Nvidia CEO sends serious wake up call to all Americans

In March 2026, federal prosecutors charged the co-founder of Supermicro with masterminding a $2.5 billion scheme to smuggle Nvidia-equipped servers to China through front companies, forged documents, and dummy servers staged to fool auditors. Jensen Huang had been watching this unfold for months. On June 24, standing before Nvidia's annual stockholder meeting, he addressed it directly.

But what Huang said went beyond the Supermicro case. He used it to make a broader argument about what America's AI advantage actually depends on and why both the shortcuts and the fears surrounding AI are missing the point entirely.

Why Huang called smuggled AI data centers a dead end

Huang told shareholders on June 24 that if a commercial opportunity conflicts with U.S. national security, Nvidia (NVDA) would prioritize American interests without hesitation.

"National security comes first," Huang said, according to CNBC.

He then turned to the smuggled chips problem directly. Companies trying to build AI data centers with diverted Nvidia hardware will find that those systems cannot function because Nvidia will not provide support or repairs for them.

More Nvidia:

• 5-star analyst revisits Nvidia stock on peculiar GPU demand situation
• Bank of America resets Nvidia stock forecast after CFO meeting
• Nvidia CEO sends strong message to stock market investors

"Advanced AI data centers are massive integrated systems that require trusted hardware, software, networking, and continuing support," Huang said.

Without ongoing software updates and technical support, systems built on smuggled chips cannot run at the scale modern AI demands. The Next Web reported that Huang told shareholders anyone trying to build around export controls would face operational failure, not just legal exposure.

Supermicro's co-founder allegedly used heat guns to swap serial numbers off packages and staged warehouses of fake servers to fool compliance auditors.

Huang's message on June 24 was that the operation was doomed regardless. Getting the chips through customs is one problem. Getting them to actually work without the company that built them is another.

What Huang's national security message means for American AI strategy

Huang has been making this argument about America and AI all year. What the June 24 meeting added was the national security frame around it.

Earlier this year, when Nvidia announced a manufacturing partnership with Corning to expand domestic optical connectivity capacity tenfold, Huang described what the AI buildout means for workers outside the tech sector.

"This is such an extraordinary opportunity because we can use these market dynamics to reinvest, revitalize American manufacturing for the first time in several generations," he said.

He pointed to rising demand for electricians, construction workers, and infrastructure specialists as evidence that the buildout was already spreading beyond Silicon Valley. The partnership centers on the optical technologies needed to connect chips inside massive AI data centers, which is precisely the kind of physical infrastructure Huang keeps coming back to.

At Davos in January, as TheStreet reported, Huang framed AI as a five-layer system: energy, chips, infrastructure, models, and applications. All five need to scale together. Countries that invest only in models and chips while neglecting energy and manufacturing will find themselves bottlenecked at a different layer.

The smuggled chip problem fits that same frame. When US executives are willing to risk federal criminal charges to route Nvidia hardware into China, it tells you something about how large the gap has become. Huang is saying the U.S. built that gap through the whole stack, not just through one layer of it.

Cho&solGetty Images

What Huang told investors about AI returns and Nvidia's position

Huang also used the June 24 meeting to address the most persistent doubt about the AI spending cycle: whether the returns are real. He did not equivocate. The question of AI return-on-investment "has been answered," he told shareholders.

He used GitHub as the example. Pull requests nearly tripled this year because of AI. Companies generating useful AI output are finding it profitable and are actively seeking more computing power to do more of it.

"Nvidia systems may not be the cheapest to purchase, but Nvidia generates the lowest cost tokens, the highest token throughput, and the most revenues," Huang said.

Nvidia generated over $96 billion in free cash flow in fiscal 2026. About 9% of revenue came from China, including Hong Kong, down from roughly 13% the prior year. The U.S. cleared Nvidia's H200 chip for export to approved Chinese customers, but Huang told shareholders no revenue has been generated from those licenses yet, and the company does not know whether China will allow imports.

As TheStreet reported in May, Huang told a ServiceNow conference that the compute needed for agentic AI has increased 1,000% compared to generative AI in just two years.

If that trajectory holds, the infrastructure buildout he is calling on America to accelerate is still in its early stages. The June 24 message was less about what has been built and more about whether the country is moving fast enough to build what comes next.

Huang has been saying versions of this all year. At Davos, at GTC, at the Corning announcement. The stockholder meeting on June 24 was the same message, delivered to the people who own the company, with the Supermicro case as the backdrop. National security first. The ecosystem is the advantage. The returns are already showing up. And the buildout is nowhere near done.

Related: Nvidia's $25B bond deal sends investors a clear signal