Some have since recovered, notably those involved in cybersecurity, many of which have reached new all-time highs. There has been enough optimism in other areas of the market to restore momentum. 

Remarkably, despite the fact that the S&P 500 has risen 11% so far this year, the 12-month forward PE ratio has fallen slightly, to around 22, thanks to a steady upgrading of earnings forecasts. First-quarter earnings growth of 26% over the previous year came in well ahead of expectations. As long as earnings continue to improve, it’s hard to see equities coming under sustained pressure. US earnings growth is forecast at 24% for calendar year 2026, followed by 13% in 2027. 

Even so, we don’t want to get carried away. Around a third of the first-quarter growth was attributable to Amazon and Alphabet marking up the value of their stakes in AI developers Anthropic and OpenAI. Certainly, both companies have generated value from their investments, but asset-based gains are not as valuable to shareholders over the long term as recurring profits from business operations, even if accounting conventions require them to have equal standing in the headline numbers. 

There’s also the fact that the spending on data centres being made by the hyperscalers is being capitalised on their balance sheets, rather than expensed through the profit and loss account (P&L). That means this expense is recorded as an asset. This is perfectly reasonable and will, in future, be recognised through depreciation charges in the P&L: the gradual writedown in the value of assets. However, that spending is being recognised as instant profit by its recipients. Some of these, such as semiconductor manufacturers, have even higher operating margins than the hyperscalers, meaning that overall margins are rising too. We emphasise that there is nothing untoward going on here, no accounting shenanigans. But the sheer amount of capex, running close to $800bn in 2026, is on a scale never before experienced, especially in such a narrow field, and is strongly boosting reported profits in the short term. 

We are always looking for reasons why profits will not be sustained. The usual suspect is an economic recession. If not caused by some exogenous factor, outside the normal economic factors (with Covid being the most extreme example in recent history), recessions tend to be caused by a tightening of monetary policy. While there is a threat of higher rates to offset the combined inflationary pressures of the war in Iran and the AI-related capex boom, futures markets currently price in only a single quarter-point increase from the US Federal Reserve (Fed) over the next year or so. 

It would be highly unusual for the Fed to initiate a tightening cycle and then stop at ‘one and done’ (it did in 1997, but was deterred from going further as Asian economies ran into difficulties). This suggests that the current pricing is wrong – but which way? We might get some clues when the new Fed Chair, Trump-appointee Kevin Warsh, attends his first meeting later this month (17 June). There’s an expectation that he will lean ‘dovish’ because that is what President Trump will have wanted to hear in the interview process for the Fed’s top job. But Warsh has a history of being ‘hawkish’, as seen during his previous tenure on the Fed’s Board of Governors (2006 –2011). Of course, he has to take account of the other voting members’ views – and they have been more divided recently. It promises to be a more interesting meeting than usual. 
 

More AI

Our longstanding attitude towards AI has been not to bet against it. There have been numerous scares in the last couple of years, ranging from concerns about prospective returns on capital to slow adoption or the risk of competition to the US industry leaders from China – the ‘DeepSeek moment’ in early 2025 (DeepSeek is China’s leading generative AI company, which unveiled a breakthrough product at that time). We have maintained, and even added to, investment exposure to AI throughout. 

It would be fair to say that some returns have been more elusive. For example, the recent explosive strength in memory semiconductor stocks has been difficult to fully participate in. This is because of the speed with which companies in a historically highly cyclical industry that had long delivered low returns on capital suddenly turned extremely profitable. This was driven by the massive increase in demand from data centres as their workload shifted from training large language models (LLMs), the basis for generative AI models such as ChatGPT and DeepSeek, to providing the computing power for AI agents. With limited production capacity for the foreseeable future and ballooning demand, it appears that the chip makers could be generating abnormal profits for some time to come. 

Another problem with emergent industries is that it’s not clear which companies are going to be the winners. At the beginning of 2025, Alphabet was seen as a loser from the growth of generative AI, owing to the risk that its advertising business would be disrupted. Far from it. But Google’s ads business is now better targeted (thanks to AI) and more profitable than ever. And it has a first class LLM (Gemini), as well as developing its own chips. Google has been by far the best performing of the Magnificent 7 US tech giants over this period. 

A battle for dominance taking place away from the scrutiny of public markets is the one between two relative newcomers, the LLM developers OpenAI (ChatGPT) and Anthropic (Claude). The former raised $122bn in new private funds in March, at a valuation of $852bn. The latter has just raked in $65bn, valuing it at $965bn. The companies have been playing valuation leapfrog for a while. Anthropic has now filed confidential documents with the US authorities to conduct an IPO, so public investors will have a chance to see the numbers and make their own judgement (having first coughed up what’s likely to be a further few tens of billions of dollars for the privilege). Open AI will, in all probability, follow suit. 

Our research suggests that the world remains early in the adoption cycle for AI-based tools, making this remains an investment opportunity not to be missed. But, if one includes the downstream beneficiaries of all this capex (construction, power, raw materials, etc), more than 40% of the global equity market could be exposed to this one theme. It’s important not to allow portfolios to become too concentrated, just in case something goes wrong. Potential banana skins include: 

• Insufficient returns leading to lower capex 
• Societal, regulatory or environmental pushback 
• Exogenous growth shocks 

• An inflation-driven increase in the cost of capital (via higher interest rates and bond yields). 

  
   

Conclusion: still dancing

Perhaps the most famous quote from the period running up to the global financial crisis in 2008 is this one from US bank Citigroup’s CEO, Chuck Prince: "When the music stops, in terms of liquidity, things will be complicated. But as long as the music is playing, you've got to get up and dance. We're still dancing." It feels a bit dangerous to invoke it now because it suggests something dreadful lies ahead. Not necessarily. 

We’ve consistently pushed back against fears of popping bubbles and imminent financial crises, but we have to acknowledge evidence of increasingly speculative behaviour. This can be seen in niche sub-sectors such as space technology and quantum computing, both long on narrative and promises, but short on current profits. Expanding margin debt, the increasing use of leveraged (loaded up with debt) exchange-traded-funds, and the explosion in the options market, all point the same way. But market peaks are usually associated with tightening monetary policy and that is not currently a problem. And, as mentioned earlier, earnings continue to grow. Still, we continue to seek to participate in these rising markets but on a prudent basis and with a diversified exposure across and within asset classes.  
 

US semiconductor industry stocks are on a tear