It says something about Nvidia’s current state when a trillion dollars in sales counts as mundane.
Nvidia hasn’t achieved that yet. But the company used its GTC conference this week to project that much in revenue for just two of its chip families for the three-year period ending in 2027. And Chief Executive Officer Jensen Huang later made clear that forecast is a floor, as it doesn’t include some of the company’s latest offerings like the new Groq chip for AI inferencing that was announced at the event.
Wall Street’s projections for Nvidia’s business were already robust. But the trillion-dollar forecast still counts as a beat, as analysts were expecting Nvidia’s total data center revenue of around $965 billion for the specified period, according to consensus estimates from Visible Alpha.
And yet, investors are yawning. Nvidia’s stock has barely budged over the two days since the conference started. That continues a subpar recent performance, with the shares slipping more than 2% since the start of the year. Nvidia’s stock has essentially flatlined even as its results—and projected business—keep surging higher. The world’s most valuable company now commands a multiple of projected earnings below the S&P 500 for the first time in more than a decade, according to data from FactSet.
There is a mix of reasons. Investors have been growing more concerned with the sustainability of AI spending. Most of the megacap tech giants draining their free cash flow to buy Nvidia’s chips have seen their stocks fare even worse this year.
The capital spending plans broadcast by those companies over the last couple of months also didn’t leave much mystery as to who would be the biggest beneficiary. Daniel Newman, CEO of market research firm Futurum, noted that even the $1 trillion forecast “did not fully answer the question the market is asking now, which is whether Nvidia has another real upside leg beyond the already massive AI infrastructure cycle that investors have been underwriting for two years.”
Nvidia’s already enormous market cap of $4.4 trillion also creates a challenge for semiconductor investors accustomed to looking for more upside potential. “Many investors, at least the ones that talk to semis analysts, want to pick stocks that they can at least create scenarios of them doubling,” Josh Buchalter of TD Cowen wrote in a report Monday.
Nvidia also now sits at a competitive crossroads, as the evolving needs of the AI computing market grow beyond the training of large language models that the company’s GPU chips were so well suited for. Nvidia spent much of this year’s conference touting its strength in inferencing, which is when AI models produce responses to queries. Cost-effective inferencing requires a high mix of CPU chips, which are the main brains of any computing system. Nvidia designs its own CPU chips now, but that has long been a market dominated by companies like Advanced Micro Devices and Intel.
Those competitors, plus a host of startups and the in-house chip design teams of major Nvidia customers like Google, Amazon and Microsoft all see an opportunity in the shift to inferencing. Nvidia’s sky-high gross profit margins of more than 70% also make the company a tempting target for challengers that feel they can underprice the leader.
But a lower price alone won’t nullify Nvidia’s strong advantages. The company is moving fast and now pushes out significant new platforms every year. That speed is most evident with the Groq chips that will be available with the Vera Rubin platforms that start shipping later this year—less than 12 months after Nvidia secured the deal with the startup. The rapid evolution of Nvidia’s systems makes it hard for competitors to keep up. “We consistently hear that customers see a competitive product as potentially lower cost, put it into use, and come back to Nvidia,” Morgan Stanley analyst Joe Moore wrote in a report Tuesday.
Nvidia’s resources are also a major advantage in a market where vital components like memory are in short supply and seeing significant price surges. The company ended its latest fiscal year in January with $95.2 billion in manufacturing, supply and capacity commitments, which was up nearly $45 billion from just three months prior and far beyond the annual free cash flow of any other chip competitor. Nvidia will likely keep dominating the AI market for the next few years. How investors will choose to value that is far less certain.
Write to Dan Gallagher at dan.gallagher@wsj.com