The author is a Reuters Breakingviews columnist. The opinions expressed are her own.

By Karen Kwok

NEW YORK, May 11 (Reuters Breakingviews) - OpenAI and Anthropic, which are already threatening to upend multiple industries, have now separately teamed up with outside investors to target consulting powerhouses. The potentially lucrative opportunity rests on a corporate structure with a long history of disappointing investors: the joint venture.

AI ventures are eager to turn cutting-edge models into durable revenue from business customers, but many companies are still fumbling for effective applications. This disconnect explains why Anthropic last week formed a new company with Blackstone BX.N, Goldman Sachs GS.N and Hellman & Friedman injecting $1.5 billion, people familiar with the matter told Breakingviews. Arch-rival OpenAI struck a similar but larger $4 billion arrangement with 19 investors, including Brookfield and TPG.

The market is enormous. Spending on AI services is on track to reach almost $600 billion this year, nearly a quarter of total AI outlays. There's plenty of scope for new ideas, too. Traditional information technology advisers depend on billing by the hour for armies of programmers, work that AI threatens to automate. Even a diversified giant like Accenture ACN.N derives roughly a third of its revenue from tech consulting and implementing applications.

The pitch from AI firms rests on scarce technical talent. They'll seek to expand their workforces to embed engineers who can re-design how businesses use AI in their work and build custom tools. Once systems are up and running, they move on, repeating the playbook across technology, accounting, and customer-service operations.

Buyout shops benefit by steering businesses they own toward these ventures, making them more efficient, while giving the AI labs a captive customer base. If the partnerships are successful, the investors can cash out through an initial public offering or sale to another company.

Although attractive in theory, history is less encouraging. Roughly half of corporate alliances fail, Boston Consulting Group reckons. Disputes over control, economics, and strategy are familiar pitfalls. Tiffany & Co. and Swatch ended up in court after their 2007 partnership collapsed, and a 2009 Volkswagen-Suzuki alliance dissolved within two years amid mutual mistrust.

Outsiders also may be more comfortable hiring a neutral AI adviser. Established consultants also will defend their turf, while other startups like Mistral and Cohere offer overlapping capabilities. Furthermore, with large-language models evolving rapidly, customers increasingly prefer a more flexible, multi-model approach rather than being locked into Anthropic or OpenAI. For all the ways AI will permeate industry, it will have a harder time solving the sticky problem of shared ownership.

Follow Karen Kwok on LinkedIn and X.

CONTEXT NEWS

OpenAI said on May 11 that it was starting a consulting business, called OpenAI Deployment, with more than $4 billion of investment from 19 partners including TPG, Brookfield Asset Management, Goldman Sachs, SoftBank, McKinsey and Capgemini.

Anthropic said on May 4 that it is partnering with Blackstone, Hellman & Friedman and Goldman Sachs to form a company focused on helping businesses leverage its AI software.