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 Felicis, Battery Ventures, Bain Capital Ventures, Menlo Ventures, Scale Venture Partners, Costanoa, Citi Ventures, Sierra Ventures and Andrew Ng of AI Fund, as well as highlights from more interviews done in 2023.
Most corporate venture firms steer clear of investing in rivals. But GV (formerly Google Ventures) isn’t most firms. Not only did it back Slack while Google had a competing product, it’s now investing in AI companies that go head-to-head with parent company Alphabet’s own efforts in the artificial intelligence race.
GV’s model — single LP, independent decisions — has allowed it to stay fast and nimble during AI’s most explosive moment. Partners Dave Munichiello and Tom Hulme are backing companies across chips, compilers and applications, and making early and late bets alike. GV’s wager is clear: The AI age rewards those who move fast, trust founders and don’t flinch at eye-watering valuations.
We recently spoke with Munichiello, who leads the digital investment practice in Silicon Valley, and Hulme, based in London, who co-lead the digital investment practice globally, about GV’s AI investment approach.
Munichiello was previously at Kiva Systems, which became Amazon Robotics. Hulme has a background in physics and was the design director at Ideo in Europe and an active angel investor before joining then-Google Ventures in 2014.
A frenzied AI gold rush on the West
“It’s an absolutely wild time on the West Coast, in San Francisco. Every weekend is five or six long walks with founders that are getting inundated with AI term sheets,” said Munichiello. “Founders are in houses filled with other AI founders. The buzz is heavy of people getting poached by the Metas or the OpenAIs of the world, while simultaneously fending off those offers and starting their own thing,” he said.
Munichiello and Hulme said they both have middle- and high-school aged children who are using the technology in their everyday lives, to write, digest and summarize complex ideas.
From widespread consumer adoption, there is a drastic push from enterprises to spend to capture this “AI magic.”
Talent competition heats up
“We think about the stack as low as chips and infrastructure, and as high as the application level, and all the different industries that application-level investments are blossoming in,” said Munichiello.
Early GV investments in AI include Lattice Data, which was bought by Apple to power Siri in 2017, video-generation company Synthesia, and data labeling company Snorkel AI. More recently, it was an investor in Thinking Machines Lab’s record $2 billion seed round at a $12 billion valuation.
GV invests independently of Alphabet, its parent and sole investor. That means it sometimes backs companies that could be seen as competitors to Alphabet-owned products. For example, GV invested in Slack, when Google had a competitive internal product, and Thinking Machines Lab competes with Alphabet’s Gemini.
As to why they invested, Munichiello said they heard from their internal recruiting team that Thinking Machines CEO and co-founder Mira Murati, formerly OpenAI’s CTO, is one of the best recruiters of deeply technical AI talent. “Amazing companies sometimes are at prices that feel very uncomfortable, and that was certainly the case with Thinking Machines,” he added.
Competition for AI research talent is intense, and model performance is quantifiable. That means “for the first time, many of the big companies are competing in the same game. So if you look at Meta, Alphabet or OpenAI, they’re all competing on having cutting-edge foundation models,” Hulme said.
“We’ve never seen companies scale as quickly as the AI-native companies we see today,” he said. “We have not seen talent move as frequently at the foundation model as the amount we’re seeing at the moment.”
The team pays attention to where the smartest researchers are going to work. “They’re not going to Big Tech. They’re going to startups, and they’re all going to San Francisco,” said Munichiello.
App layer
“We’ve been slower to invest in application-layer companies, until we know there’s real traction,” said Munichiello. “If you look at Harvey or OpenEvidence, we didn’t lead the Series A. We led a later-stage round.”
Munichiello said the question they seek to address: “Are these dollars seeking a use case, seeking an application that could be interesting in their company? Or are they locked in production, creating actual value for the company?”
To assess whether revenue is experimental or not, the team spends time understanding customers’ use — is the product beyond proof of concept, is it used daily, does it save millions of dollars in headcount while creating actual value?
In 2024, the firm led the Series C in AI legal tech startup Harvey at a $1.5 billion valuation. The company has since raised two rounds and now has a valuation of $5 billion.
More recently, GV co-led, along with Kleiner Perkins, the Series B in OpenEvidence, an AI-powered platform that helps doctors and other healthcare professionals more easily aggregate and process peer-reviewed medical literature, at a $3.5 billion valuation.
GV also co-led a $40 million Series A investment with New York-based TQ Ventures in London-based online legal platform Lawhive.
“When we look at companies that are coming in to raise, the revenue run rate is insane. These companies are growing incredibly fast, faster than ever before,” said Munichiello. “And it’s very hard to spend a lot of time looking at AI applications companies, and then go back to looking at other companies.”
“We have businesses in our portfolio like Bolt.new [a product of StackBlitz] — they built out the technology and then turned on monetization and went from zero to $40 million ARR in 12 weeks,” said Hulme “We think that’s the new normal.” GV was a co-lead in the startup’s seed round in 2022. The company was not growing meaningfully until October 2024, when it launched bolt.new to code with text prompts in a web browser, which took off.
The invest strategy at the infrastructure level
For the infrastructure layer, the firm’s calculus is different, said Munichiello. “We start to do deep technical work, and see there needs to exist a compiler that is universal, and the company needs a lot of money to get there, and really patient capital. We’re going to lead the first round. We’ll probably lead multiple rounds over time, but this will pay off 100x over time if it’s right — that’s worth deep investment from us.”
To that end, GV led the $30 million seed round in Modular, which seeks to build a unified compute layer to interface with AI hardware. The company has created a universal compiler that is a competitor to Cuda, which enables AMD‘s GPUs to run just as fast as those from Nvidia, said Munichiello. Modular has since raised a $100 million Series B led by General Catalyst.
Multistage, multisector, multigeo investors
GV has been investing for 15 years and has $10 billion in assets under management. The firm is headquartered in Silicon Valley with offices in Cambridge, Massachusetts, New York and London. Its largest exits over time include Uber, Nest Labs, Slack and GitLab, among others.
CEO and managing partner David Krane heads up the team of 21 partners. Alongside Krane, Munichiello and Hulme, the fourth managing partner is Krishna Yeshwant, who is based in Cambridge, and co-leads its life sciences practice with general partner David Schenkein.
Companies meet at least one other partner through the investment process and then usually pitch the entire partnership. Partners share ideas and thoughts with the lead, who then can make an investment decision. The team can move quickly and has been able to close a deal in a week from the first meeting with a founder.
“A and B are the sweet spot,” said Munichiello. “Our job, when we find an area that we’re excited about, is to look at the A and the B. And then to say, ‘Is this the very best way to invest in this category?’”
If the right company is at Series D or seed, the team will do that investment. Team members are not limited by locale, and can invest in the best companies across the world.
Hulme thinks more in terms of cities than countries. “For every great founder you need great operators, maybe five or 10 great operators, and the global pockets of those remain, in my opinion, in San Francisco, New York and London. Those are the cities that we’re most excited about. But we also invest in places like Tel Aviv, which are growing quickly.”
An analog business
“We’re a venture capital business, but we think of this as a human capital business. We’re as good or as bad as the people on the GV team. And interestingly, they are as good or bad or as bad as the humans they’re investing in. It’s a very analog business in that way. It’s just super-powered with technology,” said Hulme.
Munichiello said the firm’s diverse approach and flexibility is key.
“GV is a place where we can have 30 people going in 30 different directions, investing in 30 different things. And we’re not trying to drive consensus,” said Munichiello.
Related Crunchbase list:
Illustration: Dom Guzman
Clarification: This story has changed since its original publication to update the firm’s life sciences practice information.
Stay up to date with recent funding rounds, acquisitions, and more with the
Crunchbase Daily.