On May 4, 2026, Anthropic — the San Francisco AI lab behind the Claude model — announced an unusual kind of deal. Together with three Wall Street giants (Blackstone, Hellman & Friedman, and Goldman Sachs), it is launching a new company worth more than $1.5 billion. The twist: this isn't a fundraising round. It's a joint venture whose entire purpose is to embed Anthropic's engineers inside other businesses to help them actually use AI.
The capital structure tells you who matters. Anthropic, Blackstone and Hellman & Friedman are each contributing roughly $300 million. Goldman Sachs and the growth-equity firm General Atlantic are putting in $150 million apiece. A wider consortium — Apollo Global Management, Sequoia Capital, Leonard Green, and Singapore's sovereign wealth fund GIC — rounds out the rest. Blackstone, the world's largest private investment firm with $1.3 trillion in assets, led the talks and is positioned as a 'founding partner.'
Why do this now? The straightforward answer is that Anthropic is preparing for a possible IPO as soon as this year, with reported valuations as high as $900 billion. Its annualised revenue has rocketed from about $9 billion at the end of 2025 to more than $30 billion by late March 2026, much of that driven by coding tools like Claude Code. But running frontier AI is astronomically expensive — the data centres alone require massive capital — and the new venture is meant to generate the kind of recurring revenue that would justify that spending to public investors.
There's a deeper logic, though. Selling AI models is one business; getting companies to actually use them is a different, much harder one. The new firm will operate less like a software vendor and more like a forward-deployed engineering team — closer to the model Palantir popularised. Its engineers will sit inside mid-sized businesses, especially the hundreds of portfolio companies owned by the participating private-equity firms, redesigning workflows in industries like healthcare, manufacturing, financial services, retail and real estate. Goldman's Marc Nachmann argued AI engineers would help companies 'lean forward' and figure things out, while those that don't would 'be left behind.'
Private-equity-owned companies are an ideal launchpad. PE owners already pressure their portfolio firms relentlessly to cut costs and boost productivity, which is exactly what AI vendors promise. Buying a stake in the deployment vehicle gives the PE firms preferred pricing and a financial stake in the rollout they would have funded anyway. For Anthropic, it locks in distribution into hundreds of companies that would otherwise be hard to reach.
It is also a competitive race. OpenAI has launched a near-identical venture called The Deployment Company, backed by TPG, Bain Capital, and Brookfield, that reportedly raised more than $4 billion at a $10 billion valuation. Both labs have concluded the same thing: in the long run, AI revenue may not look like software licensing, but like consulting — rebuilt from the model up.
The quieter story is what this means for the existing consulting industry. McKinsey, Bain and Deloitte have spent a century billing for smart generalists in suits. If the 'advice' is really an algorithm plus an engineer who keeps it tuned, that model is under serious pressure. Watch for traditional consultancies to race to build or buy their own AI deployment arms, and for mid-sized companies — long unable to afford McKinsey-grade automation — to suddenly be able to.
An AI lab worth potentially $900 billion just teamed up with the world's biggest private-equity firms — not to raise money, but to build a consulting business that could eat McKinsey's lunch.
Anthropic, the company behind the Claude AI model, has formed a joint venture worth more than $1.5 billion with Blackstone, Hellman & Friedman, and Goldman Sachs. Anthropic, Blackstone and H&F are each putting in about $300 million, while Goldman and growth-equity firm General Atlantic are committing $150 million each. Other backers include Apollo, Sequoia, Leonard Green, and Singapore's sovereign wealth fund GIC.
The new entity isn't a normal investor — it's effectively a consulting arm. Its engineers will embed inside the private-equity firms' portfolio companies, redesigning how those mid-sized businesses work so they can actually use Claude in their day-to-day operations. The push comes as Anthropic tries to justify its enormous spending on data centres ahead of a possible IPO later this year.
AI models alone don't transform a company — somebody has to rewire the plumbing. This venture is essentially a bet that the bottleneck in AI adoption isn't the technology, it's people who know how to install it.
If you're picking a college major or a first job, this is the structural shift to watch. The 'AI will take your job' headlines are oversimplified — what's actually happening is that the people who can translate between AI models and real business workflows (forward-deployed engineers, implementation specialists, AI product managers) are about to become some of the highest-paid workers on the planet, while traditional consulting roles get squeezed.
For a century, the consulting business has been built on smart generalists in suits charging by the hour. Anthropic and OpenAI are betting that model breaks when the 'advice' is really an algorithm plus an engineer who keeps it tuned. Watch for two second-order effects: traditional consultancies (McKinsey, Bain, Deloitte) racing to buy or build their own AI deployment arms, and a wave of mid-sized companies suddenly able to automate work that only Fortune 500 firms could afford to automate before.