AI boom draws funding from tech giants, distorting traditional venture capital

Three years on from a near hiatus in the IPO cycle, venture capitalists are in a tough spot.

The private market is dotted with highly valued artificial intelligence startups, including some that are called first-generation companies, but ventures in need of an exit strategy aren’t likely to get any relief from AI anytime soon.

That’s because, unlike previous tech booms, this one isn’t centered around venture capital. Rather, it’s the industry’s biggest companies. Microsoft, Amazon, alphabet and NVIDIA — is pumping billions of dollars into fueling the growth of capital-intensive companies like OpenAI, Anthropic, Scale AI, and CoreWeave.

As the most well-capitalized companies on the planet loosen their purse strings to fund the generative AI boom, the usual pressures to go public don’t apply — and even if they did, these startups are nowhere near demonstrating the profitability metrics that public investors need to see before investing.

Tech giants have more than just capital: They also offer tangible benefits like crowd credits and business partnerships, packaging the types of incentives that venture capital firms can’t offer.

“The AI ​​startups we’ve spoken to have no problem raising capital at high valuations,” Melissa Inserra, an analyst at S&P Global Market Intelligence, told CNBC. “Many companies are reporting too much unwanted investor interest right now.”

Put all this together and venture capitalists are trying to navigate serious market distortions with no clear end in sight. According to an Aug. 29 report from PitchBook, U.S. venture capital exit values ​​are expected to hit $98 billion this year, down 86% from 2021, and venture capital-backed IPOs are expected to be the lowest since 2016. Traditional venture capitalists have been aggressively trying to get into AI, but they’ve mostly invested at the top of the so-called stack, putting money into emerging startups building applications with far less capital than the infrastructure businesses that underpin generative AI.

According to PitchBook, investors, including investments from strategic investors, have pumped $26.8 billion into 498 generative AI deals so far in 2024. This continues a trend from 2023, when generative AI companies raised $25.9 billion for the full year, and is up more than 200% from 2022.

AI’s share of total funding has surged to 27% so far this year, up from 12% in 2023, according to Forge Global, which tracks private market deals. The data shows that average funding for AI companies is up 140% this year compared to last year, while non-AI companies are up just 10%.

Chip Hazard, co-founder of early-stage firm Flybridge Capital Partners, said investment money is shifting “up the stack” and that “lasting companies will be built at the application layer.”

All of this will take time to materialize. In the meantime, startup investors continue to suffer the effects of the market volatility that began in early 2022 when a spike in inflation prompted the Federal Reserve to raise interest rates, pushing investors out of riskier assets and into more conservative investments that could ultimately yield.

Technology stocks have since recovered, led by other mega-cap names such as Nvidia, whose chips are used to train most AI models, and Microsoft. Meta And then there’s Amazon, which has since retreated slightly after hitting an all-time high on the Nasdaq in July, but initial public offerings and big-ticket acquisitions have been few and far between, and venture capital firms have given minimal returns to their limited partners.

“High interest rates have made it difficult for managers to raise additional capital without delivering LP returns, especially as more liquid, lower-risk investments now offer attractive yields,” Pitchbook said in an August report.

The only pure-play AI company that appears close to going public is chipmaker Celebrus, which was founded in 2016 and is backed by traditional venture capital firms such as Benchmark and Foundation Capital. As a semiconductor company, Celebrus has never reached the lofty valuation of AI model developers and other infrastructure companies, peaking at $4 billion in 2021 before the market fell.

Celebrus announced in late July that it confidentially filed documents for its initial public offering with the SEC. The company has not yet filed a prospectus. A Celebrus spokesman declined to comment.

As for the foundational model companies, the astronomical valuations they quickly garnered put them in a “whole different league” outside of the venture capital realm, said Jeremiah Owyang, a partner at Blitzscaling Ventures.

“Given the state of the market, it’s very difficult for venture capitalists to commit to exiting right now,” Owyang said, adding that early-stage investors might not see a return on their new investments for seven to 12 years — and that’s for companies that ultimately succeed.

Firms like Menlo Ventures and Inovia Capital are taking a different path in AI.

In January, Menlo revealed it was raising a special purpose vehicle (SPV) called Menlo Inflection AI Partners as part of a $750 million funding round for Anthropic. The deal valued Anthropic at more than $18 billion. Since Anthropic’s launch in 2021, Amazon has been the company’s main backer, pumping billions into OpenAI, as it tries to catch up with Microsoft, which is reportedly participating in the next funding round that values ​​the ChatGPT creator at more than $100 billion.

Menlo had invested in Anthropic in 2023 at a valuation of about $4.1 billion. To put more money into the bank at a higher price, Menlo would have had to put money outside its $1.35 billion flagship fund that closed last year. In an SPV raise, venture firms typically ask LPs to put money into a separate fund focused on a specific investment, rather than a portfolio of companies. Menlo has filed to raise $500 million for the SPV.

In July, rival startup Cohere, which focuses on generative AI for the enterprise, announced a $500 million funding round from investors including AMD. Salesforce, Oracle And NVIDIA now values ​​the company at $5.5 billion, more than doubling its valuation from last year.

Cohere confirmed to CNBC that part of this round, and some of its previous rounds, were raised through special purpose vehicles. Montreal-based Innovia organized the latest SPV. Shopify CEO Tobias Lutke was one of the participants.

Representatives for Menlo and Innovia did not respond to requests for comment.

Some investment banks are forming SPVs to allow multiple investors to pool capital in popular companies. JPMorgan Chase He told CNBC that clients have had “access to some major AI investments” through the bank’s Morgan Private Ventures division.

Still, an IPO will have to happen at some point for investors to profit, since the regulatory environment makes it virtually impossible for big tech companies to make big acquisitions organically. And companies like Microsoft, Alphabet, Amazon and Nvidia have plenty of patience for investments. These companies have a combined $280 billion in cash and marketable securities on their balance sheets.

Another potential way to ensure liquidity is through the secondary market, where shares are sold to other investors.

Elon Musk’s SpaceX, whose recent employee tender offer reportedly valued the company at more than $200 billion, has opened up investors to shares through secondary trading. This may ultimately be what’s in store for some investors in xAI, Musk’s 18-month-old AI startup, which raised $6 billion in funding in May and is already valued at $24 billion.

But SpaceX is an exception: For the most part, secondary deals are viewed as a way for founders and early investors to cash out some of their stake in a highly valued company, not as a way for VCs to generate revenue — for which they need an IPO.

Michael Harris, global head of capital markets at the NYSE, recently told CNBC that the NYSE is in talks with “a number of AI-focused companies” and that “we expect that pipeline to continue to build as the industry evolves.”

A select few AI companies have entered the stock market this year. Astera LabThe company, which sells data center connectivity to cloud and AI infrastructure companies, listed on the Nasdaq in March. Its valuation has fallen to about $6.5 billion from $9.5 billion on its first day of trading.

Tempus AI, a Google-backed medical diagnostics company, went public in June and its shares have risen about 50% since then, giving it a market capitalization of $8.6 billion.

But the door to an IPO has never opened, and prominent AI companies aren’t even talking about going public.

“Unless market sentiment shifts dramatically, it’s hard to see why an AI startup that can continue to grow privately on such favorable terms would seek public attention,” said S&P’s Inserra. Going public would “only increase pressure to grow revenue or reduce spending, which is not a feasible ask for many companies at this point in their maturity curve,” he said.

Many venture capitalists are optimistic that generative AI will eventually yield big benefits at the application layer, as has happened in every other notable technology cycle. Amazon, Google, Facebook These are all web applications built on top of the internet infrastructure. Uber, Airbnb and snap These are just a few of the many valuable apps built on the smartphone platform.

Gartner analyst John David Lovelock, who has had a 35-year career in tech, sees big opportunities for generative AI in the enterprise, but he said that by 2024, enterprises will spend just 1% of the $1 trillion spent on software on generative AI products.

“There’s been funding put into specific GenAI tools and some of the applications that exist,” Lovelock says, “but there hasn’t been any large-scale deployment of GenAI within a broader enterprise software product catalog yet.”