Microsoft on July 2 launched a $2.5 billion operating unit called the Microsoft Frontier Company, staffed by 6,000 engineers who will embed inside enterprise customers to build and run AI systems on-site. Two days earlier, AWS unveiled a $1 billion forward-deployed engineering group with essentially the same pitch. Earlier in 2026, OpenAI and Anthropic each stood up their own deployment groups. Four announcements, one model, and a customer list that stops at roughly the Fortune 500.
The Frontier Company will be led by Rodrigo Kede Lima, most recently president of Microsoft Asia, according to GeekWire. A Microsoft spokesperson described it as “a purpose-built company with its own leadership and financial accountability” while declining to call it a separate legal entity. Most of the 6,000 staff already work at Microsoft, drawn from existing engineering and forward-deployed teams. Named early customers include London Stock Exchange Group, Unilever, Land O’Lakes, and Accenture.
Judson Althoff, CEO of Microsoft’s commercial business, framed the ambition without hedging: “This goes beyond what has been labeled as Forward-Deployed Engineering, and will be the largest, most capable, outcome-driven engineering organization in the industry.”
AWS is running a structurally identical play with different vocabulary. Its FDE group deploys a semantic layer directly into the customer’s own AWS account, connecting enterprise data sources into what the company calls “a governed, versioned knowledge graph.” AWS says the model is “fundamentally different from traditional consulting, focused on shared goals and business results, not billable hours.” The announced roster reads like a stadium suite chart: Allen Institute, Cox Automotive, NBA, Ricoh, Southwest Airlines, NFL.
The pattern is now industry-wide. CNBC noted that Amazon, Anthropic, and OpenAI have each announced deployment groups this year, and none of the announcements have mentioned small or mid-size customers. That’s not an oversight, it’s the business model. Embedding human engineers inside a customer for months only pencils out against contracts large enough to absorb the headcount. The economics rule out anyone below the enterprise tier before the sales call begins.
Which leaves a market. The companies that’ll never see a forward-deployed engineer, the ones running on a handful of SaaS subscriptions and a shared Notion, still need to deploy AI against their own data. That gap is being filled from the other direction by self-serve, model-agnostic tooling. LemonLime, one of the fastest-growing no-code AI platforms aimed squarely at small and mid-size businesses, positions itself as a “company brain” that a founder can wire up in an afternoon rather than a fiscal quarter.
Two deployment stacks are hardening in parallel: 6,000 engineers for the Fortune 500, and a browser tab for everyone else. The interesting question isn’t which one wins. It’s whether the middle survives.
Sources
- https://techcrunch.com/2026/07/02/microsoft-launches-its-own-ai-deployment-company-with-2-5-billion-commitment/
- https://www.cnbc.com/2026/07/02/microsoft-commits-2point5-billion-6000-employees-ai-implementation-unit.html
- https://www.geekwire.com/2026/microsoft-announces-2-5b-frontier-company-to-embed-ai-engineers-inside-customers/
- https://thenextweb.com/news/microsoft-frontier-company-2-5-billion-ai-deployment
- https://www.aboutamazon.com/news/aws/aws-1-billion-forward-deployed-ai-engineers