OpenAI Launches the OpenAI Deployment Company and Acquires Tomoro to Put AI Engineers Inside Your Business
OpenAI has launched a $4B+ deployment subsidiary and agreed to acquire UK firm Tomoro, placing AI engineers directly inside enterprise organisations.
On 12 May 2026, OpenAI announced the launch of the OpenAI Deployment Company, a majority-owned subsidiary backed by more than $4 billion in initial capital. At the same time, OpenAI agreed to acquire Tomoro, a UK-based applied AI consulting firm, as the subsidiary’s first and founding acquisition.
This is OpenAI moving well beyond model development and into the business of making AI actually work inside organisations.
The problem this is designed to solve
MIT research suggests that despite close to $40 billion in generative AI spending over two years, only around 5% of enterprises have demonstrated real business returns. That gap between investment and outcome is the whole motivation here.
Most enterprise AI projects stall not because the models are inadequate, but because the hard work of connecting AI to a company’s actual data, workflows, controls, and teams gets left to the client to figure out. OpenAI’s answer is to stop leaving that to chance.
What the OpenAI Deployment Company actually does
The model is straightforward in theory, though demanding in practice. The OpenAI Deployment Company sends what it calls Forward Deployed Engineers (FDEs) directly into client organisations. These are specialists in building production AI systems, and they work alongside business leaders, technology teams, and frontline staff rather than handing over a product and walking away.
A typical engagement begins with a diagnostic to identify where AI can create the most operational value. From there, the team focuses on a small number of priority workflows, building AI systems that connect to the organisation’s data, tools, and processes in ways that employees can actually rely on day to day.
The playbook will feel familiar to anyone who has watched Palantir operate. FDEs parachute into the complexity, including legacy infrastructure, compliance constraints, and organisational friction, rather than shipping software from a distance.
Tomoro: 150 engineers from day one
Acquiring Tomoro gives DeployCo something most new ventures have to spend years building: a working delivery team on launch day.
Tomoro brings approximately 150 Forward Deployed Engineers and Deployment Specialists into the new entity. The firm has already run deployments at Tesco, Virgin Atlantic, and Supercell, where its engineers built an in-game support agent serving 110 million users in 12 weeks. That kind of track record matters when you are promising enterprise clients production-ready AI systems rather than prototypes.
The acquisition is subject to regulatory approvals and expected to close in the coming months.
The funding structure and who is involved
DeployCo launches with $4 billion in investment at a $10 billion pre-money valuation, with OpenAI retaining majority control. The syndicate is led by TPG, with Advent International, Bain Capital, and Brookfield as co-lead founding partners. Other backers include SoftBank, Goldman Sachs, Warburg Pincus, B Capital, BBVA, and Emergence Capital. Consulting and systems integration firms Bain and Company, Capgemini, and McKinsey and Company are also in the group.
That last set of names is worth paying attention to. These are firms whose business has historically been selling transformation programmes to large enterprises. Having them as partners rather than competitors is a deliberate structural choice, and it means DeployCo launches with referral networks spanning thousands of businesses from day one.
Investors receive a guaranteed minimum 17.5% return, with profits capped, and OpenAI intends to use the capital to scale operations and acquire additional firms that can accelerate the mission.
What this means for you
If you work in enterprise technology or lead digital transformation: the vendor relationship with AI providers is changing. OpenAI is not just selling API access or software licences anymore. It is offering to put its own engineers inside your organisation for full-cycle delivery. That changes the procurement conversation, the contracting model, and the expectations on both sides.
If you are currently working with an AI consultancy: keep an eye on how this affects the broader market. Accenture’s stock dropped 3% on the announcement, Cognizant fell 5%, and Infosys declined 4%. The market read this immediately as OpenAI entering the professional services space. Whether that materialises as direct competition or as a shift in how consulting firms position their own AI practices remains to be seen, but it is already moving share prices.
If you have been struggling to get AI projects past the proof-of-concept stage: the FDE model is specifically designed for that problem. Rather than expecting your internal team to figure out how to connect a frontier model to your production systems, you would be working with engineers who specialise in exactly that. BBVA is a useful reference point here: what started as a ChatGPT Enterprise rollout has scaled to 120,000 employees across 25 countries, with OpenAI’s team deeply embedded in the process.
If you are evaluating OpenAI against Anthropic or Google: note that Anthropic announced a comparable $1.5 billion joint venture focused on enterprise AI integration just days before this announcement. Goldman Sachs is the only firm backing both. The race to own the implementation layer, not just the model layer, is now fully underway.
Who is running it
Brad Lightcap, OpenAI’s COO who moved to a special projects role in April, is overseeing the venture. Denise Dresser, OpenAI’s Chief Revenue Officer and former Slack CEO, will lead commercial operations. Dresser spent more than a decade at Salesforce before running Slack, which is a useful background for building an enterprise-facing business at scale.
The bigger picture
OpenAI has always described itself as both a research and a deployment company. The Deployment Company makes the second half of that description structural rather than aspirational. OpenAI’s API market share reportedly dropped from around 50% in 2023 to roughly 25% by mid-2025 as Anthropic and Google gained ground. Competing purely on model benchmarks is not a sustainable strategy when the frontier is moving this fast and this many capable models exist.
Building an implementation layer around its models, with dedicated engineers, institutional investor partners, and a ready-made consulting workforce from Tomoro, is a different kind of competitive position. It is harder to replicate than a benchmark score, and it creates ongoing relationships with the enterprises that matter most.
You can read OpenAI’s full announcement here.