Buying an AI license is the easy part. Getting it to actually work inside a large company — that's where most enterprises stall. Microsoft's answer is Frontier Company, announced July 2: a $2.5 billion bet with 6,000 engineers deployed directly inside client organizations.
The setup is straightforward but unprecedented in scale. Microsoft engineers work not from Redmond but from inside their customers — turning AI subscriptions into running operations. Commercial Business CEO Judson Althoff described it as going "beyond what has been labeled as Forward-Deployed Engineering."
London Stock Exchange Group, Unilever, Land O'Lakes, and Accenture are the first names on the client list. Microsoft enters with an advantage rivals can't easily copy: its people are already inside thousands of Fortune 500 companies. Frontier Company formalizes what that presence is for.
Amazon announced a $1 billion program along the same lines just two days prior. OpenAI and Anthropic built similar structures earlier. The convergence isn't accidental — every major AI vendor has hit the same wall: enterprises buy AI tools but can't deploy them at scale.
The business logic runs deeper than consulting fees. Embedded engineers generate performance data and long-term lock-in, differentiating Microsoft from vendors who just sell subscriptions. If the LSE model works, it scales across thousands of clients. Microsoft already has the relationships. Now it has the unit to leverage them.
