VCs abandon old rules for a ‘funky time’ of investing in AI startups

If there’s one factor that VCs agree on when backing AI startups, it’s that AI requires a distinct funding method than prior technological shifts.

“It’s a cool time,” stated Aileen Lee, founder and managing associate of Cowboy Ventures, onstage at TechCrunch Disrupt 2025. The longtime VC famous that the foundations of investing have considerably shifted now that some AI firms are leaping from “zero to $100 million in income in a single 12 months.”

Nonetheless, Lee additionally famous that, based mostly on her agency’s analysis, Sequence A buyers aren’t simply searching for fast income progress. “It’s an algorithm with completely different variables and completely different coefficients.”

Among the elements buyers now measure, in keeping with Lee, embrace whether or not the startup is producing knowledge, the energy of its aggressive moat, the founders’ previous accomplishments, and the technical depth of the product. “Relying on what your organization is, the output of the algorithmic components goes to be completely different,” she stated.

Jon McNeill, co-founder and CEO of startup creation agency DVx Ventures, acknowledged that even startups that develop quickly from inception to $5 million in income typically wrestle to safe follow-on funding. “I believe this sport has modified, and it’s altering dynamically,” he stated.

McNeill famous that Sequence A buyers are actually making use of the identical rigorous requirements to seed-stage startups that they beforehand reserved for extra mature firms.

“I believe a number of buyers have found out that the breakout firms, generally, don’t have one of the best tech,” McNeill stated about why Sequence A VCs are trying so carefully at startups’ potential to draw and retain clients. “They’ve one of the best go-to market.”

Steve Jang, founder and managing associate of Kindred Ventures, disagreed {that a} robust go-to-market (GTM), an trade time period for gross sales and advertising, holds better weight for buyers. “I don’t suppose it’s 100% true to say mediocre expertise, nice GTM wins and raises cash and will get clients. I believe that it’s a essential requirement to have each.”

Whereas McNeill later clarified that having a stable product is necessary, he indicated that his preliminary remark was associated to the founders’ must develop an exceptionally robust gross sales and advertising technique proper out of the gate. “Traders are getting way more refined on the go-to market than they’ve previously,” he stated.

(The talk over advertising versus tech was delivered to the forefront later in the course of the convention when Roy Lee, founding father of the viral startup Cluely, said onstage that launching a product that hardly labored, even with large social media fame, could not all the time be one of the best thought.)

Aileen Lee added that AI startups are actually below stress to ship product updates and new options at an unprecedented tempo, preempting current firms that may attempt to introduce related merchandise.  “In the event you have a look at how a lot OpenAI and Anthropic are delivery, you’re going to have to determine the right way to match how a lot you ship, how rapidly and the standard of it,” she stated.

Regardless of the expectations for breakneck progress and quick product improvement, panelists agreed that the AI trade continues to be in its very early levels. As Jang put it, “There are not any clear, outright winners, even in LLMs. There are rivals nipping at their heels.”

This implies startups nonetheless have a path to unseating perceived leaders, whether or not they’re decades-old firms or fast-moving newcomers.