Venture debt firms sound first notes of caution about AI

“AI is obviously the hottest sector,” says Joe Morrison, partner at Barnes & Thornburg. “The valuations are crazy. The exit opportunities are fabulous. Everybody’s a potential customer.”

According to business and technology insight company Gartner, worldwide spending in AI is set to top USD 2tn in 2026.

The sector is almost perfectly positioned for venture debt funding. This is an industry that burns through cash quickly and when it needs to scale, it needs to scale fast. Financing is needed to pay for data centres, electricity, research and development, and attracting the top talent.

ChatGPT’s parent company is still technically loss-making, though this hasn’t discouraged the lenders. JPMorgan Chase, Citi, Goldman Sachs, Morgan Stanley, Santander, Wells Fargo, SMBC, UBS and HSBC have offered a collective USD 4bn credit line to OpenAI, at a rate of SOFR plus 100bps.

This is exactly the kind of business model that venture debt seeks. However, not everyone is convinced. “We still do not do a lot of AI operating companies,” says Trinity Capital CEO Kyle Brown. “There’s USD 100bn or so invested this year into AI companies, and we are looking at that saying, ‘I have no idea who the winners are going to be in that space or if these products are really going to survive.’”

In recent months there has been concern around the exposure that both private lenders and listed companies have to the sector. The five largest listed companies in the world — Nvidia, Microsoft, Apple, Amazon and Google — all have significant investments in AI. The same is true for the world’s largest venture debt companies.

But Sonya Iovieno, head of venture and growth banking at HSBC Innovation Banking, says that capital is still flowing into companies driving AI.

While machine learning has been trialled and financed for decades, it was the release of OpenAI’s ChatGPT in November 2022 which seems to have triggered a massive global boom. But as money continues to flood into the sector, economists have begun to warn of the possibility of a market correction. Microsoft co-founder Bill Gates has even compared the rapid rise of AI to the dot-com bubble, which famously burst in the early ’00s.

For the venture debt firms entwined with the sector, a burst bubble could cause reputational damage. They have long taken pride in their ability to maintain low loss rates and deliver competitive yields, but these metrics only work if the underlying investments are performing.

“A lot of the venture debt calculus is cashflow prospects for your next financing round and prospects for your exit,” says Morrison. “And regardless of whether or not they’re profitable or they have decent free cashflow metrics, a lot of the venture debt is covenant-lite as well.”

Venture debt firms were among the first backers of AI, and most of these firms maintain a stake in the sector to this day. At the same time, companies across the financial services sector are aware of the technology’s impact on their non-AI investments.

“How these companies are using AI, how these companies can be impacted positively or negatively from AI, has been a part of our underwriting thesis,” he said. “It’s certainly something we’re watching closely and our credit teams are doing a great job at monitoring that.”

It seems a sense of caution is emerging. As Bluestein points out, disciplined underwriting is key. But the ability to ignore scaremongering headlines about a theoretical bubble that may — or may not — be about to burst, could be just as important.