Anthropic and Hg are building a European consultancy because model access is no longer the scarce part

Anthropic and London-based private-equity group Hg have agreed to back an independent European AI consultancy, expected to launch by the end of 2026 with hubs in London and Berlin. Its consultants and engineers will work inside companies to design AI products and improve existing deployments. Hg's portfolio will provide an immediate customer base, while the business can also sell to companies outside it.

The arrangement follows Anthropic's US consultancy Ode and a deployment joint venture launched by OpenAI in May. It also moves model makers toward territory occupied by Accenture and the Big Four. The competition is not primarily over writing strategy presentations. Enterprise AI projects fail in the connective work: cleaning data, integrating permissions, changing workflows, evaluating outputs and assigning responsibility when an automated process is wrong. A model supplier that captures that work can learn where adoption stalls and make its technology the default before a procurement exercise begins.

Hg brings 150 existing AI specialists and a collection of software businesses that can be converted into repeatable case studies. Anthropic will invest from its balance sheet; Hg will invest through managed funds, including HgCapital Trust. The new company is formally independent, but its origins create an obvious incentive to recommend Anthropic. Clients should therefore distinguish advice about whether to use AI, which model to select and how to deploy it—three questions that may not have the same answer.

For Anthropic, consulting revenue is useful but the more important return is distribution. Frontier-model prices are falling, enterprise customers can switch through aggregators, and training alone does not produce defensible workflow ownership. Embedding engineers in European companies makes the model one component of a larger operational relationship. It also offers policymakers a more concrete form of European AI capacity than another pledge to provide cloud credits: local implementation talent, albeit attached to an American laboratory and a private-equity owner.