# Both AI labs quietly became consulting firms, and that tells you the model isn't the product > OpenAI and Anthropic each stood up private-equity-backed enterprise deployment firms within a week of each other. The convergence says the hard part of enterprise AI was never the model, it's getting anyone to use it. - **Pillar:** Business - **Author:** Aditya Marin Gasga (Founding Editor) - **Published:** 2026-07-01T18:25:00.000Z - **Tags:** openai, anthropic, enterprise-ai, deployment, systems-integrators, acquisition ## TL;DR OpenAI launched the OpenAI Deployment Company (DeployCo) with $4B+ and 19 investment and consulting partners, a week after Anthropic built a similar enterprise services firm with Blackstone, Goldman Sachs, and Hellman & Friedman. Both moves point to the same conclusion: enterprises can't operationalize frontier AI on their own, so the labs are racing to own the deployment layer, because the model was never the whole product. ## Key takeaways 1. OpenAI launched the OpenAI Deployment Company (DeployCo) on May 11, 2026 — majority-owned, with more than $4 billion and 19 investment and consulting partners led by TPG — and agreed to acquire Tomoro for roughly 150 Forward Deployed Engineers. 2. A week earlier, on May 4, Anthropic announced a parallel enterprise AI services firm built with Blackstone, Hellman & Friedman, and Goldman Sachs to deploy Claude inside companies. 3. Two frontier labs independently reaching the same private-equity-and-consultancy-backed services structure within a week is a signal, not a coincidence: self-serve models and APIs cannot absorb enterprise demand on their own. 4. The strategic read: if the model is a commodity, the durable moat moves up the stack to deployment, integration, and change management — the services work, not the model itself. import Figure from '~/components/article/Figure.astro'; Within a week of each other, the two leading AI labs each did the same slightly surprising thing: they became, in effect, consulting firms. OpenAI launched the OpenAI Deployment Company. Anthropic built an enterprise AI services firm with Wall Street's biggest names. Neither framed it that way, but strip the language and the shape is identical, and the fact that both arrived at it independently, at almost the same moment, is the actual story. ## What OpenAI announced On May 11, OpenAI launched the [OpenAI Deployment Company](https://openai.com/index/openai-launches-the-deployment-company/), or DeployCo, a majority-owned entity built to embed "Forward Deployed Engineers" inside customer organizations to design and ship production AI systems. It launched with more than $4 billion in initial investment and a roster of 19 partners led by the private-equity firm TPG, with Advent, Bain Capital, and Brookfield as co-leads, and Bain & Company, Capgemini, and McKinsey among the backers. Alongside it, OpenAI agreed to acquire Tomoro, an applied-AI consultancy, bringing roughly 150 engineers in from day one. The pitch is that models alone do not transform a business. OpenAI's own framing is that building powerful models is only part of the work, and that the next stage of enterprise AI will be defined by how effectively businesses can actually deploy it. So DeployCo's engineers embed inside a company, pick a few high-value workflows, and build systems around them, connecting the models to the customer's data, tools, and processes. ## Why it rhymes with Anthropic Here is the part worth noticing. A week before DeployCo, Anthropic [announced its own new enterprise AI services firm](https://www.anthropic.com/news/enterprise-ai-services-company), built with Blackstone, Hellman & Friedman, and Goldman Sachs, and backed by a consortium of alternative asset managers, to deploy Claude inside enterprises. Same structure: a frontier lab, private-equity capital, systems-integrator muscle, aimed squarely at the gap between "we have a powerful model" and "our customer's operations actually run on it." Two competitors independently reaching the same non-obvious structure within a week is a signal, not a coincidence. It means both have seen the same thing in their own enterprise pipelines: demand that a self-serve product cannot absorb, and customers who cannot operationalize the technology without hands-on help.
OpenAI — DeployCo TPG, Bain, McKinsey + Tomoro Anthropic — services firm Blackstone, Goldman, H&F Same structure: PE + consultancy-backed AI deployment services firm
## What it means The convergence resolves a question the hype tends to skip: if frontier models are so capable, why isn't adoption faster? The answer both labs are now betting on is that capability was never the constraint. Integration was. It is the same gap that strands most enterprise efforts short of production — [fewer than a quarter of AI agent pilots ever reach it](/agent-roi-pilot-production), and the ones that stall rarely stall on model quality. A model that can reason is worth little to an enterprise until someone connects it to the company's data, rebuilds the workflow around it, handles the governance, and gets employees to actually use it. That is services work, the unglamorous, high-touch, change-management labor that consultancies like McKinsey and Bain have sold for decades. The labs are now buying their way into it directly. There is a strategic read here too. If the model is a commodity that competitors can match on price and capability, the durable business, and the durable customer relationship, moves to the layer above it: the deployment, the integration, and the [ongoing operation, where the real bill for a production system actually lands](/the-full-bill-for-one-agent). Owning that layer is stickier than owning the model. Both OpenAI and Anthropic appear to have concluded that the money, and the moat, increasingly live there. For anyone deploying AI inside an organization, the useful takeaway is not about DeployCo or Anthropic's firm specifically. It is the admission underneath both: the hard part of enterprise AI is not getting a good model. It is everything that has to happen after. The labs, who know their own technology better than anyone, are telling you as much by spending billions to staff up for exactly that work. ## FAQ ### What is the OpenAI Deployment Company? A new company, majority-owned by OpenAI, launched May 11, 2026, that embeds 'Forward Deployed Engineers' inside customer organizations to build production AI systems. It launched with more than $4 billion, is led by TPG with Advent, Bain Capital, and Brookfield as co-lead partners, and includes Bain & Company, Capgemini, and McKinsey among its backers. In connection with the launch, OpenAI agreed to acquire the applied-AI firm Tomoro for roughly 150 engineers. ### How is this like Anthropic's move? Closely. A week earlier, Anthropic announced a new enterprise AI services firm built with Blackstone, Hellman & Friedman, and Goldman Sachs, backed by a consortium of alternative asset managers, to deploy Claude inside mid-sized enterprises. Two frontier labs independently reached the same structure — a PE-and-consultancy-backed services arm — at almost the same time. ### Why are the labs doing this? Because enterprise demand is outpacing what a self-serve model or API can satisfy, and because most organizations lack the in-house expertise to turn a capable model into a working production system. The labs are moving to own the deployment and change-management layer, not just the model. ### What does the convergence signal? That in enterprise AI, the model is necessary but not sufficient. The bottleneck has shifted from raw capability to integration, workflow redesign, and adoption — the unglamorous services work that consultancies have always done. Both labs are betting the money and the moat are increasingly there.