There's no one-size-fits-all strategy when it comes to venture capital investments in the booming artificial intelligence sector.

State of play: Large language model companies are still making headlines for raising piles of cash at eye-popping valuations, but it also appears a growing number of VCs are slinging some of their checks in other directions.

The big picture: At Upfront Ventures' annual confab in Los Angeles this week, a number of top VCs took the stage to share their thoughts on investing in AI.

Zoom in:

• General Catalyst: "We've decided not to invest too much at the [foundational] language model layer, to focus much more on … how do these models interface with different verticals? So what are the modern vertical [software as a service] companies gonna look like, and what are the modern vertical business function companies gonna look like," explained CEO Hemant Taneja.
• Menlo Ventures: "In the near term, it's that it's more about picks and shovels. So really the surrounding infrastructure that helps everyone build at the application layer," said partner Matt Murphy. "You have to be really careful about getting into a crowded category … [and] things that feel like a bolt-on … like if generative AI didn't exist, this application shouldn't be built."
• Lightspeed Ventures: Partner Nicole Quinn said that "on the application layer, it generally starts with entertainment, companionship, escapism.…And then once it reaches critical mass at that point, you then move to the more utility sides of applications … an AI financial manager, or an AI interior designer or a travel planner … and so when we're sifting through, that's when we get excited." She and fellow partner Mike Mignano later warned that big corporate investors don't have the same incentives and investment models as traditional VCs, making those funding rounds and their valuations tricky.
• Founders Fund: "Anything not AI," partner Brian Singerman replied when asked about contrarian areas worth investing in. He compared the current wave of self-proclaimed AI startups to the pre-IPO internet companies of the dotcom era. He cautioned that many startups today aren't really AI companies, but rather are just using the tech.

The bottom line: As in any technologic boom, some will dive in to plant their flag in what they see as the companies that will control how the industry evolves, while others will focus on clever applications once the froth dissipates.