By WSJ Staff

What's it like to make big investment bets on artificial intelligence?

Laela Sturdy is managing partner at CapitalG, Alphabet's multibillion-dollar independent growth fund. Founded in 2013 with a mission to provide capital and support to technology companies looking to accelerate their growth, CapitalG under Sturdy has made sizable investments in companies including online payment-processing platform Stripe, cybersecurity firm CrowdStrike and language-learning platform Duolingo.

Sturdy sat down with Sarah Krouse, the technology and media editor at The Wall Street Journal, at the 2025 WSJ Tech Live conference in Napa Valley, Calif., to discuss AI's potential to transform life and work, and the challenges involved in picking AI investments.

Edited excerpts follow.

Strong stomach

WSJ: If you're investing about a billion dollars a year, can you give us a sense of your approach and what your biggest bet is at the moment?

STURDY: We've done a lot in cybersecurity, a lot in data infrastructure. And of course, we've spent much of the last two years really focused on AI and looking at the multitude of investment opportunities both in AI native companies [those designed with AI at their core], as well as how AI will be tailwinds or potentially headwinds for different players across the technology ecosystem.

WSJ: Throw a rock in San Francisco and you'll hit an AI startup. How do you vet them and make sure there's actually room for them to compete and thrive?

STURDY: AI companies are growing almost five times faster than the software companies that came before them, which is good and bad. It's good because it demonstrates the market pull. You can really understand the customer value that's being provided. But things like differentiation and moats [that is, whether a company has a durable, competitive advantage] are a lot more challenging to assess.

WSJ: Do you think about the path to profitability for this generation of AI startups differently than the startups of the software era?

STURDY: We've seen a lot of variability on this front. Some of these AI companies are smaller than we've ever seen because it's easier than ever to build a company, and you can do a lot more with fewer resources. So in some places, we're seeing this massive growth, really profitable business models, smaller teams. In other cases, we're seeing business models that I would call still emerging. Many of the most successful AI companies that we're looking at this year, if you would have looked at them 12 months ago, had negative gross margins. They were burning significant amounts of capital. The investment landscape today requires an ability to underwrite that. You need a strong stomach.

WSJ: What's a big area of investment opportunity in AI right now?

STURDY: There is a lot of interesting stuff happening in autonomy, a lot of interesting stuff happening in robotics. The field of science and drug discovery has enormous, enormous potential, and we're already seeing interesting progress in AI models that are helping identify early drug candidates. If we see AI working in some of these traditionally offline industries that have required huge amounts of capital and long cycles, it unlocks enormous amounts of value.

Impact on work

WSJ: How are you seeing AI being used in the workplace now? Is it replacing jobs? It feels like we've had a wave of white-collar layoffs under the banner of AI. Is that real?

STURDY: I run an investment fund with 50 people, and since we're investing in AI, we have to eat our own dog food. So we have completely reinvented our process of how we do research, how we compare companies, how we discover interesting insights into the data. And the impact of AI is incredible.

We're able to get interesting research reports around industries in all parts of the world and sectors. I say to my team, how can we be AI-first? And I think the companies in our portfolio that are most successful leaning into AI are asking the same thing. You see companies like Stripe have an AI -based fraud model, and a company like Whatnot, which is a livestream shopping platform, build an AI application where sellers can scan their goods and automatically create listings, bringing a bunch more inventory onto the platform.

So the use cases that I'm seeing now are much more about leaning in, creating opportunities, reducing friction, augmenting human labor and making people more and more effective. But we're in early innings. I think we will see sort of some of the longer-term structural changes in the labor market.

WSJ: What is AI bad at? Has it surprised you?

STURDY: Hallucination is a challenge. A lot of judgment is still needed in taking the information that models give. You need to be very careful in sourcing and getting to the underlying data. I find it most effective as a thought partner, not someone I would trust with sort of zero-error type analysis. That's one of the big limitations.

WSJ: Do you worry about the speed at which all of this is changing and growing and the sort of froth around investment and the enthusiasm for what can be?

STURDY: There is more potential for transformation here than we even can contemplate, so I believe the enthusiasm and capital being invested into AI makes a ton of sense. Of course, there will be bumps along the way, and how the distribution of winners ends up remains a hard thing to call. But I think we're way underestimating what the impact of AI will be and how much that will transform our lives in the industries that we work in.

(END) Dow Jones Newswires

November 10, 2025 14:00 ET (19:00 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.