The AI talent wars have a new front: forward deployed engineers
OpenAI, Anthropic, AWS and Microsoft have all reached the same conclusion: building frontier models isn't enough. Winning enterprise AI now depends on the people who can actually deploy them.
Over the past few weeks, OpenAI, Anthropic, Amazon Web Services (AWS) and Microsoft have all converged on the same big billion-dollar bet: enterprise AI adoption won't succeed unless specialized engineers are embedded inside customer organizations to help companies as they rush to deploy agentic AI.
The idea may sound counterintuitive. At the same time AI companies are promising to automate white-collar work, they’re racing to hire thousands of highly paid humans to make sure the work actually, well, works. They’re called forward deployed engineers, or FDEs—customer-facing technologists who combine software engineering, systems integration and product expertise to tackle the “last-mile” challenge of deploying AI tools—from coding agents like Claude Code and Codex to custom enterprise AI systems.
Last month, AWS announced it would invest $1 billion to embed thousands of AI FDEs, while OpenAI has been making its own push, investing $150 million in a new partner program and a services firm called the OpenAI Deployment Co. that will employ FDEs on-site with business customers. Anthropic has taken a similar approach with Ode, a standalone AI services company that pairs its frontier models with teams of AI engineers to help enterprises deploy AI. And Microsoft has gone even bigger, investing $2.5 billion in Microsoft Frontier Company, a new organization that will embed 6,000 AI engineers and industry experts with customers to co-design, deploy and continuously improve AI systems.
The bottom line is that building increasingly capable models is, apparently, no longer enough. Winning enterprise AI now depends on helping customers actually deploy them.
More than consultants?
At this point, you might be wondering: Isn’t this just consulting by another name? Or, not even another name — after all, many top consulting firms have also rushed into the FDE market, including Deloitte, McKinsey, PwC and Tata Consultancy Services. I asked Francessca Vasquez, vice president of frontier AI engineering and services at AWS, exactly that. She insisted it is not.
“We’re not trying to do a rebrand of consulting,” she said. The goal, she explained, isn’t to create long-term dependence on outside experts but to help customers build production AI systems and become self-sufficient. Rather than delivering recommendations from the outside, forward deployed engineers work alongside customer teams to implement AI in production before handing off the capability.
That doesn’t mean FDEs don’t have anything in common with traditional consultants, particularly when it comes to communication, judgment and working with business stakeholders. But unlike traditional consulting engagements, she said, the work is ultimately outcome-based and engineering-led.
As foundation models become increasingly capable, she said, the real challenge is no longer model intelligence but redesigning business processes and deploying AI into the complex workflows that run large organizations.
“A lot of our demand is coming from the C-suite,” said Vasquez, who previously led a $200 million AWS strategic program that paired enterprise customers with AI experts, data scientists and strategists. She explained that AWS is embedding forward deployed engineers directly into organizations for about 45 days—including organizations like the NFL and the NBA—to build production-ready AI systems and help enterprises become self-sufficient.
A new AI talent war
There's just one problem: There aren't enough elite forward deployed engineers to meet demand. And according to a new six-month study from executive search firm Christian & Timbers, there is a tiny talent pool that companies across the AI industry are now competing to recruit. In fact, the firm estimates there are only about 2,000 “elite” forward deployed engineers in the U.S, out of a total number of FDEs of 17,000 — a number that, according to CEO Jeff Christian, falls well short of what the industry will need.
In other words, the next battle in enterprise AI may not be over who has the smartest model, but who can recruit the people capable of turning those models into measurable business results.
