Andrew Ng, AI’s biggest star, has announced that his AI Fund, an incubator for startups that supports small teams of experts who are looking to solve problems with AI, will raise upwards of $120 million in its second attempt.
A filing at the SEC shows the AI Fund’s AI Venture Fund II has so far amassed 69.75 million dollars from 13 partners – leaving around 50 million dollars to be invested. The AI Fund’s public relations declined to comment.
Ng was a well-known name in the AI community before he joined Baidu as chief scientist in 2014. He left Baidu to launch a number AI ventures including DeepLearning.ai and landing AI.
Ng launched AI Fund with $175 million in 2018, serving as the incubator’s GP, and leading its direction. On the SEC filing he is listed as the “managing partner of the general partnership” for AI Venture Fund II. The idea was to fund companies at the seed stage and Series A, allowing them to work in relative secrecy until they were ready and connecting them to Ng’s professional network.
Greylock Partners, New Enterprise Associates and SoftBank Group are among the initial backers of the AI Fund. Crunchbase lists as 38 portfolio companies. These include AI observability platform WhyLabs and Ng’s Landing AI.
AI Venture Fund II, at $120 million would be significantly smaller than the original AI Fund vehicle. It’s still more than twice what Ng initially hoped to raise for the AI Fund, which was $50 million.
Take this as a sign that the AI bubble, and in particular the buzzy generative AI segment of it, may be deflating.
PitchBook recently reported that generative AI deals at the earliest stages have declined for two consecutive quarters. They have fallen 76% from their Q3 2023 peak. The value of pre-seed- and seed-stage deals dropped to $122.9 million in Q1 2024, down from the Q3 high of $517.7 millions.
Enterprise reluctance may be at fault.
In a recent survey from Boston Consulting Group about half of respondents — all C suite executives — said they didn’t expect generative AI would bring about significant productivity gains. They were also concerned about the possibility of mistakes and data breaches arising from tools powered by generative AI. My colleague Ron Miller reported last week that businesses are finding generative AI harder to implement than they thought. And that executives are exercising caution.