The companies that build frontier AI models are no longer content to sell access to them. On 12 May, OpenAI launched the OpenAI Deployment Company, a consulting and engineering arm backed by more than 4 billion dollars in capital and valued, according to reporting, at around 14 billion dollars. Its founding piece is the acquisition of Tomoro, a London based AI consultancy with offices in Edinburgh and Manchester and an Asia Pacific base in Singapore.
On site, not just on tap
The pitch is hands on. Tomoro's roughly 150 engineers will work on site inside client organisations to design, test and deploy production ready AI systems, rather than leaving customers to puzzle out integration from documentation. It is a recognition that the hard part of enterprise AI is rarely the model itself. It is wiring the model into messy real world workflows, data and compliance regimes, the work where most corporate AI projects stall.
Anthropic is doing the same
OpenAI is not alone. Anthropic has announced a 1.5 billion dollar joint venture aimed at helping businesses embed AI more deeply into their operations, and both firms have reportedly been negotiating with private equity to buy up service providers that can guide non technical companies through adoption. Anthropic has also been running free AI fluency workshops on a multi city tour aimed at small and mid sized businesses. The labs are racing to own not just the technology but the expertise to apply it.
Why the shift, and why now
The move is partly defensive. OpenAI's share of the enterprise API market is reported to have slipped from around half in 2023 to roughly a quarter by mid 2025, as Anthropic and Google's Gemini made inroads. Selling deployment services deepens customer relationships and makes a model far harder to swap out. It also captures revenue that would otherwise flow to the big systems integrators, the Accentures and Deloittes that have built lucrative practices on top of other people's software.
A new front in the AI jobs story
For the labour market this is a quietly significant development. The consultancies and integrators employ hundreds of thousands of people whose business is helping enterprises adopt technology. If the model makers move into that work directly, with smaller teams and their own automation, the pressure will be felt first by the very firms that sell digital transformation. The AI jobs debate has mostly focused on workers displaced by AI. The rise of the lab as consultancy is a reminder that the disruption reaches into the advisory economy too.
