Insecure Culture (Updated)

I don’t want to write a lengthy post about it but there’s something about the Jensen Huang brouhaha that analysts have not commented on sufficiently. To refresh your memory, the CEO of nVidia, the company on which many LLM AI service providers depend, has been accused of having an affair with a Chinese spy, a so-called “honeypot” operation. It has subsequently been claimed that Mr. Huang has had at least one other affair and nVidia has illegally sold its chips to China.

What I wanted to point out is that whether the women with whom Mr. Huang had affairs were spies or not, indeed whether management were involved in the sale of chips to China or not, there is a culture of insecurity at nVidia, at least with respect to China. That presents risks.

Mr. Huang’s resignation would not be enough to remove those risks. Firing him wouldn’t be enough. Clearly, nVidia requires serious scrutiny by the federal government. That would consist of mandatory security audits, continuous monitoring of sensitive transactions, verification of end users, government approval of particularly sensitive transactions, and consequences for non-compliance. Refusing to submit to those it should suffer those consequences: it should be placed on the list of vendors from which the federal government cannot procure under the terms of the Federal Acquisition Supply Chain Security Act of 2018 (FASCSA) and so should companies that use nVidia chips. That includes Palantir, OpenAI, and Anthropic.

I don’t recommend nationalization. I recommend formal, systematic official scrutiny.

The problem goes beyond that. The U. S. needs more companies that are viable alternatives to nVidia. But that’s the topic of another post.

Update

nVidia has announced that Jensen Huang has resigned. Talk about closing the barn door after the horses have already bolted…