School might be out but it looks like investors will return from the summer facing a market still concentrated in a single trade.
Information technology and communication services have generated roughly three-quarters of the S&P 500's aggregate earnings growth, and the ‘Magnificent Seven’ (though that collective name seems less and less relevant) continue to outrun the broader index.
Against this backdrop, what can we learn? To me, three lessons stand out. The first is that selectivity will matter more than playing any given theme.
For much of the past year, exposure to artificial intelligence has been rewarded almost indiscriminately. During July's second quarter reporting season, earnings were on track to grow 36% year-on-year and 85% of companies beat expectations that had already risen going in.
Despite this, a few of the largest technology companies still struggled. Beating forecasts was no longer enough. Investors wanted evidence that continued increases in AI capital spending would generate an attractive return. Where they did not find it, they sold up.
What does this mean? As the market increasingly tries to distinguish the winners and losers from AI, I think diversification across the AI supply chain remains increasingly important.
The AI theme spreads across semiconductors and infrastructure; from hyperscalers to software and hardware companies, and it is still unclear where the greatest returns on AI investment will ultimately accrue.
Maintaining exposure across these different parts of the AI ecosystem will therefore reduce reliance on any single expression of the AI theme.
The second lesson is not to confuse regional diversification with diversification away from the AI theme. In the US, AI exposure is concentrated in hyperscalers and software platforms. In emerging markets, it is concentrated further upstream in semiconductor manufacturing.
The contrasting performance in July – and the opposite pattern earlier in the year, when semiconductors were outperforming while hyperscalers struggled – illustrates how regional diversification can provide exposure to different parts of the AI value chain.
When concerns around China’s technological progress and export controls resurfaced in July, SK Hynix fell 35% and Samsung Electronics 21%, dragging South Korea down 17.1% and Taiwan 5.3%. China itself rose 9% over the month, but that gain was easily overwhelmed.
For investors, the lesson is that regional diversification can broaden exposure across the AI value chain, but seemingly diversified regional allocations can still leave portfolios heavily exposed to the same underlying theme.
The third lesson is the one most likely to be underappreciated and it points beyond AI altogether. July's volatility did not begin with earnings, it began with oil.
Rising tensions between the US and Iran briefly pushed Brent crude above $100 per barrel, reviving inflation concerns, lifting government bond yields and prompting a broadly hawkish tone from central banks even as they held rates steady.
Oil is now back down to $80 a barrel, but what we learnt should be remembered: inflation does not stay solved and the structural backdrop makes this more than a passing concern.
Unfavourable demographics, deglobalisation and sustained investment in AI and energy infrastructure all point towards inflation settling at a higher average level than investors became accustomed to during the 2010s.
Looking to the rest of the year, the risks ahead are genuinely two-sided. A reversal in AI enthusiasm would slow growth and the durability of AI earnings remains an open question. Persistent inflation would do the opposite kind of damage.
These are distinct risks and they call for distinct hedges. Government bonds remain the most effective protection against a growth slowdown while real assets offer resilience should inflation prove stickier than markets expect.
Of the two, I think the inflation leg is the risk that most portfolios are least prepared for. After three decades in which low and stable inflation flattered both stocks and bonds, many investors remain positioned for a world that may no longer exist.
Real assets deserve renewed consideration precisely because that structural risk has been underappreciated for so long.
None of this argues for abandoning the AI trade. It argues for recognising how much influence it has over portfolios and for building in protection against the two very different risks the months ahead could bring.
As the summer showed, when many investors are pulling in the same direction, the ground beneath them tends to be less solid than it appears.
Aaron Hussein is global market strategist at JP Morgan Asset Management. The views expressed above should not be taken as investment advice.