Huang’s exit raises fears of a funding gap for AI startups

Nvidia has invested tens of billions of dollars in startups in recent years and took part in 105 funding rounds this year, often as the largest investor, according to PitchBook. “If that direction changes, it’ll be a big blow to so many young companies,” said Maria Davidson, founder of Relational Capital. Dawn Capital partner Shamilla Bankiya said Nvidia had helped carry “almost every AI-native startup from application to neocloud.”

Investors say Huang personally drove many of the company’s startup bets. Umesh Padval, managing partner at Seligman Ventures, said significant investment decisions often came from Huang, whom he has known since serving on the board of Mellanox Technologies. Nvidia acquired the networking chipmaker for approximately $7 billion in 2020. The company’s interim CEO, Colette Kress, is expected to make changes to its startup investing. One investor described her as “much more conservative” and doubted she would commit as much money to early-stage companies.

Nvidia’s portfolio includes neoclouds, inference providers and open-weight models—areas that can be expensive for smaller companies to develop. The company backed CoreWeave while it was private, participated in two Crusoe rounds, including its $1.4 billion Series E, and invested in Mistral AI, Lambda and Black Forest Labs. It also agreed to buy Hugging Face for $13 billion, in a deal expected to close in 2027. Last year, Nvidia put $800 million into Reflection AI’s $2 billion Series B, according to The Wall Street Journal.

The company’s investments helped startups get established, but also reinforced reliance on Nvidia chips and its CUDA software platform, which lets developers run AI models on its GPUs. A pullback could give rival chipmakers and cloud providers an opening. Battery Ventures general partner Dharmesh Thakker said competitors would need a strong product, ecosystem and financial capacity to challenge CUDA. He doubts, however, that Google, Amazon, Microsoft or AMD will put hundreds of millions into pre-revenue infrastructure startups as Nvidia did. Investors expect Nvidia to be more selective, with smaller checks for existing portfolio companies and possible pressure on valuations. Still, few TechCrunch sources believe a slowdown alone will stop the broader AI boom. Former Nvidia employee Amulya Vishwanath said the ecosystem is bigger than one person and that founders and leaders “must be held accountable.”