AI Services M&A: What Do the Top AI Companies Acquire?

Something fundamental shifted in the AI services market in the spring of 2026.

Within roughly six weeks, Anthropic launched a $1.5 billion AI services joint venture with three of the largest financial sponsors in the world, OpenAI launched its own $14 billion OpenAI Deployment Company and bought a 150-person London consulting firm called Tomoro to anchor it, and Accenture closed a roughly £740 million acquisition of Faculty.

All directed at the same prize: the implementation layer that turns frontier models into production enterprise systems. For founders running AI services businesses, the strategic map has been redrawn.

In this note, we analyze the recent acquisitions of top investors, the financial profile of the targets and what that means for the next generation of IT services businesses.

Anthropic Acquires Fractional AI

Anthropic’s approach is instructive. In May 2026, Anthropic partnered with Blackstone, Hellman & Friedman, and Goldman Sachs to launch a new AI-native enterprise services company, backed by approximately $1.5 billion of committed capital, designed to embed Anthropic engineers and Claude models directly into mid-size enterprise operations. It is, in plain terms, Anthropic competing with the Big Four consulting firms and doing so with PE patience capital rather than building a delivery footprint organically.

On May 21, 2026: Anthropic announced its founding acquisition of Fractional AI, a San Francisco-based applied AI services company founded in 2024. Fractional AI had quickly become one of the leading end-to-end AI implementation partners for US enterprises; its team and delivery capabilities will serve as the operational centerpiece of the new venture. Terms were not disclosed.

The symmetry with OpenAI’s move on Tomoro is almost too clean to be coincidence: both labs have now anchored their services plays with a founding acquisition of a small, senior, applied AI services firm.

Two months earlier, in March 2026, Anthropic also had committed $100 million to the Claude Partner Network, a formalisation of the consulting relationships it had been quietly assembling. Accenture, Deloitte, Cognizant and Infosys are anchor partners.

OpenAI Acquires Tomoro

OpenAI took a similar path. On May 11, 2026, it launched the OpenAI Deployment Company, a TPG-anchored vehicle that opened with $4 billion of initial investment from 19 investors, a $14 billion valuation, and a guaranteed 17.5% return structure for capital partners. Its founding acquisition was Tomoro (https://tomoro.ai/), a London-based applied AI consulting and engineering firm of roughly 150 forward-deployed engineers and deployment specialists.

Tomoro was created in 2023 “in alliance with OpenAI,” and its client list reads like the case-study brochure every AI services founder wants: Virgin Atlantic, Supercell, Fidelity International, Tesco, Red Bull, Mattel, the NBA.

What Tomoro’s Numbers Tell Us

Tomoro has not filed a profit and loss account (it qualifies for the UK small companies exemption), but its publicly filed balance sheets at Companies House contain enough information to triangulate.

Net profit for FY26 is the change in retained earnings: £8.2 million, with a corporation tax accrual of £1.6 million implying pre-tax profit of roughly £9.8 million.

Revenue we triangulate three ways. The VAT creditor of £797k represents roughly one quarter of net UK VAT, implying around £18 million of UK VATable revenue annualised at year-end (international services are zero-rated, so total is higher). Trade debtors of £5.89 million at a plausible 60-75 day payment terms would imply £28-36 million revenue, but that likely overstates the full year because debtors reflect the elevated Q4 run-rate. Per-head premium AI consulting in London bills 200-300k per practitioner; at an average headcount of 89, that implies £18-27 million.

The three methods converge on a central estimate of around £20 million in FY26 revenue, with a plausible range of £18-25 million. That is 4-5x year-on-year growth, with a profit margin of north of 40%.

That is the profile OpenAI bought. By the time of the May 2026 transaction, Tomoro had scaled to roughly 150 forward-deployed engineers implying the FY27 run-rate is materially above the FY26 figures, very plausibly £35-50 million.

OpenAI did not disclose a separate price for Tomoro, but the firm is the cornerstone of a $4 billion vehicle valued at $14 billion.

Accenture and the Legacy Integrators

Accenture is the other major buyer founders need to map. Its scale is the starting point: fiscal 2025 revenue of $69.7 billion. The firm executed dozens of acquisitions and actively continues to deploy capital.

The most informative deal is Faculty. Announced January 6, 2026 and closed March 16, 2026, Faculty reported revenue of £41.7 million for the period ending March 2025, with a run-rate of more than £50 million annually. Reported headline price was $1B.

At those figures, the multiple lands at roughly 15x run-rate revenue and well above the 11-13x EV/EBITDA band where public IT consulting peers trade. The 400-person Faculty team joined Accenture and CEO Marc Warner became Accenture’s CTO.

The two earlier European AI services tuck-ins are quieter but instructive. 

Afterwards, Accenture closed 2 other AI deals: Halfspace (Denmark, ~80 practitioners, founded 2015) was acquired in March 2025 and Keepler Data Tech (Spain, ~240 practitioners, founded 2018) was acquired in April 2026.

Inside Halfspace: What Accenture Actually Bought

Halfspace files audited accounts at the Danish Business Authority, and they paint a strikingly different picture to Tomoro. Not hyper-growth and software-like margins, but the steady, profitable, decade-old delivery shop that European integrators are increasingly willing to pay a premium for.

A few things stand out. Halfspace’s gross profit grew ~35% year on year from €6.1m to €8.2m), while headcount grew 51% from 39 to 59. Net profit stayed remarkably steady around €1.3m. Note that the Danish “bruttofortjeneste” line is net revenue minus direct and external costs; assuming external costs of 10-15%, implied FY2024 revenue is roughly €9.5m, or about €160k per head.

Accenture took 100% ownership on 27 February 2025, after which Halfspace ran an 8-month stub period (Jan–Aug 2025) to align with Accenture’s August year-end. So, the small net loss in the stub period is the likely the footprint of post-acquisition integration rather than underlying performance.

The contrast with Tomoro interesting. Tomoro is the unicorn: 2.5 years old, ~£20m of revenue at 40%+ margins, £225k revenue per head. Halfspace is the mainstream comp: 10 years old, €8-9.5m of revenue at 75 senior practitioners across the Nordics.

The other tier-one buyers are running the same playbook. IBM Consulting acquired Hakkoda, a data and AI consultancy spanning the US, Latin America, India, the UK and Europe, in April 2025. Capgemini acquired data specialist Syniti in 2024. Deloitte committed roughly $3 billion, PwC $1 billion (Agent OS), EY $1.4 billion (EY.ai), and KPMG launched Workbench.

Cognizant, Infosys, TCS and Wipro, Microsoft’s anointed “Frontier Firms”, have collectively deployed over 200,000 Copilot licences and are scouting bolt-ons across Europe and India.

What This Means for AI Services Founders

Three things matter for anyone running an AI services business right now.

First, the buyer set has tripled. Twelve months ago, the realistic acquirer pool for an AI services boutique was Accenture and a few of the Big Four. Today it is Accenture, IBM, Capgemini, the Big Four, the Indian Tier-1s, OpenAI’s Deployment Company, and an Anthropic-backed PE joint venture. Buyer diversity directly drives valuation; this is the strongest competitive tension AI services founders will see in the cycle.

Second, Tomoro is the signal, and the multiples are real. A two-and-a-half-year-old London services firm with £20m of revenue and a small, but excellent client logos list became the cornerstone of a $14 billion vehicle. Faculty’s reported £740 million acquisition price implies roughly 15x run-rate revenue.

These are not the multiples a services business trades at in a normal cycle they are the multiples a strategic asset trades at when buyers are racing to consolidate a category. The boutique e-commerce services boom of 2017-2021 lasted about three years before multiples normalised; AI services may follow the same shape, even though the long-term opportunity is real.

Third, the firms being bought look like you. Halfspace built dashboards and pipelines for Nordic energy and banking clients. Keepler wrote cloud-native architecture in Madrid. Faculty did applied data science consulting. Tomoro’s founders are former OpenAI staff who hired senior, kept the team small, picked enterprise logos carefully, and executed.

There is no Nobel-track research and no frontier model. There is a sharp thesis, a credible team, and clients willing to pay for it. If you run a 20-250 person AI services firm with real implementations in production, you look more like an acquisition target than you may realise.

If you are running an OpenAI or Anthropic deployment business and want a candid read on where you sit in this market, we are happy to talk.