The AI trade is no longer about owning more Nvidia
Nvidia’s latest earnings kept the AI spending boom intact, but the stock’s muted reaction said something more important: the market is no longer paying up just for obvious exposure. From here, the better trade looks like selective owners of AI bottlenecks and monetization layers, not simply adding to the biggest GPU winner.
The message from Nvidia’s post-earnings tape was not that AI is fading. It was that the easy phase of the trade is over. When a company can still post huge growth and keep the capex story alive, yet fail to spark the usual euphoric response, the market is telling us the debate has shifted from demand to economics. The question now is not who sells into AI, but who captures durable margin pools as hyperscalers pour hundreds of billions into infrastructure.
That is why we think the AI trade needs a reframe. Nvidia remains the central franchise in the buildout, and the numbers still justify respect:
trades at 33.19x earnings with 65.5% revenue growth and a remarkable 63.0% net margin. But that is exactly the point. Everyone already knows Nvidia is the core beneficiary, and after a run to a $5.24 trillion market cap, the bar is no longer “great quarter.” It is “great quarter, plus evidence that upside is still underappreciated.” The muted reaction after May 21 suggests investors are becoming less interested in owning the obvious winner at any price and more interested in where the next layer of AI economics sits.
If that sounds like a call against semis, it is not. This week’s tape still says semiconductors are leading. What has changed is the selectivity. MU is up 139.5% year to date, MRVL 117.9%, and AMD 107.6%, versus NVDA at 14.4%. That is not random dispersion. It is the market rewarding exposure to memory, networking, and alternative compute paths where expectations were lower and operating leverage to the capex wave may still be expanding. When leadership broadens away from the flagship name even as the theme remains intact, that usually signals maturation, not collapse.
The best way to read that shift is through bottlenecks. AI infrastructure is not just a GPU story; it is a systems story. Networking has become one of the clearest examples. ANET is not cheap at 52.22x earnings, but it is growing revenue 28.6% with a 38.3% net margin, which is exactly what a scarce-enabler profile looks like. Broadcom is the more forceful version of the same idea. AVGO screens expensive at 80.08x earnings, yet its appeal is not simple multiple expansion; it is that Broadcom sits across custom silicon, connectivity, and software, which gives it more than one way to monetize AI demand. That matters in a market increasingly asking who gets paid repeatedly, not just who ships the most coveted component.
NVDA: 33.19x P/E, 65.5% revenue growth, 63.0% net margin
AVGO: 80.08x P/E, 23.9% revenue growth, 36.6% net margin
ANET: 52.22x P/E, 28.6% revenue growth, 38.3% net margin
MU: 34.07x P/E, 48.9% revenue growth, 41.5% net margin
AMD: 156.07x P/E, 34.3% revenue growth, 13.4% net margin
Micron may be the cleanest example of why “own more Nvidia” is no longer enough as a framework. If AI infrastructure spending is set to exceed $700 billion this year, the memory content of that buildout is not optional. MU trades at 34.07x earnings with 48.9% revenue growth and a 41.5% net margin, while its TickerSpark Score reflects one of the stronger combinations of cyclical recovery and AI leverage in the group. That does not make Micron safer than Nvidia; it makes it more interesting in a market now looking for less crowded ways to express the same theme. The same logic applies to Marvell, where 42.1% revenue growth and a 117.9% YTD move point to investor appetite for names tied to AI data movement and custom infrastructure rather than only front-end compute.
Yes, Nvidia bulls have a fair rebuttal: the company still has the best economics in the group, and a 33.19x P/E is not outrageous against 66.0% EPS growth and 63.0% net margins. If AI capex keeps accelerating, the cleanest way to own that may still be the company selling the most indispensable chips. But that argument misses what the tape is saying. A stock can remain a fantastic business and still become a less compelling trade when ownership is crowded and expectations are already built around dominance. Markets do not stop rewarding quality; they stop overpaying for certainty when adjacent beneficiaries start to show they can capture meaningful slices of the same spend.
AMD shows both the opportunity and the risk in this next phase. On one hand, a 107.6% YTD gain tells us investors are eager for credible alternatives to Nvidia. On the other, AMD at 156.07x earnings and just 13.4% net margin is a reminder that not every “next Nvidia” narrative deserves the same confidence. The better hunting ground is not simply second-place GPU exposure. It is businesses with structural leverage to AI scaling where the economics are harder to displace: memory, networking, custom silicon, and the connective tissue around hyperscaler capex. That is why Broadcom, Arista, and Micron fit this moment better than a generic basket of AI hopefuls.
The AI trade is still alive. What is dying is the lazy version of it. Nvidia’s earnings showed demand remains real, but the stock’s reaction showed the market wants more than proof that spending continues. It wants evidence of incremental upside, cleaner valuation support, or exposure to parts of the stack where the margin pool is still being discovered.
What we would watch from here is simple: whether AI capex keeps broadening into networking, memory, and custom infrastructure with the same force it first hit GPUs. If that broadening stalls, Nvidia can reassert itself as the only game that matters. But if the current tape holds, the winning AI trade from here is less about owning the biggest name and more about owning the chokepoints and monetizers around it.
Our take, not advice. This is opinion commentary — informational only, not personalized investment recommendations. Markets carry risk. Do your own research and consider your own situation before any trade.
▌The Daily Briefing · Free
A new stock idea, every evening.
One stock worth watching each weekday, plus the analysis behind it. Free, in your inbox.
