Two rival AI labs. Same strategy. Same day. This isn’t a coincidence — it’s a declaration of war on a $700 billion industry.

On May 4th, Anthropic announced a $1.5 billion joint venture with Blackstone, Hellman & Friedman, and Goldman Sachs to embed its engineers and Claude directly into enterprise operations. Hours earlier, Bloomberg reported OpenAI is finalizing a nearly identical play: a $10 billion venture called “The Development Company,” backed by TPG, Brookfield, Bain Capital, and 16 other investors.

The message is unmistakable: the future of AI revenue doesn’t look like software subscriptions. It looks like consulting — rebuilt from the model up.

The $6-to-$1 Problem

Here’s the number that explains everything. For every dollar companies spend on software, they spend six dollars on services to implement it. That ratio has made McKinsey, Accenture, and Deloitte enormously wealthy. It’s also made enterprise AI adoption painfully slow.

The bottleneck isn’t the models — those have been impressively capable for over a year. The bottleneck is the scarcity of engineers who can actually deploy frontier AI at speed. Blackstone’s President Jon Gray said it plainly: the venture aims to break down “one of the most significant bottlenecks to enterprise AI adoption.”

Sequoia partner Julien Bek went further in a recent essay: the world’s next great company won’t sell software at all. It will sell outcomes — legal services, financial analysis, insurance processing — delivered by AI but billed like consulting.

These ventures are that thesis with a $11.5 billion checkbook attached.

How It Works: Palantir’s Model, Supercharged

Anthropic’s venture mirrors what Palantir pioneered with “forward-deployed engineers” — highly skilled technical staff who embed directly within client organizations. Except instead of Palantir’s analytics platform, these engineers bring Claude and Anthropic’s entire AI stack.

The anchor partners are each contributing roughly $300 million, with Goldman adding about $150 million. The rest of the cap table reads like a private equity who’s-who: Apollo, General Atlantic, GIC, Sequoia, Leonard Green.

OpenAI’s “Development Company” runs the same playbook at larger scale — $4 billion raised from 19 investors against a $10 billion valuation. The investor lists don’t overlap, effectively splitting Wall Street’s PE universe into two competing AI camps.

The Built-In Client Pipeline

The genius isn’t just the money — it’s the distribution. Blackstone alone manages over $1 trillion in assets across hundreds of portfolio companies. Goldman, Apollo, General Atlantic, and the rest collectively control thousands more. Every one of those companies is a potential customer, and the investors have every incentive to push adoption.

Fortune reported last November that 85% of PE buyers now factor AI capabilities into company valuations. Firms that fail to integrate AI risk being penalized at exit. So when your PE sponsor says “we’ve got an Anthropic team ready to deploy Claude across your operations,” that’s not a suggestion — it’s a mandate dressed as an opportunity.

Why McKinsey Should Be Losing Sleep

Traditional consultants sell expertise and frameworks. They send MBA graduates who interview stakeholders, build slide decks, and recommend strategies. Implementation falls to yet another firm. It’s expensive, slow, and the results are mixed.

The Anthropic model flips this entirely. Instead of consultants who understand your business and then recommend AI tools, you get AI engineers who own the model and build solutions directly into your workflows. As Anthropic described one scenario: an engineering team sits down with clinicians and IT staff to build tools that fit workflows people already use.

Goldman’s Marc Nachmann said the venture would “democratize access to forward-deployed engineers” for companies that can’t afford the talent to build AI systems independently. Translation: we’re going to undercut the Big Three on price while delivering more capable solutions.

The Regulatory Wildcard

The timing gets interesting. On the same day these ventures launched, the New York Times reported the Trump administration is now considering pre-release government vetting of AI models — a sharp reversal from its hands-off stance.

The catalyst? Anthropic’s own Mythos model, which cybersecurity experts say can identify and exploit zero-day vulnerabilities in every major operating system and browser. Anthropic flagged this themselves, noting Mythos crosses a threshold where AI can “surpass all but the most skilled humans at finding and exploiting software vulnerabilities.”

So Anthropic is simultaneously asking Wall Street to bet $1.5 billion on deploying its AI across thousands of companies and telling the government its most powerful models pose unprecedented cybersecurity risks. Both things can be true — Claude and Mythos are different models — but it illustrates the tightrope these companies are walking.

What Happens Next

The next 12 months will test whether AI labs can actually deliver embedded enterprise transformation at scale. My bet: the ventures succeed in PE-backed companies where sponsors can force adoption, and struggle in the open market against Accenture’s 750,000 employees and McKinsey’s century of brand equity.

But the bigger picture is already clear. The era of AI labs as pure technology companies is over. Anthropic and OpenAI are becoming services companies — and they’re bringing $11.5 billion in fresh capital to prove it.

The consulting wars have officially begun.