We reach August with markets that are holding up despite the noise around them. The war with Iran is still open, oil is a worry, semiconductors have corrected sharply from their highs and US equity valuations are high. For us, though, what actually moves markets is something else.
At the start of the year, when defining our market strategy, we identified three major drivers for 2026: inflation, artificial intelligence and liquidity. But we also gave a lot of importance to corporate earnings. In fact, we already said that they had been one of the major drivers of the stock markets during 2025, and we warned that one of the risks for this year would be precisely that they might not be able to sustain that pace of growth. We think we are reaching a point where this issue will become increasingly important.
Corporate earnings under scrutiny
We have talked many times in recent months about whether or not there is a bubble in artificial intelligence. Our opinion has not changed substantially: we do not think that, for now, we are facing a bubble like the one seen with dot-com companies in the year 2000. It is true that share prices have risen a great deal, but so have corporate earnings. Ultimately, it is companies that are listed on the stock market and, over the medium and long term, the value of these companies depends mainly on their ability to generate earnings. A bubble appears when share prices move excessively far from this reality. And the data we are seeing, for now, continues to be good. The percentage of S&P 500 (the main US index) companies beating earnings estimates is at exceptionally high levels. But this is precisely where we think we need to start being a little more demanding. After the accumulated gains, it will no longer be enough for earnings to be good: they will need to keep justifying the expectations the market has already priced into share prices.
There is one figure we think we need to watch especially closely over the coming months: the gap between the earnings companies have already reported and the earnings analysts expect them to generate in the future. If we measure the S&P 500’s valuation against the earnings analysts expect over the next twelve months, the P/E ratio (the relationship between a company’s share price and its earnings) stands at around 21 times. That is a high valuation, true, but only slightly above the 30-year average of around 19 times. If instead we use already-reported earnings, we get closer to 28 times. This gap matters because the market is assuming that future earnings will keep growing strongly (the P/E ratio’s denominator will increase).
We believe this points to two possible scenarios:
- That companies keep increasing earnings and meet or beat estimates. In this case, the gap between P/E ratios would close naturally.
- That analysts have become too optimistic and companies start reporting results below what the market expects. In this case, the adjustment would come not through earnings but through share prices.
Therefore, we do not think there is a bubble in the US stock market today, but there are elevated expectations that leave earnings less and less room to disappoint. And this distinction matters. It is not the same to think the market is about to collapse as it is to assume that, after the accumulated gains, a correction falls perfectly within the range of normal, without implying a prolonged bear market.
Semiconductors: correction, trend change or rotation?
A good example of what we have just explained can be found in semiconductors.
The Philadelphia Semiconductor Index has corrected by roughly 20% from its highs. Seen in isolation, that can look like a significant fall. But it needs to be put in context: over the previous twelve months it had risen by roughly 180%. We interpret the move more as a correction of previous excesses than as a bubble bursting. There has been a lot of optimism, leverage and very high expectations. We see the disappearance of part of those excesses as healthy, if anything.
In this sense, we think semiconductors will remain essential for the development of AI. However, that does not necessarily mean their share prices need to keep rising at the same pace. We believe the economy will keep demanding computing capacity and will therefore consume more semiconductors. So, we think we should stay exposed to this subsector of technology. That said, since the market also knows this, much of this story is already priced in. That is why we think we are probably entering a different phase. And here there is a question we are starting to watch with particular interest: the possibility that part of the capital starts looking for the next beneficiaries of the AI revolution. Until now, the revolution that AI represents at an economic level has focused market attention first on software, and then on computing capacity (semiconductors) and data centres. But all these data-centre chips need an enormous amount of energy to run. We think the next major bottleneck for AI could be not just the ability to manufacture semiconductors, but the ability to generate, transport and guarantee enough energy to power all this infrastructure. So, we could be looking at an interesting rotation within the same technological revolution: not necessarily exiting AI, but shifting from some of the big beneficiaries of the first and second phases towards the sectors that need to make the next one possible: energy, power grids, generation and associated infrastructure could gain increasing prominence. This does not mean we think one should exit software, semiconductors or data centres, but rather that we think the chain of AI beneficiaries can broaden.
This is an idea we have mentioned before and one that guides our analysis of how this technology is developing: first, those who build the infrastructure benefit; then, those who enable that infrastructure to run; and finally, those who use it to boost their productivity. That is why, rather than interpreting the semiconductor correction as the end of the AI story, we think it may be worth starting to ask ourselves where the next bottleneck will be and, therefore, where the next investment opportunity might appear.
Iran: what if the conflict lasts longer than we think?
Since the conflict began, we have worked on the assumption that it would be relatively short. We still think that is the most likely scenario, but as time goes by we need to factor in the possibility that neither side manages to prevail decisively and the conflict turns into a war of attrition, similar to Russia-Ukraine.
We believe Trump has strong incentives to try to resolve the situation before the US midterm elections in November. A prolonged conflict, especially if it keeps petrol prices high, carries an obvious political cost in the United States. But we also think ending things halfway is a problem: withdrawing without a minimally favourable agreement could be perceived as a defeat. And a much more forceful military intervention would also carry significant risks. This could end up leading to an intermediate situation: military and economic pressure on Iran, negotiations, pauses, new episodes of tension, and a resolution that keeps being pushed further down the road.
From the markets’ point of view, however, we still think the truly important variable is not so much the conflict itself as the price of oil, its knock-on effect on inflation, and the path of interest rates. As long as crude stays at levels that do not create new inflationary problems, we think markets can live with a low-intensity conflict for quite a while. The problem would arise if oil sustainably rose above levels close to or above $90 and this started feeding through into inflation expectations and rate policy.
We would therefore return to the first of the drivers we defined at the start of the year: how inflation can move markets.
This is where an interesting connection appears with what we have just explained about AI. If energy becomes an increasingly sought-after resource to power data centres and, at the same time, we have geopolitical tensions that can put pressure on oil and other energy sources, the energy sector could go from being simply a hedge against inflation to also becoming a structural piece of technological development. This is an issue we think is worth following very closely.
In conclusion: we remain constructive on markets, but we think that after the accumulated gains we need to be more vigilant. Valuations are elevated and, precisely for that reason, corporate earnings become even more important. This was one of the issues we had already identified at the start of the year, and it now enters a phase where it will be especially important to check whether corporate reality continues to validate market expectations. As long as earnings keep growing and meeting expectations, we see no reason to talk about a broad-based bubble or to radically change our view. But the higher valuations are, the smaller the margin for disappointment.
In technology, we think the AI revolution remains fully in force. The semiconductor correction looks to us more like a normalisation of excesses than the end of the AI cycle. But it may well indicate that we are entering a new phase, one in which the spotlight gradually widens towards the new bottlenecks created by this revolution, and we think energy could be one of the most important ones.
And on the geopolitical front, we are factoring in the possibility that the conflict with Iran could last longer than we initially expected. But we will keep paying less attention to headlines and more to the variables that can genuinely affect portfolios: oil, inflation, interest rates, volatility and corporate earnings.
We wish you a very good rest of the summer!