As the AI World Turns: The Jensen Huang Episode
Nvidia's board is investigating reports that Chinese intelligence targeted Jensen Huang. The stakes reach from Washington's chip controls to the financing of the next AI data center.
Nvidia's board has opened an independent investigation into reports that Chinese intelligence targeted its founder and CEO, Jensen Huang. The reporting centers on Alice Huang, a former executive at Megaspeed, a Singapore-based Nvidia customer. The New York Times reported that U.S. intelligence assessed her as working for China. Beijing dismissed the allegation. NVIDIA Newsroom
On November 8, Nvidia published a message in which Jensen Huang acknowledged being unfaithful to his wife and pledged to cooperate with the investigation. His message didn't name Alice. The Times reported that he acknowledged to the board that he had been intimate with her. The board expressed confidence in Huang and said it had no reason at that point to believe Nvidia secrets or U.S. national security had been compromised. Nvidia shares closed down 4.3% on Monday, November 9. NVIDIA Newsroom
Why Jensen Huang is worth getting close to
If the U.S. intelligence assessment is right, the choice of target tells you what Beijing considers valuable. Huang runs the company whose chips train and run many of the world's most advanced AI models. Washington restricts exports of some of those chips to China. The man running Nvidia can talk to customers about what they need and to Washington about what he's allowed to sell them. Three House committee chairs have asked him to brief them behind closed doors about his interactions with Alice Huang and Nvidia's security safeguards. apnews.com
Think about what a routine customer conversation could tell you. Who's buying enough chips to build a serious cluster? Where are those chips going? When will they arrive? A delivery schedule can tell you when a competitor might have the computing power to train its next model. Advance insight into export policy could tell you where to place your next order. Those are reasons access would be valuable. What Alice Huang actually obtained, if anything, is a question for the investigation.
The company store has a credit department
When we were talking about the Nvidia company store, we followed the money from the chipmaker to the companies buying its chips. A GPU can ship, the invoice can get paid, and you can still ask how much of that demand depends on Nvidia helping the buyer line up financing. This is where the Jensen story picks up that thread. If a lender gets nervous about the company backing the deal, the next GPU order can get harder to finance even when the customer still wants it. I Owe My GPUs to the Company Store
Nvidia disclosed $36 billion in commitments as of July 26 under arrangements with select AI cloud partners. In these deals, Nvidia can commit to paying for computing capacity, giving the operator a source of revenue that can help it secure financing. Nvidia says its commitments decrease as third-party customers or Nvidia itself use that capacity. A new facility gets built, the operator can buy equipment, and Nvidia takes on some of the risk while the operator lines up customers. investor.nvidia.com
That's where the company store still needs customers with their own money. Nvidia's backing can help get a facility built. Over time, that facility has to earn enough from people paying to use it. There's a big difference between a line out the door for GPU time today and enough paying business to support the next building for years.
Could the Jensen story move the AI buildout?
The credit market was already getting skittish. Reuters reported on November 9 that some lenders were pulling back on new AI deals even when big technology companies stood behind them. Look at Anthropic for a separate example of how quickly expectations can move. Reuters reported in September that it was seeking a valuation of about $2 trillion and that Nvidia was considering investing up to $10 billion in its IPO. By November 6, Bloomberg reported that some prospective buyers were talking about $1.2 trillion to $1.3 trillion ahead of formal pricing. reuters.com
When we were talking about OpenAI's boardroom drama, the question was who could make the next big decision. Nvidia's version is what the people outside the boardroom do if the inquiry uncovers a security breach or puts Huang's leadership in doubt. Does a lender ask Nvidia for a bigger guarantee? Does a cloud operator shrink its order or postpone construction? That's where the soap opera could hit the balance sheet. As the OpenAI Turns
The data center takes years to build. The appetite for financing it can change over lunch.
Rich Washburn is a technologist, strategist, and Founder & Chief AI Architect of ARIA AI Labs, working at the intersection of AI, infrastructure, communications, and capital. He also serves as Managing Partner and Chief AI Officer at Eliakim Capital.
Comments