OpenAI Launches Subsidiary DeployCo with Over $4 Billion in Investments
OpenAI has founded a company for deploying AI in the corporate environment, attracting initial investments from TPG, Bain, Goldman, and McKinsey totaling over $4 billion. The company also acquired the engineering firm Tomoro to embed AI specialists directly with clients.
DeployCo: How Sam Altman Is Building an Army of Forward Deployed Engineers, Changing the Game in Corporate AI
The Gist: What's Really Happening
OpenAI has launched a subsidiary, OpenAI Deployment Company (DeployCo), with initial investments exceeding $4 billion and simultaneously acquired the engineering firm Tomoro, bringing in about 150 deployment specialists. On the surface, it's a routine expansion into consulting. In reality, it's the largest assault on the business model of traditional IT consultants and an attempt to solve the main problem of generative AI: the gap between "the model can do everything" and "business doesn't know what to do with it."
The key term here is Forward Deployed Engineers (FDE). These are not tech support or product managers. They are engineering special forces who physically embed themselves in client teams, overhaul infrastructure, and rewrite business processes around a specific OpenAI model. The model is not sold as a boxed product; it is literally woven into the enterprise's nervous system.
The bet is not just on market share but on architectural lock-in of the client for decades to come. A company whose internal processes are tied to a specific FDE implementation from OpenAI won't be able to simply "switch to Anthropic" when they release a better model.
Timeline and Context
2023: ChatGPT boom. Corporations launch pilot projects en masse, but by year-end, it turns out 70% of pilots never reach production. The reason is not model quality but companies' inability to restructure processes.
2024: Alliances form. Tomoro is created as an applied AI consulting firm in partnership with OpenAI. Meanwhile, Palantir, which has long practiced the FDE model in the public sector, lands record contracts, proving the approach's effectiveness.
2025: The model race slows. The quality gap between GPT, Claude, and Gemini begins to narrow. Differentiation shifts from "whose model is smarter" to "whose service deploys faster and generates real money."
March 2026: OpenAI raises a $122 billion round at an $852 billion valuation, revealing monthly revenue of $2 billion and 900 million weekly users. The corporate segment now generates 40% of revenue, up from 30% a year earlier.
Late April – Early May 2026: Anthropic announces a joint venture to install and sell AI tools to corporate clients. OpenAI's response: launch DeployCo with a budget an order of magnitude larger than the competitor's initiative, and acquire the ready-made Tomoro team.
May 12, 2026: Official announcement. OpenAI reveals the partner lineup: 19 investment and consulting firms led by TPG, Advent, Bain Capital, and Brookfield. Investors also include Goldman Sachs, SoftBank, and McKinsey.
Who Wins and Who Loses
OpenAI wins. DeployCo solves three problems at once. First, it creates a direct channel to influence corporate architecture, blocking competitors. Second, it monetizes not only models but also their deployment process—the margins on this business can be higher than API access. Third, FDEs form a feedback loop: engineers in the field see real client needs and relay them directly to the product team, bypassing traditional sales filters.
TPG, Bain Capital, and Brookfield win. These funds gained access to a deal with a guaranteed annual return of 17.5% over five years. At current rates, these are exceptionally favorable terms—effectively, OpenAI pays a premium for strategic partnership. Additionally, the funds own portfolio companies that will get first access to DeployCo.
McKinsey and Bain & Company win. Their inclusion as investor partners is a preemptive move: instead of fighting DeployCo as a competitor, they've embedded themselves inside and gain privileged access to cutting-edge models. This is a classic "if you can't beat them, join them" scenario.
Accenture loses. The world's largest technology consultant, with annual revenue over $60 billion, is absent from the partner list. DeployCo directly threatens their business model of "selling methodology and implementation," replacing months of consulting with weeks of work by FDE engineers armed with knowledge of the model's internal architecture.
Accenture, Deloitte, Capgemini (in terms of independence) lose. Even Capgemini, listed among the investors, finds itself in an ambiguous position: it is both a partner and a competitor of DeployCo. The FDE model radically shortens the value chain, displacing traditional consulting layers.
Anthropic loses. DeployCo, with a budget of over $4 billion at launch, instantly devalues Anthropic's joint venture in the eyes of the market. OpenAI's investment scale is an order of magnitude larger, and the ready-made Tomoro team gives a head start that Anthropic will take months to catch up.
What the Media Isn't Saying
Insight One: DeployCo is not consulting; it's infrastructure capture. Traditional consulting ends with a report and recommendations. The FDE model ends with deep integration of the OpenAI API into the core of the client's business processes. When the accounting system, HR workflows, and supply chain are tied to a specific model with a unique call architecture, switching providers becomes not a "subscription change" but a surgical operation costing millions of dollars. DeployCo is building an economic moat around the client that only widens over time.
Insight Two: $4 billion is not a deployment budget; it's an acquisition budget. OpenAI states outright that the money will go toward "scaling operations and acquiring companies." Tomoro with its 150 engineers is just the first swallow. In the coming months, we will see a series of acquisitions of AI consulting boutiques worldwide, each bringing not only talent but also exclusive contracts with major local clients. This is a classic roll-up strategy from private equity textbooks, applied to AI services.
Insight Three: Conflict of interest among investors. Brookfield invested $500 million in DeployCo. But Brookfield also manages a portfolio of companies, many of which will become DeployCo clients. This creates a situation where the investment manager pays itself: money invested in DeployCo returns through payments from portfolio companies for deployment services, while simultaneously increasing the value of Brookfield's stake in DeployCo. This is not a scheme or fraud—it's a closed-loop value creation spiral available only to players of Brookfield's scale.
Insight Four: FDEs are not engineers; they are scouts on client territory. The constant physical presence of engineers inside client organizations gives OpenAI unique information about corporate pain points, budget cycles, and internal politics. This data cannot be obtained through API metrics or surveys. In the hands of competitors planning to launch their own products, such information is worth billions.
Forecast: Next 30 Days and 90 Days
30 days (through mid-June 2026). DeployCo will start with a dozen large FDE team placements in companies from TPG and Brookfield portfolios—quick wins that will create case studies for the external market. Tomoro will gain "preferred partner" status with OpenAI, and its clients (Tesco, Virgin Atlantic, Supercell) will become public references for DeployCo.
I expect a wave of FAANG migration: the most ambitious AI engineers from Google, Microsoft, and Meta will start moving to DeployCo, attracted by the unique combination of startup risk, big tech salaries, and direct access to cutting-edge models.
Accenture will urgently announce its own AI deployment unit, but without its own models, it will look like a belated imitation. IBM will try to play the independence card: "we work with all models, unlike DeployCo."
90 days (through mid-August 2026). DeployCo will begin a second wave of acquisitions: targets include AI consulting firms in Germany, Japan, and Brazil, where the corporate AI market is growing rapidly and local players with expertise are scarce.
The key question is how OpenAI's leadership will react to the inevitable internal conflict. The company simultaneously sells APIs to everyone (including direct competitors of DeployCo) and competes with its own clients through DeployCo. Sooner or later, an integrator like Slalom or Thoughtworks will realize it is competing not with another consultant but with a subsidiary of the model provider that has access to non-public roadmaps and internal architecture. This will lead either to a formal separation of DeployCo with a Chinese wall, or to an exodus of independent integrators from the OpenAI ecosystem in favor of Amazon Bedrock or Google Vertex AI.
Finally, if DeployCo shows even 20% quarterly revenue growth in the first two reporting periods, it will trigger a revaluation of the entire AI services sector. Today, the market values the future of AI through the lens of "who will make the best model." DeployCo is trying to prove that the real money is not in the model but in who controls its deployment in the Fortune 500. If this works, it will change the balance of power in the industry more than any technological breakthrough of 2026.
— Editorial Team
No comments yet.