The prediction I'd love to see you stress-test: which 2025 "AI-native" Series A and B rounds get quietly repriced in 2026 H2? The Q1 barbell funded roughly five platforms with $242B and left the remaining ~1,500 AI-native startups fishing in single-digit billions. Half of those cohorts are heading into next-round meetings underwriting GPT-6-tier growth and will meet investors re-underwriting to Q4 2023 SaaS comps. Are any of the GPs in your network already running that conversation with their portfolios, or is it still on the "next quarter" list?
Your pivot away from annual 'State of AI' reports to predictions makes sense - the landscape shifts too fast for static snapshots. I've been taking a similar approach with monthly analysis rather than trying to be comprehensive. One thing I've noticed tracking the agent space is that the companies making the most noise aren't always the ones shipping the most useful stuff. There's a weird inverse correlation between marketing volume and production deployment in my experience. How are you validating your predictions over time? Are you keeping a scorecard or just qualitative tracking?
great essay, set of ideas and predictions
The prediction I'd love to see you stress-test: which 2025 "AI-native" Series A and B rounds get quietly repriced in 2026 H2? The Q1 barbell funded roughly five platforms with $242B and left the remaining ~1,500 AI-native startups fishing in single-digit billions. Half of those cohorts are heading into next-round meetings underwriting GPT-6-tier growth and will meet investors re-underwriting to Q4 2023 SaaS comps. Are any of the GPs in your network already running that conversation with their portfolios, or is it still on the "next quarter" list?
Your pivot away from annual 'State of AI' reports to predictions makes sense - the landscape shifts too fast for static snapshots. I've been taking a similar approach with monthly analysis rather than trying to be comprehensive. One thing I've noticed tracking the agent space is that the companies making the most noise aren't always the ones shipping the most useful stuff. There's a weird inverse correlation between marketing volume and production deployment in my experience. How are you validating your predictions over time? Are you keeping a scorecard or just qualitative tracking?
As someone building in this space, I found the framing more valuable than the forecasts. The perspective is what sticks.