In the world of venture capital, the pace of artificial intelligence is forcing investors to look at their holdings with fresh urgency. Eric Acher, co-founder of the Brazilian firm Monashees, warned that AI is changing the "half-life of a thesis" and that deals that seemed solid in 2024 could appear outdated by 2026.
Why portfolio reviews matter now
According to Kamran Ansari, founder of Kapital Ventures and venture partner at Infinity VC, only businesses that hold regulated licences or own proprietary data enjoy any real insulation. "Absent those two things, everything else feels exposed," he told a recent audience.
Lily Lyman, managing partner at Underscore VC, explained that the traditional binary view of performance, on track or off track, no longer applies. "Portfolio reviews right now aren't just a question of 'are you performing?' or 'are you not performing?' It's more like: what track are you on, and what does that mean?"
Four lanes for AI-focused startups
Lyman identified four categories that investors are currently sorting through:
- Startups betting on the soaring AI consensus.
- Efficient-growth companies that are scaling without heavy AI reliance.
- Firms seeking product-market fit in traditionally tough sectors.
- Companies caught in the crosshairs of large model providers such as OpenAI and Anthropic.
In the fourth bucket, the market has shifted so dramatically that a startup's original value proposition may no longer hold. "The question becomes: do you have something that's valuable? If so, is it people, a product, or distribution? What do you lean into in order to try to recoup value?" Lyman said.
Real-world examples of rapid change
One high-profile case is Perplexity, a search-engine startup that surged in popularity before Google introduced a comparable AI-driven product. "Perplexity was molten-lava-hot, but I don't think it's that special anymore because Google caught up extraordinarily fast," Ansari observed.
Zachary Aarons, co-founder of MetaProp, added that software firms must "agentify" their platforms or risk extinction, especially in mission-critical fields like construction where reliance on generic foundation models is insufficient.
How much of the portfolio is at risk?
During a EuroHerald Brainstorm Tech session, private-equity veteran Robert F. Smith noted that a small but real slice of his software holdings no longer have a viable future in the AI-driven environment. Ansari estimated that in his own portfolio, an extra 10-20% of companies are especially vulnerable beyond the normal venture-capital attrition rate. An anonymous early-stage VC echoed a similar figure of roughly 10%.
"As an investor, I'd welcome more companies saying: 'you know what, we're 18 months in, we still have a bunch of cash, it's not going to work,'" Ansari suggested. "Why don't I return 60 or 70 cents on the dollar, and when I have something new, I'll come back to you to raise money again?"
He argued that a partial return of capital leaves a better taste than draining a startup's resources to zero and then delivering a final, bleak email.
What comes next for VCs?
Investors are being urged to take stock now, because the next wave of AI breakthroughs will again reshape the landscape. The advice is clear: conduct thorough portfolio reviews, identify which companies own defensible data or licences, and consider graceful exits for those that cannot adapt.
In the words of the classic film that opened this discussion, if you don't pause to look around, you could miss the next big shift.