Startup 2025: What AI-Focused VCs Are Looking For

Y Combinator founder Paul Graham advises startup founders to “live in the future, then build what’s missing.”

I had the privilege of glimpsing the future through a series of interviews with investors on the bleeding edge of the AI landscape. Insights from these candid conversations laid the foundation for “Startup 2025: Building a Business in the Age of AI,” the AI startup report that Snowflake is publishing today.

We selected eight investors from across the United States and Europe to speak with, seeking individuals we knew would bring a unique vantage point on the realities, challenges and opportunities of building an AI startup in today’s rapidly evolving world:

• Akash Bajwa, Principal, Earlybird Venture Capital

• Rohini Chakravarthy, Managing Partner, NewBuild Venture Capital

• Patrick Chase, Partner, Redpoint Ventures

• Sakib Dadi, Partner, Stage 2 Capital

• Mats Olov Magdesjö, Operating Partner for Analytics, EQT Ventures

• Liam Mulcahy, Go to Market, Kleiner Perkins

• Shravan Narayen, Partner, IVP

• Sam Teden, Investor, Anthos Capital

I spoke to them in my capacity as director of the Powered by Snowflake Startup Program, which helps give startups the resources they need to build and grow data-intensive applications and products. Most of the investors in this report have worked with our program to nurture early-stage companies, and they gave us their front-line perspectives on a tech revolution.

What we found is a mixed bag. Certain well-worn, universally held, MBA-approved beliefs and concepts have not withstood today’s new climate. For example, annual recurring revenue and net dollar retention, two foundational metrics when it comes to analyzing any software company’s success and trajectory, have a more nuanced value when it comes to evaluating the performance of AI startups. This shift is driven by “AI tourists,” as Redpoint Ventures partner Patrick Chase puts it — folks who are more than happy to try a new AI solution but less likely to remain users than your historical software buyer.

Other traditional beliefs and best practices are more true than ever. Go tackle a large market, make your early adopters wildly successful, and build a competitive moat as you scale. And don’t burn too much cash while you’re at it.

The report is an exercise in extracting signals from the noise, of which there is more than ever, to help arm founders with a coherent strategy on which to build their business.

What I found fascinating was how quickly opinions and beliefs evolved in the larger marketplace over the three months we spent crafting this report. In that time, suspicions of an AI-fueled bubble swelled, and blind optimism around the business value that may one day come from AI morphed into pragmatic realism around the quantifiable business impact AI solutions are driving today.

Both Sam Teden of Anthos Capital and Mats Olov Magdesjö of EQT Ventures voiced an aversion to companies pushing "AI for AI's sake" and talked about how they look for startups with real value propositions — using AI as the best solution to fundamental business challenges.

Of the many hype cycles the tech industry has seen (and helped fuel), this one certainly feels different. And this report has only furthered that belief for me. The opportunity is vast and growing, the winners will win at a scale we’ve never seen, and the short-term lens of peaks and valleys in the space will, in my view, even out into an aggressive hockey stick growth of adoption and business value.

Read the full report to hear eight AI-focused investors talk through their approach to investing, measuring and fostering growth and tackling go-to-market challenges. They also explain exactly what captures their hearts (and checkbooks) in a pitch meeting. And if your startup is building something impactful, AI-centric or not, we’d love to hear from you. Our Powered by Snowflake Startup Program helped hundreds of startups get started in the Snowflake Data Cloud. Reach out to us here.