A few months ago, partners at Bessemer Venture Partners gathered together at a partner offsite event in Montana and discussed one big topic: artificial intelligence.
“The number one thing that was on every partner's mind in each of their respective categories...from health care to cyber to frontier tech, was how A.I. is impacting their areas,” partner Talia Goldberg told me yesterday. It was an unusual moment of unanimity within the VC firm, she says: “Very rarely is it…so clear and so obvious across our firm that this is so hugely important,” she says. It “became almost a no-brainer decision for us internally.”
As a result, the firm has decided to go big: Bessemer is directing $1 billion of its deployable capital to invest in A.I., the firm told me—a hefty sum that Goldberg says is Bessemer putting out a signal to founders that they're serious about A.I., they have cash for it, and they’re ready to spend it. The $1 billion is not a new fund, but a commitment from current capital, which they will invest across stages, Goldberg said. She added that their fund structure gives them the flexibility to invest “even beyond a billion dollars in this area,” adding that LPs are excited. (Bessemer doesn’t have a target of how many startups they want to invest in, she says.)
The reasoning, and the message, is simple: “A.I. is here, it is massive,” Goldberg says. She notes that the speed at which A.I. is and will affect virtually every sector the firm is investing in is exciting, and also cited things like model performance improvements and falling costs as reasons for Bessemer’s timing. “We expect that over the next many years, you're going to see totally new industries emerge and transform based on A.I.—and so there's a whole lot of good reason for all the attention that it's getting,” she later added.
In general, it’s pricey to invest in A.I. I recently wrote about how valuations are soaring, while startups like OpenAI are reportedly burning lots of cash training their models. But Goldberg points out that while A.I. is expensive, “it's also getting cheaper, and it is getting more and more accessible. We think that this [announcement] really reflects the magnitude of the potential opportunities for A.I.-native companies and new categories, specifically,” she says, referring to companies where A.I. is at their core versus those that use A.I. as a feature or to alter existing products. “We are seeing growth and adoption at rates that we've just never seen before.”
It’s not surprising Bessemer wants to telegraph loud and clear that it’s got a ton of cash to funnel into A.I.: While the firm has been making bets on the space for some time, including on the likes of buzzy A.I. copywriting startup Jasper, competition seems to be heating up. Bessemer reportedly recently beat out other VCs to lead a Series B round in EvenUp, an A.I. startup that helps personal injury lawyers compile claims. (Goldberg said via email that “nothing in that report has been solidified or publicly confirmed.”)
For her part, Goldberg says she’s excited to invest in areas where A.I. is changing search and information discovery beyond just classic Google queries, and also in the way humans and A.I. interact (for example, sales reps using A.I. to be more efficient, or even people interacting with A.I. for therapy or friendship—an application she argues is less hyped). She notes that Bessemer partners who focus on other areas like health care and cyber are also looking into how A.I. will impact those spaces.
One question I have here is, if even engineers at Google believe the company and other A.I. model developers like OpenAI don’t have a competitive moat with open source models gaining traction, how are VCs thinking about their prospective investments’ ability to win? Goldberg says the recently leaked Google engineer document sparked internal discussions at Bessemer, but that it highlighted her view: “A model alone is not a product,” she says. What matters is “being able to bring it all together and to build a product that people really want to use. Because where do moats come from? They come from also having the ability to continually improve your model...to get distribution and adoption for it, to build the business,” as well as having things like a good go-to-market strategy, she added.
That’s part of the checklist Bessemer is looking for: Those companies that are making “really great products that are solving problems or creating new categories that have durable business value, and can become category leaders over time,” Goldberg says.
To be sure, lots of VCs are hopping on the A.I. bandwagon as hype in the space has reached fever-pitch levels. But not all of them have $1 billion to do it.
See you tomorrow,
Anne Sraders
Twitter: @AnneSraders
Email: anne.sraders@fortune.com
Submit a deal for the Term Sheet newsletter here.
Jackson Fordyce curated the deals section of today’s newsletter.
