Equity indices are sitting at all-time highs across the board. The S&P 500 and NASDAQ have both reached records, with the NASDAQ alone climbing 26% since the end of the first quarter. On the surface, this looks like a vote of confidence from investors. Beneath the surface, however, the dynamics fueling that rally deserve closer scrutiny. When markets push to extremes, the right question is not whether to celebrate but whether there are risks accumulating that ought to be managed.
The Earnings Story and Its Hidden Plumbing
The most obvious driver of this rally is earnings. The Magnificent 7 companies that have reported so far have posted enormous gains in both top-line revenue and bottom-line profit, with Nvidia still to come. By any traditional measure, these are spectacular, highly profitable enterprises.
Yet a closer look at the balance sheets reveals something more subtle. Across the five hyperscalers, "other income" jumped 34%. That line item represents investments in outside businesses, and it turns out many of those outside businesses are, in turn, investing back into one another. This is the circular financing dynamic that surfaced earlier in the year and now appears to be reasserting itself in a new form. The companies remain genuinely profitable, but a growing share of reported earnings is being recycled through an increasingly interconnected web. The effect is to reconcentrate both current earnings and future growth into a tighter and tighter cluster of names.
The danger here is not that any individual company is weak — it is that if one domino does not fall the way the market expects, the concentration of exposure could amplify the downside well beyond what the headline numbers would suggest.
The AI Demand Question
The justification for the enormous capital expenditure cycle these companies are running is real demand for artificial intelligence. The clearest evidence of this is in the cost of tokens — essentially the cost of computing power — which is currently elevated precisely because compute capacity is scarce. Demand is genuinely outrunning supply. As long as that AI demand stays strong and rooted, the spending will continue to flow, and ultimately it will not matter much who is investing in whom, because the money will land in the broader economy and feed real growth.
The catch is the conditional. If AI demand were to slacken, the concentration that has built up in the NASDAQ, and especially within the Magnificent 7, would suddenly look much more dangerous. That is the central reason to diversify away from the most crowded positions at this moment, even if the underlying narrative remains intact.
Broadening Leadership Is Already Happening
The good news is that the broadening of leadership is not purely hypothetical — it is already underway in the sectors that benefit from the AI build-out. If you intend to build a hundred data centers, you have to move an enormous amount of earth, which means revenue is flowing from the hyperscalers onto the balance sheets of industrials such as Caterpillar. The trickle-down extends further into adjacent businesses, including Johnson Controls, where the demand for climate control and HVAC systems is enormous given the need for tightly managed environments inside modern data centers.
All of that spending is genuinely showing up and broadening earnings into other corners of the market. But it is worth remembering that every one of these earnings streams ultimately traces back to the same source: demand for AI computing. The structural integrity of that demand is what holds the entire broadening narrative together.
Where Execution Still Stands Out
Within the technology stack, the differentiator going forward will not be who can report a big top-line number — almost everyone can do that right now — but who can convert those revenues into returns for shareholders. Free cash flow is the metric that matters, because it feeds future profitability and ultimately drives capital returns.
Dell stands out on this measure. The company has been a strong performer this year, and the reasons for that performance look durable. Its margins, profitability, and free cash flow speak to a company executing well. It is increasingly participating in the broadening demand for hardware and memory, showing up in different layers of the technology stack, while its long-standing strength in PCs continues to provide a stable foundation.
A Different Path for the Financials
In financials, JP Morgan Chase is an interesting case precisely because it has lagged this year. The setup for the back half depends less on the Federal Reserve aggressively cutting short rates by a hundred basis points and more on what happens at the long end of the yield curve. Bringing the 10-year Treasury yield back below 4% is likely to be a key, if quietly held, objective for the next chair of the Federal Reserve — a role widely expected to go to Kevin Warsh.
Every time long rates ease, capital formation accelerates. A lower long-term cost of capital opens the door to more IPOs, particularly from smaller companies looking to exit the private holdings where they have been parked for years. The largest, most anticipated names will likely come public in the back half of the year regardless, but a return to a more normal IPO environment requires a more accommodating long end.
If that scenario plays out, JP Morgan is exceptionally well positioned to participate in what it does best: deal formation and investment banking. The investment banking revenue that has been sitting on the sidelines for an extended stretch could finally start flowing back, and JP Morgan would capture a disproportionate share.
The Takeaway
Record highs are not, in themselves, a reason to retreat. The companies driving the index higher are genuinely profitable, and the demand for AI compute that underwrites the capital cycle is genuinely strong. But the structure beneath the surface is more fragile than the headlines suggest, with concentration risks compounded by the circular flow of capital among a tight group of hyperscalers.
The prudent response is not to abandon the leaders but to broaden the portfolio toward companies that are either picking up the second-order benefits of the AI build-out — industrials, climate controls, hardware execution stories — or that stand to gain from a lower long end of the yield curve and a thawing capital markets cycle. In a market this top-heavy, the path forward is diversification rooted in real cash flows and real beneficiaries, not just exposure to the narrative everyone already owns.