Anthropic and Hg launch a consultancy, and IT services lose the AI rollout layer

Outcome-priced delivery for mid-market buyout funds is the more crowded of the two ideas. Lab-backed ventures from OpenAI and Anthropic, the reported Hg and Anthropic consultancy, PE firms hiring their own forward-deployed engineers, gain-share boutiques and many IT services firms all compete for the same rollout work. Pricing on outcomes alone sets a small team apart, but labs with far more money can copy it once they choose to. The independent AI value audit has the clearer opening. I found no established standard or independent provider that checks after the fact whether a rollout delivered the savings it promised. Existing offerings look forward: readiness assessments, opportunity maps and pre-deal technical diligence from firms like Zaigo, DataNorth, Exadel and Deloitte. The funded players are the deployers grading their own work, and that conflict grows as labs and their investors run more rollouts. To win, a new entrant would have to be clearly independent of every lab and integrator, have a repeatable method for measuring savings against a baseline that CFOs and operating partners trust, and target two moments: the hold-period review of a rollout and the claims made about AI savings when a company is sold. The main risk is that Big Four and assurance firms, which already own PE relationships and audit credibility, add this service, so speed and owning a recognised standard matter.

• Hg and Anthropic (Hg Catalyst, plus a reported new European AI consultancy)

  In September 2026 Hg expanded its Anthropic partnership through Hg Catalyst. It reports more than 1,600 live AI projects across its portfolio, carrying over $260m in budgeted EBITDA impact. The FT, as relayed by secondary outlets, reports that Anthropic and Hg will back an independent European consultancy to implement Claude. I found no official release for the consultancy, so it is reported rather than confirmed.

• OpenAI's Deployment Company (with TPG, Brookfield, Advent, Bain Capital)

  A joint venture with PE backers, reported at over $4bn in commitments and a $10bn valuation. It charges portfolio companies for implementation and integration, using its own staff plus engineers seconded from OpenAI. Reports describe fee-based work, not outcome pricing.

• Anthropic's PE-backed deployment vehicle (Blackstone, Goldman Sachs, Hellman & Friedman)

  Reported at about $1.5bn in commitments. Anthropic says it will embed its own applied engineers with customer teams to deploy Claude in core operations, which competes directly with forward-deployed delivery for PE portfolios.

• H.I.G. Capital in-house forward-deployed AI engineers

  A large PE firm hiring its own forward-deployed engineers to deploy AI and drive measurable EBITDA impact across portfolio companies. This shows funds bringing the delivery layer in-house.

• Gain-share AI team described by Wryland

  Describes a nine-person AI team paid through gain-share, recovering fees only on certified results above a baseline locked early in the engagement. This is the outcome-priced PE model already in practice, but it comes from a single Substack source.

• Neurons Lab

  An AI specialist firm marketing fast AI rollouts across PE portfolios.

• HTEC

  An engineering services firm positioning itself on AI value creation during the PE hold period. It is one of the IT services players this trend squeezes.

• MetaCTO

  Sells AI value creation to PE-backed companies. It ties metrics to EBITDA across cost, revenue and quality and benchmarks them before implementation, which overlaps partly with the measurement idea.

• Deloitte AI for private equity

  A Big Four practice offering AI due diligence and use cases sequenced to drive measurable EBITDA within the hold period. It says nothing about independently verifying those gains, but is the most likely incumbent to add it.

• Zaigo

  Launched in September 2026 to technically validate AI claims in PE investment memos. It is pre-deal diligence rather than post-deployment measurement, but it is the closest match to the audit idea.

• BSI independent AI verification

  A standards body making the case for independent verification of AI performance claims. It focuses on model performance, not financial results or EBITDA.

• DataNorth AI assessment for PE and VC

  A forward-looking AI readiness and opportunity assessment for PE portfolios. It does not measure whether a finished rollout delivered.

• Exadel AI maturity assessments for PE

  Another readiness and maturity assessment service sold to PE firms.

Checked 2026-10-23