OpenAI is offering private equity firms a guaranteed minimum return of 17.5% and early access to unreleased models as it races to outbid Anthropic for joint venture partnerships with major buyout firms, according to a Reuters exclusive published Monday citing multiple sources familiar with the talks.
Both companies are courting private equity firms to form joint ventures that would push their AI tools into hundreds of portfolio companies simultaneously, creating customer lock-in at a scale neither company could achieve through direct sales alone. OpenAI is in advanced talks with TPG, Advent International, Bain Capital, and Brookfield Asset Management to raise approximately $4 billion at a pre-money valuation of roughly $10 billion for its venture. Anthropic has been courting Blackstone, Hellman & Friedman, and Permira for a parallel structure.
The key difference: OpenAI’s deal includes the 17.5% guaranteed floor, seniority over other joint venture partners, and downside protection. Anthropic’s enterprise-focused deal offered no such guaranteed returns, Reuters reported.
Why PE Firms and Why Now
The joint venture structure solves a specific financial problem for both companies. Customizing AI models for enterprise clients requires expensive upfront engineering work, including deploying dedicated teams to integrate models into existing systems. By pushing those costs into a separate joint venture entity, OpenAI and Anthropic can keep their core balance sheets cleaner ahead of expected IPOs, potentially as early as this year.
“There’s a big race to lock in as much enterprise, as many desks as possible,” Matt Kropp at Boston Consulting Group’s AI unit told Reuters. Once a company has a customized AI model integrated into its workflows, switching costs become prohibitive.
The ventures would generate revenue through implementation service fees, revenue sharing on deployed products, and co-ownership of new products built for portfolio companies.
Not Every Firm Is Buying
At least two major PE firms walked away from both deals. Thoma Bravo, one of the world’s largest software-focused buyout shops, declined after managing partner Orlando Bravo questioned the long-term profit profile. Many of Thoma Bravo’s portfolio companies are already using AI tools without committing capital to a joint venture structure.
Other skeptics argued that large PE firms already have direct access to both OpenAI and Anthropic through standard commercial channels. The joint venture layer adds cost and complexity without necessarily improving access to the underlying technology.
Some investors also noted that with technology valuations down, these partnerships may not generate meaningful additional revenue. The real upside, several sources told Reuters, would depend on securing board seats, equity stakes, or other economic terms available only to lead partners.
How the PE Strategy Reshapes the OpenAI-Anthropic Rivalry
The PE strategy marks a new front in the OpenAI-Anthropic rivalry. Anthropic has historically held stronger positioning among enterprise buyers, particularly in regulated industries. OpenAI’s aggressive sweeteners suggest the company recognizes that gap and is willing to pay a premium to close it.
The timing adds pressure on Anthropic specifically. While OpenAI negotiates PE deals, Anthropic faces a federal court hearing Tuesday over the Pentagon’s designation of the company as an “unacceptable national security risk.” Fighting a legal battle with the Defense Department while simultaneously competing for PE capital creates a two-front resource drain that OpenAI can exploit.
For the broader AI market, the PE joint venture model represents a bet that enterprise AI adoption will follow the same pattern as enterprise software: whoever gets embedded first wins the account for years. The difference is that OpenAI and Anthropic are willing to guarantee returns to make it happen, a level of financial aggression unusual even by Silicon Valley standards. Valley standards.