Nvidia Earnings Test Whether AI Spending Boom Can Survive Market Doubts

Nvidia Earnings Test Whether AI Spending Boom Can Survive Market Doubts

Nvidia reports quarterly results Tuesday evening with investors hunting for proof that the AI infrastructure buildout—now in its third year—can justify the trillions in market value it's created, even as Big Tech customers quietly hedge their bets on alternative chip suppliers.

The timing is awkward. Nvidia's stock ticked higher this morning ahead of the announcement, but the company faces its most skeptical audience since the generative AI boom began in 2023. Weeks of tech selloffs have left CFOs and investors asking the same question: Are we still in an infrastructure gold rush, or are we watching the final innings of a capital spending cycle that's about to roll over?

Here's what makes this earnings call different from the victory laps Nvidia has been running since ChatGPT launched. The company still controls the vast majority of the market for GPUs—the chips that train and run large AI models. Its CUDA software platform remains the industry standard, the thing that keeps developers locked into Nvidia hardware even when alternatives exist. And its product roadmap—Hopper chips giving way to Blackwell, with Rubin already on deck—suggests a company that has no intention of slowing down.

But the cracks are starting to show, and they're the kind of cracks that finance leaders notice first. Meta, which days earlier committed to deploying millions of Nvidia GPUs, announced this week a multibillion-dollar deal to buy chips from AMD instead. The kicker: Meta gets the option to take up to a 10% equity stake in AMD, mirroring a similar investment AMD secured from OpenAI last October. That's not a hedge—that's a strategic pivot dressed up as diversification.

The cloud giants are making similar moves, even as they remain among Nvidia's largest customers. Amazon has begun deploying thousands of its own AI chips across a sprawling network of data centers in Indiana, where they're being used by Anthropic. Google has struck a series of deals with Anthropic and is reportedly supplying chips for several of the startup's new data centers in New York, Texas, and elsewhere.

Then there's the inference problem. A growing field of startups is building chips specifically for inference—the process of generating outputs from trained AI models—rather than the training workloads where Nvidia dominates. Nvidia has moved to hedge against that threat, entering a high-profile licensing deal with Groq, one of the most well-known inference startups, and bringing CEO Jonathan Ross and other staffers to Nvidia in the process. The deal is nonexclusive, which means Groq can still sell to Nvidia's competitors, but it's a telling sign that even Nvidia sees the inference market as a potential vulnerability.

Analysts expect data center revenue, adjusted earnings per share, and gross profit margin to rise when Nvidia reports after the bell. The question is whether those numbers will be enough to convince investors that the AI spending boom has staying power, or whether we're watching the market slowly price in a future where Nvidia's dominance starts to erode at the edges.

For CFOs watching their cloud bills and weighing their own AI investments, the subtext matters more than the headline numbers. If Meta and Amazon are building escape routes from Nvidia dependence, what does that say about the durability of the current infrastructure stack? And if the answer is "not much," what does that mean for the AI budgets finance leaders are being asked to approve for 2026 and beyond?