OpenAI and Anthropic Strike at Consulting in May 2026

TL;DR

On May 4, 2026, OpenAI and Anthropic each announced separate private-equity-backed enterprise AI services ventures on the same day. OpenAI's "The Deployment Company" is a $10 billion vehicle anchored by TPG with 19 investors and a 17.5% guaranteed annual return for backers. Anthropic's unnamed firm — formed with Blackstone, Hellman & Friedman, and Goldman Sachs — is backed by roughly $1.5 billion in committed capital. Both ventures copy Palantir's forward-deployed engineer model and target the $375 billion global management consulting market.

What You'll Learn

In this post you will learn what The Deployment Company and Anthropic's new firm actually are, how the two ventures are structured and funded, why both labs picked private equity as their distribution channel, what the 17.5% return guarantee at OpenAI signals, how this targets McKinsey, BCG, Bain, and Accenture, and what it tells us about where the AI industry thinks revenue is going next.

The Twin Announcements

Within hours of each other on May 4, 2026, two of the leading frontier AI labs revealed near-identical structures aimed at the same prize: turning their models into deployed enterprise systems faster than the market can absorb them.

Anthropic's announcement came through a Business Wire press release naming Blackstone, Hellman & Friedman, and Goldman Sachs as founding partners in a new AI-native enterprise services firm.¹ OpenAI's was reported by Bloomberg the same day, with no formal company press release — only confirmation from people familiar with the deal that "The Deployment Company" had been finalized and that Brad Lightcap, OpenAI's longtime COO who recently shifted into a "special projects" role reporting directly to Sam Altman, would oversee it.²

The pattern repeats: an AI lab takes super-voting control or otherwise anchors the entity, a coalition of large alternative asset managers commits capital, and the resulting entity sells AI deployment services into the asset managers' portfolio companies. The PE firms get a captive distribution channel and a return profile that looks more like infrastructure than venture; the labs get a path to embed engineers inside thousands of operating businesses without building a 100,000-person services arm themselves.

OpenAI's "The Deployment Company": By the Numbers

The Deployment Company is a Delaware-domiciled joint venture that will be majority-owned and controlled by OpenAI through super-voting shares.² The headline numbers, as reported by Bloomberg and the Financial Times:

The 17.5% return guarantee is the most unusual element. Venture-style structures rarely promise a return floor at all, and certainly not one that high — it pushes the deal closer to an income-oriented infrastructure investment with equity-like upside. That structure tells you something about how OpenAI is pricing the risk: the company is willing to backstop the downside in exchange for guaranteed access to a captive customer base inside TPG, Brookfield, Bain, Advent, and the rest of the consortium's portfolios.

Anthropic's New Firm: By the Numbers

Anthropic's venture is smaller, more diversified across founding partners, and — for now — unnamed publicly. The Business Wire release describes it as "a new AI-native enterprise services firm" that will work with mid-size businesses to bring Claude into their core operations.¹

Notice what is missing from Anthropic's structure: no publicly disclosed return guarantee, no headline "vehicle valuation" comparable to OpenAI's $10 billion figure, and no named operating leader from inside Anthropic. The firm is also explicitly described as "standalone" with Anthropic resources embedded — closer to a partnership model than the OpenAI super-voting structure.

The two designs reflect each lab's broader position. OpenAI just closed a $122 billion funding round at an $852 billion post-money valuation on March 31, 2026 and is generating roughly $2 billion in monthly revenue with enterprise making up more than 40% of that mix.⁷ Anthropic, in talks for a roughly $50 billion round at a valuation in the $850 billion to $900 billion range as of late April 2026, hit $30 billion in annualized revenue in early April after a $30 billion Series G round closed at a $380 billion valuation in February 2026.⁸

The Forward-Deployed Engineer Playbook

Both ventures explicitly model themselves on Palantir's forward-deployed engineering pattern, in which an engineer is embedded with a single client to build production-ready workflows on top of the company's platform.⁹ The economic logic is the same in both cases: services are not the primary revenue stream, they are the means to drive product adoption.

The pattern matters because it inverts the traditional enterprise software sales cycle. Instead of selling a license and hoping for renewal once the customer figures out integration, an engineer — paid by the AI lab or its services entity — sits inside the customer's operations, identifies where the model has the most leverage, and ships the integration. The customer's incremental cost is consumption-based usage of the underlying model.

That is the bet OpenAI and Anthropic are both making: that the bottleneck in enterprise AI adoption is not model capability or licensing terms but the absence of engineers who can translate a generic API into a working system inside a specific business. This is the same shift that is putting AI agents at the center of software development workflows — once the model itself is no longer the differentiator, deployment skill becomes the product. The 86% of consulting buyers who actively seek AI-enabled services and the 66% who say they would stop working with firms that fail to incorporate AI suggest the demand is real.¹⁰ The question is who captures it.

Why Private Equity?

The choice of distribution partner is the most strategic part of both deals. PE-owned portfolios offer four properties no other channel can match at the same scale.

The first is reach. Blackstone alone owns or has stakes in hundreds of operating companies across software, healthcare, financial services, logistics, and industrial sectors. Bain Capital, TPG, Brookfield, Advent, Apollo, General Atlantic, and the rest of the named investors collectively touch thousands more.

The second is incentive alignment. PE firms make their returns by improving operating margins at portfolio companies before exit. AI-driven workflow redesign — eliminating headcount, compressing process cycle times, automating reporting — fits that thesis cleanly. The PE owners are already paying consultants to do exactly this work; routing the spend through an AI-native services firm they partly own is a natural substitution.

The third is contract velocity. PE-owned companies have a shared executive sponsor — the PE firm itself — who can mandate adoption across the portfolio in a way that is impossible in the broader enterprise market. A single decision at the GP level can become 50 customer engagements in a quarter.

The fourth is the OpenAI-specific 17.5% return guarantee.⁴ That number is high enough to attract PE capital that would otherwise sit in middle-market buyout funds, and structured so the LPs treat the venture as a yield-bearing asset rather than a venture bet. It is what makes a $10 billion vehicle assemblable in months rather than years.

What This Targets in the Consulting Industry

The global management consulting market was valued at roughly $358 billion in 2025 and is projected to reach $375 billion in 2026, according to Mordor Intelligence.¹¹ The Big Three plus Accenture together account for a meaningful share of that revenue. Industry estimates put McKinsey's annual revenue at $15 billion to $16 billion in recent years, broadly flat over the past half-decade. BCG just reported $14.4 billion for 2025, up 7% year over year and its 22nd consecutive year of growth — with about 25% of that revenue, or roughly $3.6 billion, coming from AI work. Bain has been estimated at about $7 billion as of mid-2025.¹²¹³

The growth pocket inside that market is AI specifically. Accenture's fiscal 2025 results — $69.67 billion in revenue, 7% growth, $2.7 billion in generative-AI revenue (tripled year over year), and $5.9 billion in genAI bookings — show how lucrative AI engagements have become for the largest player.¹⁴ BCG's AI revenue accelerated from a small base to roughly $2.7 billion (about 20% of revenue) in 2024, then to roughly $3.6 billion (about 25% of revenue) in 2025.¹⁰¹³ McKinsey says about 40% of its projects are now AI-related.¹⁰

That growth mix is why this announcement matters. OpenAI and Anthropic are not attacking the entire $375 billion consulting market. They are attacking the AI-services slice that is currently growing fastest inside it, and doing it with three structural advantages no consultancy can match: lower cost of model access, faster integration cycles using engineers rather than analysts, and direct equity exposure for the PE firms that already pay consulting bills.

This also is not the labs' first move into the consulting channel. In February 2026, OpenAI announced partnerships with McKinsey, BCG, Accenture, and Capgemini to distribute its Frontier AI agent platform.¹⁵ May 4 is not OpenAI replacing those partnerships — it is OpenAI adding a parallel channel that does not require splitting margin with anyone.

What the Big Consultancies Have Already Done

The consultancies are not standing still. Reporting in December 2025 indicated McKinsey was planning to cut roughly 10% of its workforce, primarily in non-client-facing functions, phased over 18 to 24 months — on top of a gradual headcount drop from approximately 45,000 employees in 2022 to about 40,000 by mid-2025.¹⁶ BCG grew its workforce to 33,500 in 2025 and added AI engineers, data scientists, and IT architects alongside its traditional consulting teams; the firm also became the only premium consulting firm certified for ISO/IEC 42001 (the international standard for AI management systems).¹³ Accenture has built one of the largest in-house AI talent pools in the industry, exceeding 85,000 AI and data professionals by March 2026 — already past its previously stated goal of 80,000 by the end of fiscal 2026.¹⁷

The strategy on the consulting side has been to position as the "operating model partner" — the firm that figures out where and how to deploy AI agents inside an organization, rather than the firm that builds the agents themselves. That positioning works as long as the labs are willing to sit one layer below in the stack. The May 4 announcements suggest they are not.

What This Tells Us About Where the AI Industry Is Going

Three things stand out.

The traditional enterprise software sales cycle is now considered too slow. Both labs have concluded that the friction between a $20-per-seat-per-month subscription and a deployed AI workflow inside a business is the bottleneck — and they are willing to back a forward-deployed engineering model with significant capital (OpenAI up to $1.5 billion, Anthropic about $300 million as one of three anchor investors) rather than try to sell faster. That is a major shift from the seat-licensing model that defined the SaaS era.

The AI labs are vertically integrating into services. The unbundling era — model-as-a-platform with a partner ecosystem — is being replaced by a bundled model where the lab owns or controls the engineers who deliver the model. Coexistence with consulting partners will continue, but the labs will increasingly capture the highest-margin engagements directly.

Private equity is now an AI distribution channel. The PE firms are not just providing capital; they are providing customers. That makes the line between enterprise software and operational consulting blurrier than it has ever been, and it makes the labs' relationships with the largest PE firms a strategic asset on par with their compute partnerships. Compare this to the broader shift toward AI-powered enterprise workflows and the trend is consistent: AI is moving deeper into operations, not staying at the surface.

Bottom Line

May 4, 2026 was the day the AI industry stopped pretending its enterprise revenue would arrive through the usual SaaS pipes. OpenAI and Anthropic each put their balance sheets behind the same thesis: that the gating factor for AI adoption is forward-deployed engineering capacity inside operating businesses, that PE-owned portfolio companies are the fastest path to that capacity, and that the existing consulting industry — even the AI-services slice growing fastest within it — is too slow, too expensive, and too disintermediated from the underlying model providers to capture the next wave on its own.

The interesting question now is not whether the consultancies fight back; they will, and Accenture's $5.9 billion in genAI bookings shows they are doing more than survive. The interesting question is whether enterprise AI revenue ends up looking more like SaaS, more like infrastructure, or more like a hybrid where the lab, the PE owner, and the operating company share the same engineer.