Nvidia’s rivals won’t win overnight, but Huang’s exit speeds up every custom-chip program

In a bonus episode of WIRED’s Uncanny Valley, editors Brian Barrett, Zoë Schiffer and Leah Feiger worked through the founder’s resignation (our coverage), what interim CEO Colette Kress means for Nvidia, and whether competitors can capitalize. Their answer, roughly: not quickly, but over time, yes.

Schiffer’s point holds. When a social platform stumbles, a rival can ship a competitor quickly. Advanced chips take years of research and manufacturing. In the short run, customers who want to diversify away from Nvidia mostly have nowhere better to go. The hosts described the U.S. economy as heavily tied to a few companies, naming Nvidia and Anthropic, so a leadership shock at the chipmaker ripples into every lab, cloud and startup.

Barrett framed the longer-term story well. Big buyers are already designing their own chips to tune hardware to their software, much as phone makers did after Apple’s processors showed what was possible. A scandal like this turns a side project into a priority, unlocking budget and urgency at any AI lab that has been dabbling. Nvidia itself has hedged by financing custom silicon through MediaTek (our analysis). A market with more than one viable supplier is healthier, more resilient and harder to weaponize in geopolitics. Israeli investors make a similar case about a real but narrow window (coverage).

The hosts saw Kress as a classic CFO and a steady hand who knows the business deeply (profile). Barrett expected business as usual, including continued advocacy for selling to China, and argued she can make that case without her predecessor’s baggage. That is plausible, but the China question is now tied up in congressional inquiries that an interim leader can’t easily steer (Democrats’ letter). The hosts also expect China hawks to use the moment to push for tighter trade rules, a debate that will shape where AI chips can be sold more than any product launch this year.

Barrett noted the founder owns about 3.5% of Nvidia, which is why the board could act at all, unlike companies where founders hold controlling votes. There was also no succession plan and around 60 direct reports. The lack of a bench was a risk shareholders carried for years with little complaint, and the market priced it abruptly (the trillion-dollar slide). The lesson for every AI company, especially labs heading to public markets, is to plan for key-person risk before a crisis. The hosts noted Nvidia policy bars relationships with direct reports; the founder resigned after admitting an affair with an alleged Chinese intelligence asset and amid reports of at least two more alleged relationships with employees. More in our NVIDIA coverage.