news

over the past three years, the ipo market has been almost stagnant, leaving venture capital firms (vcs) in a difficult situation. although the valuations of generative ai startups have soared in the private market and are even regarded as "epoch-making companies",however, it will be difficult for venture capital firms to make profits from these ai companies in the short term.。

the reason behind it,one is that there is an oversupply of funds in the private equity market, and ai companies are not in a hurry to go public to raise funds.unlike previous technology booms, the core of this ai boom is not venture capital companies, but rathermicrosoft, amazon, google parent alphabet and nvidiasuch tech giants are driving ai development, investing billions of dollars to support capital-intensive companies such asOpenAI、Anthropic、Scale AI和CoreWeavethese giants not only have ample funds, but also provide practical benefits such as cloud service points and business cooperation.these resources are difficult for traditional venture capital firms to match.。

second, even if there is an opportunity to go public, these ai startups currently haveprofitabilityit is still far from meeting the expectations of public investors. to successfully go public, companies usually need to prove that they have a stable profit model, but these ai startups are still in the development stage and are still some distance away from being able to meet the needs of public investors.

the reason why venture capital is difficult to make a profit: excess funds in the private equity market

according to melissa incera, an analyst at s&p global market intelligence, ai startups are not having trouble raising money. instead, they are overfunded, and many companies have too many investors actively looking for them.

in 2024, investors have poured $26.8 billion into 498 generative ai deals, continuing a growth trend from 2023, when generative ai companies raised $25.9 billion.this is an increase of more than 200% compared to 2022.。

according to forge global, ai’s share of all fundraising has increased from12% in 2023 to 27% in 2024in addition, the single-round financing amount of ai companies is higher than last year.increased by 140%, while funding for non-ai companies grew by only 10%.

in this context,venture capital firms are facing a serious imbalance in the marketaccording to a report by pitchbook, the u.s.venture capital exit valueit is expected that only$98 billion, down from 202186%, while the number of venture-backed ipos is expected to reachthe lowest level since 2016。

chip hazard, co-founder of flybridge capital partners, noted that investment funds are shifting to “higher-level technologies,” meaning that the focus of investment is shifting from large capital-intensiveinfrastructure ai enterprisesmoved toapplication layer, that is, the company that develops the specific application.

in early 2022,fed raises rates as inflation surges, prompting investors to move away from high-risk assets and toward conservative investments that can provide stable returns in a high-interest rate environment. this market shift has caused a big impact on startups and venture capital firms, as funds tend to flow to low-risk investments rather than early-stage venture investments.

although the nasdaq hit a new high in july 2023, driven by a rebound in technology stocks such as nvidia,ipos and high-priced acquisitions are still rare，this makes it difficult for venture capital firms to use these methods tolimited partners (lps) bring returns.

PitchBookthis phenomenon was explained in the august report:

“due to the current high interest rate environment,vc managers are having trouble raising more money without significant returnsbecause driven by high interest rates,liquid, low-risk investments“with attractive yields now also available, many investors prefer these conservative investments over risky start-ups.”

the only pure ai company close to going public isCerebras, a chipmaker founded in 2016 and backed by traditional venture capital firms such as benchmark and foundation capital. as a semiconductor company, cerebras does not have the high valuations of ai model developers and other infrastructure companies. its 2021 valuation is as high as$4 billion，after that, as the market declined, the valuation did not increase.

cerebras said at the end of july 2023 thathas secretly submitted listing documents to the u.s. securities and exchange commission (sec), but has not yet publicly released its prospectus.

jeremiah owyang, partner at blitzscaling ventures, said thosebasic model company(such as openai, anthropic) quickly achieved extremely high valuations, which put them into a "completely different level" and far beyond the investment scope of traditional venture capital firms. in other words, the valuations of these companies are too high for venture capital firms to continue to participate in them.

he also pointed out that due to the current market environment,it is difficult for venture capital firms to promise an exit in the short termfor early investors, they may need7 to 12 yearsit takes time to see a return on investment in new projects, and this is only after these companies finallysuccessif these companies are not successful, investors may face the risk of losing money.

vcs’ response: special purpose vehicles (spvs)

venture capital firms like menlo ventures and inovia capital are taking a different approach to ai.

in january 2023, menlo announced that it had participated in a round of financing for anthropic through the special investment vehicle (spv) menlo inflection ai partners. the total amount of this round of financing was us$750 million, and the transaction valued anthropic at more than us$18 billion. since its establishment in 2021, anthropic has been mainly supported by amazon in response to microsoft's large investment in openai. microsoft is also rumored to participate in the next round of financing, which will value openai at more than us$100 billion.

menlo invested in anthropic at a valuation of $4.1 billion in 2023. but in order to continue investing despite the significant increase in valuation, menlo had to raise additional funds by creating a special purpose vehicle (spv). through an spv, venture capital firms usually raise funds from limited partners (lps) specifically for a specific investment rather than diversifying investments into multiple companies. menlo plans to raise $500 million for this spv.

in july, cohere, another startup focused on providing generative ai for enterprises, announced the completion of a $500 million financing round, with investors including amd, salesforce, oracle and nvidia. this financing brought cohere's valuation to $5.5 billion, more than doubled from last year. cohere confirmed that this round of financing and some previous financing were also conducted through spvs and organized by inovia.

an spv is a financial instrument that allows investors to pool funds into a specific project. in addition to venture capital firms, jpmorgan chase also said that clients have organized spvs through its morgan private ventures division to help clients participate in several leading ai investments.

however, if investors want to get returns, they still need the company to go public (ipo) in the end, because the current regulatory environment hardly allows large technology companies to conduct large-scale mergers and acquisitions. at the same time, companies like microsoft, google's parent company alphabet, amazon and nvidia are not in a hurry to cash out. they have a total of $280 billion in cash and securities on their balance sheets and can wait patiently for investment returns.

vcs are optimistic about the application layer of generative ai

despite the challenges of going public, venture capital firms are still confident about the future of generative ai, especially its potential at the application layer. looking back at history, similar situations have occurred in almost every major technology cycle. applications such as amazon, google, and facebook are all developed based on internet infrastructure, while successful companies such as uber, airbnb, and snap have risen on smartphone platforms. venture capital firms believe that generative ai will also bring huge returns at the application layer.

gartner analyst john-david lovelock, who has 35 years of experience in the it industry, believes that generative ai has great potential in the enterprise. however, he said that in 2024, enterprise spending on generative ai products will only account for 1% of the total software spending (us$1 trillion).

he also mentioned that some companies are currently spending money on generative ai tools and a few existing applications, but large-scale integration of generative ai into enterprise software products has not yet happened. in other words, although some companies are already using generative ai, the overall promotion is far from being launched.