AWS is the first major cloud provider to build its own forward-deployed engineering army, committing $1 billion to embed AI engineers directly inside enterprise customers. The new Forward Deployed Engineering organization announced Tuesday puts small teams of AWS engineers inside client companies for roughly 45-day sprints, building custom agentic AI systems. Francessca Vasquez, AWS VP of Frontier AI Engineering and Services, said the unit can compress what once took months into days.
"It's like the next inflection point," Vasquez told SiliconANGLE. The FDE model was pioneered by Palantir more than a decade ago, but it has become the defining deployment strategy of the AI boom. OpenAI launched a $4 billion joint venture with TPG and Advent International earlier this year.
Anthropic followed with a $1.5 billion venture backed by Blackstone and Goldman Sachs. Both are structured as external companies with private equity capital.
AWS is taking a different approach. The $1 billion comes from internal resources, not outside investors.
The org sits inside AWS as a dedicated business unit, not a joint venture. Vasquez described it as the first time the company has consolidated its AI deployment capabilities under a single rubric. "We've had capabilities over the years, but structurally this is like getting everybody together in one business unit with a common rubric of deployment," she said.
Teams of five to six engineers will embed with customers, working alongside AI agents and partnering with business, engineering, and security staff. AWS plans to seed the unit with thousands of FDEs through a mix of external hiring and internal reallocation.
The pricing model breaks from traditional consulting. AWS will charge fixed, outcome-based fees rather than hourly billing, a structure designed to incentivize speed and results over billable hours.
Vasquez laid out a 45/45/45 metric: ideate in 45 minutes, validate in 45 hours, ship in 45 days. When a sprint ends, the team leaves behind a semantic layer and knowledge graph the customer can use to build independently. "The currency that the customers are always talking about right now is speed," Vasquez said.
Early customers include the NBA, NFL, Allen Institute for AI, Cox Automotive, and Ricoh. The NFL has already used FDE teams for fan-facing products including NFL Fantasy AI and NFL IQ. Brian Mitchell, AWS Director of Product and Solutions, said the unit targets what AWS calls the "last mile" problem of AI adoption. "Customers aren't asking what AI can do anymore," Mitchell said. "They're asking, 'How do I make it part of my business that actually runs?'" The unit will prioritize regulated industries including financial services, healthcare, and the public sector, where deployment complexity is highest and the need for embedded expertise is greatest.
AWS generated $37.6 billion in Q1 revenue, up 28% year over year, giving it the balance sheet to fund a $1 billion internal bet on AI deployment. The question for competitors is whether a hyperscaler's internal org can match the speed of PE-backed FDE ventures that answer to outside investors.
