In the span of a few days in early June, AI and chip stocks lost $1.3 trillion in market value. U.S. semiconductor names got hit hard. Investors who had been riding the AI wave suddenly started asking whether the trade had run too far ahead of reality.

Jensen Huang was in Seoul on June 8 for a series of business meetings. Instead of staying quiet, he walked out and told investors what he thought they should do with a down market: buy it.

What Jensen Huang said about the AI boom and the stock market sell-off

“We’re at the beginning of it, and whatever happened to the stock market, you should be very happy because now you can buy at a discount,” Huang told reporters in Seoul, according to Bloomberg. “Everybody should be very excited.”

The trip itself was part of the story. On the same day, Nvidia and SK Hynix, one of the world’s largest memory chip manufacturers and a key supplier to U.S. AI data centers, announced a multi-year partnership to design future generations of AI memory chips.

More Nvidia:

• Bank of America sees Nvidia’s next $20 billion business
• Morgan Stanley says Nvidia stock remains top pick despite headwind
• Citi sends strong signal to Nvidia investors amid rumors

That deal signals that Nvidia is still building, still signing long-term agreements, and still betting on the infrastructure buildout that Huang has been describing for the past two years. You don’t sign a multi-year chip partnership on the day you think the cycle is turning.

Huang also called AI infrastructure “a foregone conclusion,” comparing it to the early days of the internet.

The framing is deliberate. He’s not saying AI will happen; he’s saying it’s already happening, and the only question is how much capacity gets built, Seeking Alpha reported.

The Nvidia AI infrastructure numbers behind Huang’s confidence

Huang’s confidence isn’t just rhetorical. Nvidia’s Q4 fiscal 2026 revenue came in at $68.1 billion, up 73% year over year, as TheStreet reported. Full-year revenue for fiscal 2026 came in at $215.9 billion. Data center revenue was $62.3 billion in Q4 alone, making up 91% of total quarterly revenue.

The company had $500 billion in AI chip bookings covering 2025 and 2026 combined, and CFO Colette Kress confirmed that figure has since grown as full-year orders for Nvidia’s next-generation Rubin chips came in.

At Nvidia’s annual GTC conference in March, Huang went further. He predicted that Nvidia’s AI processors would generate $1 trillion in sales through 2027, according to Bloomberg. CFO Kress added on the May earnings call that hyperscaler capital expenditures are expected to exceed $1 trillion in 2027 alone, putting AI infrastructure spending on track to hit $3 trillion to $4 trillion annually by the end of the decade.

These aren’t projections built on hope. They reflect committed orders from cloud providers and enterprise customers who have made AI a core part of their capital plans.

There’s a Bloomberg analysis worth noting alongside those numbers. It found that up to half of US data centers planned for 2026 are facing delays or cancellations, with only 5 gigawatts under construction out of 12 to 16 planned. Huang’s implicit argument is that those delays actually reinforce his point: the world still needs far more compute than it’s building, which means the demand doesn’t go away, it just stacks up.

If Huang is wrong and AI spending slows, it’s not just NVDA that feels it. AMD feels it. Broadcom feels it.

Spencer/Getty Images

Why Nvidia stock is a proxy for the entire AI trade

Nvidia briefly became the world’s first $4 trillion company in July 2026, with its stock hitting $164.42 intraday. Broadcom, AMD, Super Micro, and Marvell all rose in the same AI wave. When Huang speaks at earnings calls or events, those stocks get a read-across. They’re all in on the same buildout, and his read on that buildout carries through all of them.

If Huang is wrong and AI spending slows, it’s not just NVDA that feels it. AMD feels it. Broadcom feels it. The data center power companies, the networking names, the cloud providers sitting on hundreds of billions in AI capex commitments. They all lose the bet at the same time.

Calling the sell-off a buying opportunity is Huang putting his credibility behind the whole chain, not just his own stock.

Huang’s answer to all of this is one he’s been giving all year. The AI buildout isn’t a product cycle that rises and falls in a few quarters. It’s a decade-long infrastructure project, and the world is still early in it.

He’s compared it to the internet, which kept growing for decades after people thought the initial boom was over. A bad week on the Nasdaq doesn’t change what’s being built underneath it.

What Nvidia investors should watch as the AI boom continues

The key test for Huang’s thesis isn’t what he says in Seoul. It’s what the hyperscalers say in their next earnings calls. If Microsoft, Amazon, Meta, and Alphabet hold their AI capex guidance, the sell-off looks like exactly what Huang called it: a discount on something still growing. If any of them pulls back significantly, the conversation changes.

Huang has already described what he’s watching for on Nvidia’s own side. He said Nvidia expects to face a backlog of chip demand and that the company is constrained not by demand, but by its ability to manufacture and deliver at scale.

That’s a very different problem from facing a slowdown. It’s the problem of growing faster than your supply chain can keep up with.

For investors sitting on Nvidia stock after the tech rout, the choice the Seoul comments were designed to address is a familiar one: Do you trust the long-term thesis, or do you follow the short-term price action?

Huang’s answer was unambiguous. The AI buildout is still at the start. The market just gave you a sale.

Related: Citi sends strong signal to Nvidia investors amid rumors