The private equity playbook just got rewritten: build your own models or become a data source for someone else's.
The Summary
- EQT CEO Per Franzen says scale is now mandatory in private markets as AI transforms how deals get sourced, priced, and managed
- Firms without proprietary AI infrastructure will become data contributors to competitors who do
- The new competitive moat: internal AI expertise combined with strategic external partnerships
The Signal
EQT manages over $250 billion across private equity, infrastructure, and real estate. When its CEO tells the Milken Conference that scale is now essential, he's not philosophizing. He's describing the new table stakes for surviving in private markets.
The logic is straightforward. AI agents need data to learn. In private equity, that data lives in deal flow, portfolio company operations, and exit histories. The more deals you see, the better your models get at spotting mispriced assets, operational inefficiencies, and exit windows. Smaller firms with thinner datasets can't train models that compete.
"Firms without AI infrastructure will become data sources for competitors who do."
But Franzen isn't pitching a build-everything-yourself strategy. EQT is investing in AI infrastructure while simultaneously forming external partnerships. This hybrid approach matters because private equity firms aren't technology companies. They can hire data scientists and build proprietary tools for deal sourcing and portfolio management. They can't compete with Anthropic or OpenAI on foundation models.
The middle path: build internal systems for competitive advantage in areas like due diligence automation, portfolio company performance tracking, and market signal detection. Partner externally for the underlying AI capabilities those systems require.
Key strategic shifts this creates:
- Small and mid-market PE firms face a scale problem they can't solve with better networks alone
- Data becomes a moat, which means firms that share portfolio data in consortiums may be funding their own obsolescence
- The "operator PE" model gets turbocharged when agents handle routine portfolio company optimizations
This isn't about replacing deal professionals with chatbots. It's about creating a structural advantage in pattern recognition. When your agents have seen 10,000 SaaS acquisition scenarios and your competitor's have seen 300, you price risk differently. You move faster. You see opportunities in datasets that look like noise to others.
The Implication
If you're running a PE firm under $5 billion AUM, you have two moves. Partner with a larger platform that will share AI infrastructure in exchange for your deal flow data. Or accept that your edge has to come from hyper-specialization in sectors where relationship depth still beats algorithmic pattern matching.
For everyone else, watch where EQT and peers deploy capital in AI infrastructure companies over the next 18 months. They're not buying exposure to the AI economy. They're buying the tools that determine who survives the reshaping of private markets.