Investments in generative AI startups—those developing AI-powered products to generate text, audio, video, and more—usually are not slowing down. But they’re consolidating right into a shrinking variety of early-stage startups.
In the primary half of 2023, from January to July 16, 225 startups raised $12.3 billion from VCs, in line with Crunchbase data shared with TechCrunch. If the trend continues, generative AI corporations are on the right track to around 21.8 billion US dollars they increased in 2023.
The total for the primary half of 2024 is broken down into phases:
- 198 Angel/Seed deals: $500 million
- 39 early-stage deals: $8.7 billion
- 18 late-stage deals: $3.1 billion
The clear winners have been early-stage startups reminiscent of Elon Musk's xAI (which raised $6 billion in May), China's Moonshot AI ($1 billion in February), Mistral AI ($502.6 million in June), Glean ($203.2 million in February) and Cognition ($175 million in April). According to Chris Metinko, analyst and senior reporter at Crunchbase, investors appear to be betting on large startups that they imagine have a high probability of success, while letting those they’re less sure about “wither away” within the early stages.
“Some enterprise capitalists expect that the legal and regulatory dilemmas that AI corporations may face within the U.S. and abroad will result in a decline in AI funding,” Metinko told TechCrunch. “Others point to the undeniable fact that the largest winners of the basic infrastructure layer within the mobile revolution greater than a decade ago were established technology corporations.”
Metinko points out that the fate of many generative AI corporations – even the best-funded ones – is uncertain.
Generative AI models are typically trained using data reminiscent of images and text taken from public web pages. The corporations claim that the fair use provision protects them from legal challenges if that data seems to be copyrighted. However, it is just not yet clear whether the courts will ultimately rule in favor of the generative AI corporations. This is probably going why some have begun to enter into licensing deals with copyright holders.
Regardless of the end result of a legal case, it’s becoming increasingly difficult to acquire high-quality training data. costlier as startups exhaust the net's offerings and more developers select to forestall crawlers from scraping their data. (A evaluation estimates that the marketplace for AI training data will grow from $2.5 billion to $30 billion inside a decade.) The means of modeling is just not getting any easier or cheaper either: According to a recent study by Stanford reportThe training cost for OpenAI's GPT-4 was $78 million, while the associated fee for Google Gemini was $191 million.
Given the numerous initial investment required to construct flagship models, it is just not surprising that few generative AI startups are profitable—not even big names like OpenAI and Anthropic. According to The Information, OpenAI, which is claimed to be a turnover of around 3.4 billion US dollars in sales, could lose $5 billion this 12 months.
Investors in generative AI are apparently playing the long game – especially big tech investors like Google, Amazon and Nvidia, who view investments in generative AI as strategic bets. But could the bubble soon burst? If generative AI startups cannot overcome the existential challenges they face, this looks like an actual possibility.
