Models & Research

After Huang: AI startups worry Nvidia's money is drying up

3 min read AI-generated

105 funding rounds this year alone, often as the lead check. Investors say many of those calls were Jensen Huang's personally.

Featured image for "After Huang: AI startups worry Nvidia's money is drying up"

Most coverage since Jensen Huang resigned from Nvidia has been about affairs, spy allegations and the hunt for a successor. On Friday TechCrunch asked VCs and founders a different question: what happens to the startups Nvidia has been funding? Their answer is blunt. The industry just lost one of its biggest checkbooks.

How much Nvidia put in

According to PitchBook, Nvidia took part in 105 funding rounds this year, often as the largest investor. “Nvidia is carrying almost every AI-native startup from application to neocloud,” says Shamillah Bankiya of Dawn Capital. Umesh Padval of Seligman Ventures, who knows Huang from the Mellanox board, says most of the big investment decisions came from Huang himself.

The list is long. Nvidia backed CoreWeave while it was still private, joined Crusoe’s $1.4 billion Series E, holds stakes in Mistral AI, Lambda and Black Forest Labs, and is buying Hugging Face for $13 billion. It wrote an $800 million check into Reflection AI’s Series B before the startup had meaningful revenue.

One VC describes the pattern: infrastructure startups that needed $100 million to $500 million worth of GPUs would often raise half of it from Nvidia at valuations of $1 billion or more. Other investors piled in because Nvidia was already on the cap table.

Kress is seen as more cautious

Several investors describe interim CEO Colette Kress as far more conservative. “I don’t think she’s going to commit the same amount of money to keep the early-stage ecosystem going,” one says. The expectation is that Nvidia gets pickier about new bets and writes smaller checks into its existing portfolio. Dharmesh Thakker of Battery Ventures expects Google, Amazon, Microsoft or AMD to compete for Nvidia’s role, but doubts they’ll put hundreds of millions into pre-revenue infrastructure companies.

Open models lose their champion

Nathan Lambert, who writes about open models at Interconnects, thinks the damage lands hardest there. Open models were one of Huang’s most ambitious founder-style bets, he writes, and that funding will very likely dry up. Inference startups already struggle to get enough chips because Anthropic and OpenAI are buying so much capacity. Huang bought the ecosystem time and was willing to use his clout when open models came under political pressure.

Less competition from below for Anthropic

Directly, little changes for Anthropic: it pays for its own compute and is close to its IPO. Indirectly, the balance shifts. If Nvidia money for neoclouds and open models gets scarcer, Claude faces less cheap competition from below, and the big closed labs with their own access to chips get more entrenched. Lambert says as much: Anthropic currently looks like one of the greatest businesses ever assembled.

Sources

NvidiaJensen HuangStartupsOpen SourceFunding