Nvidia's Jensen Huang Says Nobody Uses AI Better Than Meta, Right After Meta Stock Dropped 7%

Jensen Huang doesn't hand out compliments for free. So when the Nvidia CEO told CNBC that nobody uses AI better than Meta, people noticed the timing as much as the words.

Meta's stock had just dropped 7%. Investors were spooked by the company's 2026 capital spending plan, which Meta has set between $125 billion and $145 billion, almost all of it going into AI infrastructure. That's an enormous number even for a company Meta's size, and Wall Street wanted a reason to believe it'll pay off.

Huang gave them one. His argument wasn't just cheerleading. He pointed to something specific: Meta already made the hard technical leap from old-school CPU-based systems to generative AI architecture. That's not a promise about the future. It's a completed transition with real products behind it, according to Huang, including better Instagram and Facebook content recommendations, sharper ad targeting for advertisers, and new AI tools that let brands generate ad creative automatically.

That distinction matters. Plenty of companies in this AI cycle are selling a vision. Huang's point, as reported by Crypto Briefing, is that Meta is one of the few that can point to actual revenue and engagement numbers tied to its AI deployments, not just a roadmap.

The Obvious Conflict of Interest

Nvidia builds the GPUs that power Meta's AI infrastructure. Meta is one of Nvidia's biggest customers. When Huang says Meta uses AI better than anyone, he's also saying the hardware underneath it works great, and he's talking his own book.

That doesn't make him wrong. But it means his praise should be weighed the same way you'd weigh a supplier praising his best customer. Nobody covering this story, including Crypto Briefing and Cryptonomist, pretended otherwise. Both outlets flagged the self-interest angle directly. Give them credit for that.

The Market Reaction Was Small, Not a Rally

Meta shares climbed nearly 1% in premarket trading the day after Huang's interview aired. That's a bounce, not a comeback. The stock was still down net from its pre-decline level. Calling this validation of Meta's spending plan would be overselling what actually happened. A modest, single-day premarket tick followed a week of investor anxiety.

Wall Street's broader position on Meta hasn't changed much regardless of Huang. Of 64 analysts covering the stock, 57 rate it Buy or higher, with an average price target of $826.75, implying roughly 38% upside from where shares traded around the time of Huang's comments. That bullishness predates Huang's CNBC appearance. It reflects existing analyst models on Meta's ad business, not a reaction to one interview.

The Bigger Fight: Open Versus Closed AI

There's a second Nvidia story running in parallel that gives Huang's Meta comments more context. According to The Daily Upside, Nvidia has spent months positioning itself firmly on the open-source side of an industry split, versus OpenAI and Anthropic's closed-model approach.

Huang used Nvidia's March developer conference to pitch Nvidia chips as ideal infrastructure for open-source tools. He opened a personal account on X largely to promote open-source AI. This week Nvidia released a new open-source model, Nemotron 3.5 Lightning, along with a routing tool called NeMo Switchyard that directs AI tasks to whichever model handles them best.

Bill Wong, an AI research fellow at Info-Tech Research Group, told The Daily Upside that Nvidia's motive here is straightforward: "Software traditionally is where the value is, and that is also where a lot of the costs are. What Nvidia is doing is removing that cost," freeing up enterprise budgets to spend more on hardware, Nvidia's actual business.

Nicolas Sauvage of TDK Ventures put it differently to The Daily Upside: "Open models do not eliminate value. They change where value is captured." His bet is that the future isn't one model winning outright, but systems that dynamically pick whichever model fits a given task.

Put those two threads together and Huang's Meta praise looks less like a standalone opinion and more like part of a consistent strategy. Cheaper, more open AI models push spending back down to hardware and infrastructure. Meta, with its enormous capital budget and open-weight Llama models, fits that thesis better than almost any other big tech company.

What's Still Unresolved

Huang's claim that Meta's AI spending is already showing up in earnings is checkable, and Meta's own quarterly results will be the actual test, not a CNBC soundbite. The open question is whether $125 billion to $145 billion in annual AI capex produces returns that justify it, or whether Meta is locking in spending commitments based on a moment of investor anxiety that a well-timed endorsement helped smooth over. Meta has not disclosed a breakdown of what portion of that budget is contracted versus discretionary, so the actual downside risk if AI returns disappoint remains unclear.