On October 23, 2026, the Financial Times reported, citing sources, that Anthropic and London-based private equity firm Hg have agreed to launch a consultancy by the end of 2026. Per the Techmeme summary of the FT story, the new company will embed AI consultants and engineers into European businesses and advise them on how to deploy AI. We could not open the FT article, and we did not find a statement from Anthropic or Hg, so this post separates what has been reported from what has been confirmed.
The idea fits a pattern we have been following: AI labs are moving from selling model access to selling deployment. Our explainers on forward deployed roles and on the frontier deployed engineer describe why, and the OpenAI forward deployed engineering post shows the rival lab's version.
Claimed vs verified at a glance
| Claim | Who says it | Status |
|---|---|---|
| Anthropic and Hg agreed to launch a consultancy | FT, via sources, summarized by Techmeme | Reported; no company statement found |
| Launch by the end of 2026 | FT summary | Reported |
| It will embed AI consultants and engineers into European businesses | FT summary | Reported |
| Hubs in London and Berlin | Secondary coverage of the FT report | Reported; not seen in the FT text |
| Independent company, Anthropic funds it, Hg invests through managed funds including HgCapital Trust | Secondary outlets repeating the FT | Reported; unverified by us |
| Ambition of a business worth over $10 billion in five years | One secondary outlet citing the FT | Single-source; treat with caution |
| Work would extend to Hg-owned companies | Secondary coverage | Reported |
| Anthropic has an existing, similar U.S. venture | Fortune and WSJ (May 2026) | Reported earlier, see below |
What exactly is reported?
The core of the story is short. The FT's sources say the two have agreed to set up a company that will place AI consultants and engineers inside European organizations. The summary frames it as advice on how to deploy AI, which suggests the service is deployment help rather than a model product.
Several outlets repeat the FT with more detail, including GuruFocus and Invezz. Those details include the London and Berlin hubs, an independent company structure, funding from Anthropic with Hg investing through funds it manages, and an ambition to build a business worth more than $10 billion within five years. These come from outlets that we could not always open or cross-check, and several repeat each other. They may be right, but we have not seen them in a primary source, so treat each as reported.
What is not confirmed?
- No company announcement. We found no Anthropic or Hg press release, newsroom post or executive statement.
- No named leadership, headcount or pricing. Nothing we reviewed says who will run it, how many people it will employ, or what it will cost clients.
- No legal structure. "Independent company" and who owns what are secondhand.
- The $10 billion figure. It appears in a single secondary report citing the FT. An ambition is not a valuation, and it should not be read as one.
- The relationship to earlier ventures. Whether this is a European arm of an existing structure or a separate company is not clear in the reporting we could open.
How does it compare with Anthropic's May venture?
In May 2026, Fortune reported that Anthropic had formed a joint venture with Blackstone, Hellman and Friedman and Goldman Sachs, describing it as a new AI-native enterprise services company. Fortune, citing the Wall Street Journal, put the committed capital at about $1.5 billion, and listed other backers including General Atlantic, Leonard Green, Apollo, Singapore's GIC and Sequoia. The stated aim was to embed Anthropic engineers and models in the core operations of mid-size businesses, in competition with large consulting firms.
If the FT is right, the Hg arrangement is the same idea pointed at Europe, with a European private equity partner whose portfolio gives it a ready list of software and services companies. Hg is known for backing software businesses, which is why a deal that extends work to Hg-owned companies, as secondary coverage describes, would give a new consultancy a built-in first set of clients. That is an inference from the reporting, not something either company has said.
Why would Anthropic do this?
Three pressures explain the logic, and none requires speculation about secret terms.
Deployment is the bottleneck. Anthropic's CFO has been quoted saying enterprise demand for Claude is significantly outpacing any single delivery model, and a Blackstone executive described forward deployed engineer scarcity as a major bottleneck to enterprise AI adoption, as we noted in our frontier deployed engineer explainer. Anthropic has also put money into training such engineers, so a delivery company is a way to scale people the lab cannot hire fast enough.
Europe is a growth market with its own rules. European buyers face data residency, procurement and regulatory questions, and often want a local partner. Anthropic has been building a European presence, and our coverage of the region includes Metis Reasoning's London lab and NVIDIA's European partner announcements.
Services lock in usage. A consultancy that builds workflows around Claude makes Claude the default. The risk, which analysts will raise, is conflict of interest: advice from a firm tied to one model vendor is not neutral.
What questions should buyers ask if the consultancy launches?
A consultancy tied to a model maker is a different proposition from an independent integrator, and procurement teams should treat it that way. These are the questions worth putting in a request for proposal before any contract.
Is the advice model-neutral? If the firm is funded by Anthropic and staffed with people who work closely with Claude, will it recommend other models when they fit better? Ask for a written policy and for examples of projects delivered on other vendors' models.
Who holds the data? Embedded engineers will see internal systems. Ask where client data is processed, whether it can touch Anthropic's own systems, and how that interacts with data residency promises. For European clients the answers need to line up with GDPR obligations and the EU AI Act timeline.
How are conflicts handled between the consultancy and Anthropic? If a client's needs conflict with Anthropic's product roadmap or pricing, whose interest comes first? An independent legal entity helps on paper, but contracts and governance documents show the reality.
What happens when the engagement ends? Embedded teams build systems that clients depend on. Ask about documentation, handover, code ownership and whether the work runs without the consultancy's people on site.
How are results measured? The deployment gap is real, and surveys keep showing that many organizations see little return on AI spending, a theme we covered in the Gartner CIO survey post. A consultancy that wants a $10 billion business, if that ambition is accurate, will need to show measured outcomes, not demos.
What does it mean for people who build with AI?
For most builders, nothing changes today. The practical effects, if the report holds, are further out.
- Enterprise teams in Europe may get a new path to dedicated Claude deployment help, which could compete with systems integrators and with the deployment arms of other labs.
- Consultancies and integrators face another well-funded competitor with direct model access.
- Freelancers and agencies who sell Claude implementation should watch whether the venture hires aggressively and what it charges, since it will set expectations for rates.
- Workers should note the broader labor question. Anthropic has published its own worker retraining evidence review, and consulting is one of the professions AI is expected to reshape.
Why does a private equity partner matter?
Private equity firms own many mid-size companies that run on aging processes and have an incentive to raise margins. For a portfolio owner, an AI deployment partner is a lever applied across many businesses at once: one playbook for support, finance operations or document handling, repeated company by company. That is the commercial logic behind pairing a model maker with an owner such as Hg, and it is also why the May venture drew in firms like Blackstone and Hellman and Friedman.
It cuts both ways. Portfolio-wide rollouts can produce real savings and learning, and they can also produce job cuts presented as efficiency. Readers following that debate may want our coverage of Barclays and AI-linked job cuts and the Goldman Sachs displacement estimate. Neither is about this venture, but they describe the environment in which European employers will evaluate it.
How can you check this story yourself?
- Look for the original FT article and read its attribution wording.
- Check Anthropic's newsroom and Hg's news page for an announcement.
- Watch Companies House and German commercial register filings for a new entity.
- Check job boards for London and Berlin roles that mention the venture.
- Note which details appear only in secondary outlets, and wait for a primary source before quoting them as fact.
We will update this post if either company confirms or corrects the report.
Related reading
- Forward deployed roles and the future of work
- What is a frontier deployed engineer?
- OpenAI forward deployed engineering and the deployment gap
- Anthropic worker retraining evidence review
- Metis Reasoning: a London lab, claimed vs verified
- Uber puts all 30,000 employees on ChatGPT for work
This post reflects reporting available on October 23, 2026. Details come from the FT as summarized by other outlets and may change when the companies comment.