Forward-Deployed Engineers: $9 billion to confirm what those on the ground already knew
In two months, the four American AI giants have invested nearly $9 billion based on a single conviction. And that conviction is not “a better model.”
An Unprecedented Convergence
Microsoft has just announced the creation of Microsoft Frontier Company: $2.5 billion and 6,000 experts (engineers, domain experts, consultants) deployed directly within client organizations to design, deploy, and continuously improve their AI systems. The approach is deliberately multi-model (OpenAI, Anthropic, Microsoft, open source), with a firm commitment: clients’ data and intellectual property will not be used to train the models. Among the first clients are the London Stock Exchange, Unilever, and Novo Nordisk.
Microsoft is neither the first nor the only one. OpenAI launched its “Deployment Company” in May, backed by more than $4 billion and a consortium led by TPG. Anthropic partnered with Goldman Sachs, Blackstone, and Hellman & Friedman on a $1.5 billion initiative targeting mid-market companies. AWS committed $1 billion two days before Microsoft’s announcement.
This model has a name: Forward Deployed Engineering, a concept developed twenty years ago by Palantir. These engineers don’t simply deliver technology; instead, they set up shop at the client’s site until value is realized.
Why now?
An MIT study (the NANDA project) has shown that 95% of generative AI pilots in companies have no measurable impact on results. The problem has never been the power of the models. It lies in deployment, team training, and change management.
In other words: American giants are now confirming—with investments in the billions—what those on the ground have long observed. Value isn’t created by the model itself, but by the way an organization adapts it to its own context.
A regional counterpart already exists
These mega-structures are primarily aimed at very large global organizations. But the approach they embody—expertise embedded within teams, long-term support, and independence from model providers—has a regional counterpart.
This is precisely the model that Darest IT&AI has been following for more than two years in French-speaking Switzerland: experts who work within organizations themselves, from strategy definition through to the integration of solutions into the business environment, guided by a steadfast conviction: an AI project succeeds when it transforms work practices, not when it merely checks off a technological box.
What the announcements in recent weeks confirm is that this community-focused approach is not merely a modest alternative to the model of the industry giants. It is the very essence of that model, applied on the scale that matters to businesses in our region: a scale based on trust, an understanding of the local context, and a long-term commitment.
The billions invested across the Atlantic validate this approach. In French-speaking Switzerland, it’s already being put into practice.
Darest IT&AI supports companies in French-speaking Switzerland through their digital transformation and adoption of AI, from strategy to deployment.