Seventy-seven days.

That is the gap between OpenAI's announcement of "Frontier Alliances" — a multiyear partnership with Accenture, BCG, Capgemini, and McKinsey to deploy OpenAI's enterprise platform — and yesterday's launch of the OpenAI Deployment Company, a $14 billion private-equity-backed joint venture that will compete with those same four firms on those same enterprises.

Anthropic announced a parallel structure eight days earlier: a $1.5 billion joint venture with Blackstone, Goldman Sachs, and Hellman & Friedman. Anthropic, Blackstone, and H&F are each putting in roughly $300 million; Goldman is also an investor. The Fortune headline that day: "Anthropic takes shot at consulting industry."

Two foundation-model companies, eight days apart, backed by four of the largest financial-services firms in the world, launched dedicated entities to compete with the consultancies they had spent the previous twelve months partnering with.

The Number

The OpenAI Deployment Company is being capitalized with $4 billion in fresh investment against a $10 billion pre-money valuation — a $14 billion post-money — to operate with roughly 150 specialists, the headcount of the AI consulting firm Tomoro that OpenAI acquired as the new entity's operating core.

OpenAI Deployment Company post-money valuation
Specialists acquired with Tomoro

$14 billion for 150 specialists is $93 million per head.

Accenture's market cap is roughly $200 billion against about 800,000 employees — $250,000 per head, fully loaded. Take only the 80,000 staff Accenture committed to its AI practice in 2023, and treat them as the relevant comparison: at the per-head ratio implied by the OpenAI Deployment Company, that practice would be worth $7.4 trillion. The number is absurd. The point of citing it is that the public-market price of an AI consultant and the private-market price of an AI consultant are no longer in the same order of magnitude.

The asset being priced inside the Deployment Company is not the 150 specialists. The asset is OpenAI's position as the only vendor that can deploy GPT-class models without paying a license fee — a position no consultancy can replicate at any headcount.

The Calendar

Frontier Alliances launched on February 23.

The first reporting that OpenAI was in talks with private-equity firms — TPG and Advent — to form a joint venture appeared in Reuters on March 17. Twenty-two days after announcing the consulting partnership, OpenAI was already raising the money to compete with it.

Bloomberg confirmed the $10 billion pre-money figure on May 4, the same week Anthropic finalized its Wall Street JV. The OpenAI Deployment Company launched seven days later.

The partnership and the competing entity were not sequential phases. They were parallel tracks. Frontier Alliances was the public-facing distribution strategy while the PE-backed competitor was capitalized in private. The consultancies signed multiyear deals with a counterparty that was already raising $4 billion to replace them.

What the Consultancies Were For

The case for AI consulting in 2024 and 2025 was straightforward. Enterprises had to deploy generative AI, the deployment was complex, and the model providers — small companies measured against Accenture's 800,000 — lacked the headcount, the methodology, and the procurement footprint to do enterprise integration at scale. The natural division of labor was: model providers ship models, consultants ship implementations.

That division produced real revenue. Accenture's three-year deal with Anthropic in December 2025 made the consultancy one of Anthropic's top three enterprise customers and committed Accenture to training 30,000 of its staff on Claude. Frontier Alliances three months later did the equivalent for OpenAI. Model providers had distribution. Consultancies had revenue. Boards on both sides nodded along.

The division of labor lasted 77 days as an articulated OpenAI strategy. What changed was not the consultancies' execution. What changed was the calculation that the deployment layer is where the margin lives — and the foundation-model companies decided, with PE money explicitly authorized for the purpose, that they would keep that margin themselves.

Frontier Alliances was the public distribution strategy while the PE-backed competitor was capitalized in private.

The Coverage Shift

Coverage of Accenture in the tech press averaged one to two articles per quarter from 2023 through mid-2025 — the baseline of a large enterprise IT vendor that mostly stayed out of the news. In Q1 2026, it was seven, in the same quarter that included Frontier Alliances and the news that Accenture was tying promotions to weekly AI tool logins. The firm became, briefly, the AI deployment story.

That kind of regime change in coverage usually reflects a regime change in market position. In this case it does — but not in the direction the coverage implied. Accenture was not becoming the AI deployment company. It was being briefly cast as the AI deployment company in the months between the model providers' "we need partners" position and the model providers' "we have $5.5 billion in PE money and a joint venture" position. The narrative arrived and reverted inside a single quarter.

Accenture's stock underperformance through 2025 — paired with successive rounds of layoffs blamed on AI skill mismatches — suggests the market priced the regime change long before the press named it. The press was the lagging indicator. The PE round was the leading one.

Seventy-Seven Days

"Multiyear" was the qualifier in OpenAI's February press release. Seventy-seven days was the actual term. The consultancies were the distribution layer of last quarter. The PE-backed joint ventures launched this week are the distribution layer of the next one. Goldman, Blackstone, Hellman & Friedman, TPG, and Advent decided which.

Seventy-seven days is a quarter, less three weeks. It is shorter than the onboarding cycle for a single enterprise Frontier Alliances client. The partnership did not fail. It did not have time to.