Amazon Web Services just committed $1 billion to a new initiative that essentially turns its engineers into AI consultants who show up at your office and build things for you. The program, announced June 30, creates what AWS calls a Forward Deployed Engineering organization, a model that Palantir pioneered over a decade ago and that two of the biggest names in AI have already copied this year.
The concept is straightforward: instead of selling customers cloud tools and wishing them luck, AWS will send small teams of engineers, organized into pods of five to six people, to work directly alongside clients for roughly 45-day engagements. Their job is to build custom AI systems and, critically, teach the client’s own teams how to maintain them after the engineers leave.
The Palantir playbook goes mainstream
AWS is the first major hyperscaler to formalize this approach, but it’s not the first AI company to make the move in 2026. Both OpenAI and Anthropic launched their own forward deployed engineering programs in May, about six weeks before Amazon’s announcement.
Who’s signing up
