Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Nvidia, the world’s most valuable chipmaker, has become one of the most prolific investors in artificial intelligence start-ups this year, seeking to capitalise on its position as the dominant provider of AI processors.
Silicon Valley-based Nvidia said on Monday it had invested in “more than two dozen” companies this year, from big new AI platforms valued in the billions of dollars to smaller start-ups applying AI to industries such as healthcare or energy.
According to estimates by Dealroom, which tracks venture capital investments, Nvidia participated in 35 deals in 2023, almost six times more than last year.
That made the chipmaker the most active large-scale investor in AI in a banner year for dealmaking in the sector, outstripping Silicon Valley’s largest venture firms such as Andreessen Horowitz and Sequoia, according to Dealroom, excluding small-scale accelerator funds such as Y Combinator that place many smaller bets.
“Broadly, for Nvidia, the number one criteria [for making start-up investments] is relevancy,” Mohamed Siddeek, head of its dedicated venture arm NVentures, told the Financial Times.
“Companies that use our technology, who depend on our technology, who build their businesses on our technology . . . I can’t think of a situation where we’ve invested in a company that did not use Nvidia products.”
Between NVentures and its corporate development team, Nvidia’s portfolio now includes Inflection AI and Cohere, two of the biggest rivals to ChatGPT maker OpenAI.
Other investments are in Hugging Face, a provider of data and tools for AI developers that was valued at $4.5bn in August, and CoreWeave, a cloud infrastructure company focused on high-performance computing applications that rely on chips such as Nvidia’s graphics processing units.
Its most recent investment was in Mistral, the Paris-based AI start-up that was valued at €2bn this month.
The one thing the companies have in common is that they are all Nvidia customers, whether using its GPU chips or its software.
Nvidia’s H100 GPU has become one of the most sought-after products in Silicon Valley this year. The powerful processor helps creators of “large language models” — the underlying platform that powers “generative AI” services such as ChatGPT — to train their systems much more quickly than using traditional server chips.
Nvidia invests from its own balance sheet, writing cheques running to tens of millions of dollars. NVentures looked to “generate healthy returns” from its investments, while its corporate development team could invest for more strategic purposes, Siddeek said. It both leads rounds itself and invests alongside venture capital firms.
Some start-ups hope that accepting investment from Nvidia will offer a closer relationship with the group. “They’re obviously a very strategic partner; you want to be a priority for them as they roll out new chips,” said one VC who shares investments with Nvidia.
Another investor added: “It’s the smartest go-to-market strategy to really lock up the market . . . [Nvidia] know the money will come back eventually.”
Nvidia denied that it sought special terms with start-ups to ensure they used its chips. “We try to be as investee-friendly as possible,” said Siddeek. “We don’t have any conditions per se.”
Inflection AI — which in June announced a $1.3bn fundraising round jointly led by Nvidia, Microsoft, Bill Gates and others — boasts that it has access to 22,000 H100 GPUs, thanks to its alliances with the chipmaker and CoreWeave. Late last year, CoreWeave said it was among the first companies to receive H100 shipments, alongside cloud giants Amazon, Google, Microsoft and Oracle.
Siddeek denied the group’s portfolio companies receive preferential access to its chips. “We don’t help anybody jump the queue,” he said.
Read the full article here