Where Active AI Investor Felicis Is Placing Its Bets

Editor’s note: This article is part of an ongoing series in which Crunchbase News interviews active investors in artificial intelligence. Read previous interviews with Battery Ventures, General Catalyst, Bessemer Venture Partners, Accel, Insight Partners, Index Ventures, Sequoia Capital, Section 32, M12 and Sapphire Ventures as well as highlights from these stories from 2023.

Felicis was one of a clutch of seed funds founded in the mid-2000s to write smaller checks to a new cohort of startups that were started after the dotcom bust. Now two decades later, the firm is still making seed investments, but pulling from deeper pockets to fund later-stage rounds, too — and this time it’s competing to invest in the next generation of AI companies.

The San Francisco Bay Area-based firm was founded in 2006 by Aydin Senkut and raised its first external fund, of $41 million, in 2010. It’s grown significantly since then, with the announcement of its ninth fund in March 2023, an $825 million fund.

Sundeep Peechu, general partner at Felicis

We recently spoke with general partner Sundeep Peechu, an engineer by training who joined the firm in 2010 as it raised its first external fund, about where Felicis is investing in artificial intelligence these days.

“We think of AI as a sequence of capabilities that have been slowly building over a period of time,” he said.

The firm was an early investor in design tool Canva and productivity tool Notion, both of which have integrated AI for their large customer bases. It was also an early investor in Cruise, founded in 2013 and one of the pioneers in autonomous vehicle technology.

Felicis likes to invest in companies that it can hold on to for a decade-plus, Peechu said. “It cannot be a flash in the pan that basically solves a problem now, but that problem disappears because the sands have shifted three years away.”

Felicis AI investments since 2022

Since 2022, the firm has invested in foundation model companies Runway, for AI video creation, and Poolside, for building AI for software developers.

The firm is also an investor in open-source database Supabase and Weights & Biases, which is used by AI teams for model oversight.

Other recent investments include defense tech startup Vannevar Labs, which sells cutting-edge technology to intelligence agencies.

Two portfolio companies with early traction are Deep Infra and Flower Labs, both of which assist companies with training on specific data sets using large language models.

Peechu sees a shift to a focus on investing in applications utilizing this technology where “there’s a lot of services work that might be semi-automatable and with a human at the very end, basically adding a level of oversight,” he said.

He anticipates multimodal AI is coming next. That technology integrates text, image, sound and video or some combination of those into one interface.

Peechu notes we’ve seen false dawns before in areas such as natural language processing or computer vision that showed rapid progress and then reached a plateau before the next discovery.

“But in five years, I’d be shocked if people say ‘I’m investing in AI,’ ” he said. “It is just going to be the foundational layer of almost everything.”