Private Equity’s AI Strategy: Outsourcing to OpenAI, According to Lee McCabe

In a recent LinkedIn post, Lee McCabe discusses a significant shift in private equity’s approach to artificial intelligence, highlighting recent deals where firms are effectively outsourcing their AI strategies to providers like OpenAI, rather than building in-house capabilities as previously claimed.

McCabe points to the recent closing of “The Deployment Company,” a $10 billion vehicle anchored by TPG and including major players like Brookfield, Advent, and Bain Capital. He notes the unique structure of this deal, which guarantees OpenAI a 17.5% annual return for five years.

“The structure is what’s interesting. OpenAI guarantees the consortium 17.5% per annum for five years. OpenAI engineers get embedded inside portfolio companies. PE pays for the privilege.”

The “Value Creation Theatre” of AI Integration

Lee McCabe argues that this model represents a departure from the narrative private equity firms have been presenting to their limited partners (LPs). For years, PE firms have asserted their ability to develop and implement AI strategies internally within their portfolio companies. However, McCabe suggests that the reality is often different.

He elaborates on a pattern he has observed:

“I’ve watched this pattern for years. A firm raises a tech-enabled value creation fund. Hires three Operating Partners with consulting CVs. Builds a deck called something like the AI Acceleration Programme. Two years in, nothing material has changed in the portfolio. So you import a vendor, slap a partnership banner on the LPAC slide, and call the next vintage data-driven.”

This approach, according to McCabe, is less about genuine AI integration and more about creating a compelling story for future fundraising efforts, a phenomenon he terms “value creation theatre.” He points out that the economic implications of such guaranteed returns on services revenue are often overlooked by the industry.

Broader Industry Trends and the AI Gap

McCabe extends his analysis to include other major players, mentioning Anthropic’s similar deals with Blackstone, Hellman & Friedman, and Goldman Sachs. He highlights that both OpenAI and Anthropic have essentially adopted a service model similar to Palantir’s, which was quickly embraced by PE buyers.

He further emphasizes a critical gap he and his colleagues at Claymore Partners frequently observe:

“Something we keep seeing at Claymore Partners is the gap between the AI line in the value creation plan and the person who actually owns a P&L for it. Almost nobody has one.”

This lack of clear ownership and accountability for AI initiatives within portfolio companies underscores McCabe’s critique. He suggests that while PE firms may be acquiring a narrative of AI prowess, the tangible results and P&L impact are often deferred or uncertain, serving the immediate need for a convincing story in investor pitches rather than driving fundamental business transformation.