Inside EQT's AI Summit 2025
TLDR
- EQT gathered 75+ portfolio tech leaders and concluded that AI success depends on orchestration, culture change, and governance, not model development.
Key Takeaways
- Companies should orchestrate existing AI models rather than train new ones; agents are “doers” to be onboarded and performance-evaluated like employees.
- AI-native orgs win through continuous adaptation and learning speed, not raw scale or compute.
- Safety, evals, and interpretability are framed as competitive moats, not compliance overhead; governance accelerates adoption rather than slowing it.
- Culture and change management outrank tooling; leadership quality determines whether AI investments compound or stall.
- Incumbents are deploying AI for near-term business value while AI-natives race to lock in distribution, proprietary data, and feedback loops.
Why It Matters
- EQT’s framing of agents as employees with onboarding and performance reviews signals a shift in how enterprise AI deployment will be structured and measured.
- The incumbent vs. AI-native race is a concrete strategic fork: incumbents optimizing existing revenue vs. new entrants building data moats that get harder to close over time.
- With 75+ CTOs and CPOs aligned on governance-as-speed, enterprise AI buyers are likely to reward vendors who ship reliable, interpretable systems over those shipping raw capability.
Victor Englesson and William Evensen, EQT Ventures · 2026-04-01 · Read the original
