The End of Consulting as We Know It

The AI Companies Have Decided to Become Consulting Firms

On May 4 — the same day, almost the same hour — the two biggest companies in AI announced enormous deals. Both are joint ventures with Wall Street private equity. Both have the same goal: installing their AI directly into the hundreds of portfolio companies those firms own.

OpenAI unveiled the $10 billion 'Deployment CompanyThe Deployment Company​'; Anthropic announced a $1.5 billion joint venture of its own. The two announcements landed minutes apart.

This is not just financing news. It marks a change in the very identity of what an AI company sells.

How the Two Deals Differ

First, the facts.

On Anthropic's side. Anthropic, Blackstone, and Hellman & Friedman are each putting in roughly $300 million. Goldman Sachs joins as a founding investor with a $150 million commitment. A16z, General Atlantic, Leonard Green, Apollo Global Management, Singapore's sovereign wealth fund GIC, and Sequoia Capital round out the $1.5 billion total. 

On OpenAI's side. Nineteen investors, $10 billion in total. TPG anchors the deal, joined by Brookfield Asset Management, Advent International, Bain Capital, and Goanna Capital. The private equity consortium will invest roughly $4 billion over five years, while OpenAI retains strategic control through super-voting shares. 

The most striking difference is the terms OpenAI offered. OpenAI has guaranteed its private equity backers an annual return of 17.5% over five years. What makes that strange is that guaranteed returns are almost unheard of in venture investing. \"This guaranteed return floor is abnormal by ordinary venture standards. Private equity vehicles don't typically receive explicit annual return promises from an operating partner, and OpenAI is not in the habit of writing structurally subordinated paper.\" 

Put simply, OpenAI has attached something close to a bond guarantee to this deal. If it succeeds, the private equity firms take more of the upside; if it fails, OpenAI covers the floor. Anthropic made no such guarantee.

Why Private Equity?

Both companies arrived at the same conclusion: conventional enterprise sales cannot spread AI fast enough.

\"Both companies concluded that traditional enterprise software sales cycles are too slow to capture the next wave of AI adoption. Both are betting that private equity firms — which hold hundreds of operating companies and have the structural ability to mandate adoption across their portfolios — are the most efficient distribution channel.\" 

The logic of the bet is simple. Selling Claude or ChatGPT into companies one at a time takes too long. Private equity is different. Blackstone owns more than 230 companies; KKR, more than 200; Carlyle, more than 220. When a private equity firm says \"adopt,\" its portfolio companies adopt. A single deal opens a pipeline for installing AI across hundreds of companies at once.

AI's Consulting Turn

The real shock of these deals is not the money — it's the business model. What the two joint ventures will actually do is consulting.

The Deployment Company's operating model resembles Palantir's. It dispatches engineers inside client organizations to implement AI systems using OpenAI's tools, automate workflows, and restructure operations. Instead of selling software licenses and leaving implementation to the client, the Deployment Company delivers technical expertise directly at the point of transformation. 

Anthropic is taking the same approach. \"The new company is an independent entity with Anthropic engineering resources embedded directly inside teams. The structure mimics Palantir's forward-deployment model, combining model ownership with implementation capability to undercut traditional consulting firms.\" 

The meaning is unmistakable. AI companies are no longer content to build and sell models — they intend to be the consulting firms that wire those models into their customers' operations. It is a declaration that they will do the work McKinsey, Accenture, and BCG used to do.

A remark from the Goldman Sachs side captures the intent precisely: the venture will \"democratize access to forward-deployed engineers\" for companies that can't afford the talent or the consulting fees to build AI systems on their own. 

The biggest casualties of this shift are the incumbent consultancies. The deal puts the Claude maker in direct competition with the world's largest consulting firms in the lucrative market for enterprise AI transformation. 

The old arrangement looked like this: AI companies built the models, and consulting firms applied them inside client companies. They were partners in a division of labor. OpenAI maintained 'Frontier Alliances​Frontier Alliances​' with BCG, McKinsey, Accenture, and Capgemini. Anthropic had similar partnerships.

These joint ventures break that division of labor. If AI companies do the consulting themselves, what is left for the consulting firms? Can a consultancy that cannot build models compete with a model company's joint venture?

The signal has already arrived. \"After a pilot lifted productivity, Baldwin Insurance Group expanded Claude across the entire company and became the first named client of the $1.5 billion joint venture.\" An insurer became a direct customer of Anthropic's joint venture — with no consulting firm in between.

OpenAI vs. Anthropic: Two Different Bets

The two companies are pointed in the same direction, but their bets differ in meaningful ways.

OpenAI is going bigger ($10 billion), with a more aggressive financial structure (the 17.5% guaranteed return) and a tighter focus on private equity portfolios. Anthropic is smaller ($1.5 billion), anchor-investor-led, with no guaranteed return.

\"OpenAI's structure is larger in absolute capital, more aggressively financialized, and more concentrated on private equity portfolios. Anthropic's is smaller and anchor-investor-driven.\" 

The difference reflects where each company stands. 

\"As of early 2026, Anthropic's Claude is winning roughly 70% of new-business matchups against OpenAI in the professional segment.\" 

Anthropic already holds the lead in the enterprise market, so a smaller deal suffices. OpenAI has ground to make up, which is why it needed bigger, more aggressive terms. The 17.5% guarantee is the mark of that chase.

Three Things This Means

First, AI's business model is shifting from selling to embedding. Selling a license and leaving implementation to the customer is too slow. AI companies now go inside the customer's walls to build the systems, redesign the operations, and stand up the automation. Revenue comes from transformation services, not license fees.

Second, private equity becomes the core channel for AI distribution. Private equity firms are no longer mere investors — they have become distribution infrastructure. The thousands of companies held by giants like Blackstone, KKR, and Carlyle become targets for AI adoption all at once. \"Private equity firms collectively own or control hundreds of operating companies across healthcare, logistics, manufacturing, and financial services.\" 

Third, the final pre-IPO housekeeping is underway. \"Ahead of its IPO, Anthropic announced a $1.5 billion joint venture with Blackstone and Goldman Sachs.\" Both companies are preparing to go public, clarifying their revenue structures and locking down enterprise channels along the way. The joint ventures are part of that cleanup.

What This Means for Korean Companies

This shift may be playing out between American Big Tech and Wall Street, but it bears directly on Korean companies.

AI adoption is accelerating on a global scale. Once private equity portfolio companies start rolling out Claude and ChatGPT at speed, Korean companies in the same industries will have to keep pace. When a competitor cuts operating costs by 30% with AI, standing still is not an option.

The definition of the consulting market is changing. Korea has a sizable consulting market for digital transformation and AI adoption, and the top of that market is being restructured. When model companies do the consulting themselves, the gap widens between those who own models and those who don't.

The criteria for choosing an AI partner are changing too. What companies need is no longer a simple right to use a model, but a partner that can come into their operations and build the systems. The selection criterion becomes not just model performance, but the model company's implementation capability.

Not the End of Consulting — Its Redefinition

One thing should not be misread. These deals are not killing the consulting industry. They are redefining what consulting means.

Anyone can do the former. Only companies that own models can do the latter. So the model companies take the top layer of consulting, while incumbent consultancies get pushed down into change management, reorganization, and human resources.

This is what the restructuring of an industry's value chain looks like. And at the apex of that restructuring, two AI companies announced deals pointing in the same direction on the same day.

May 4 may well prove to be an inflection point for the AI industry: the day AI moved from tool to infrastructure, from selling models to integrating operations — and the day it became explicit that private equity is the channel for that move.