FDEs Are the New Billable Hour. Here's the $9.75B Proof.

In the last 30 days, Tata Consultancy Services publicly committed to a team of up to 8,900 forward-deployed engineers for AI deployment. OpenAI's Deployment Company closed its second FDE-services acquisition — Northslope — and expanded its bench to "hundreds" of forward-deployed engineers. Microsoft launched Microsoft Frontier Company with $2.5 billion and 6,000 embedded engineers. Amazon put $1 billion into a dedicated AWS FDE unit. The pattern is no longer a trend. It is a re-pricing of how AI work gets done inside the enterprise, and it left a $9.75 billion-shaped hole at the top of the stack.

The people being paid to embed inside your organization are no longer generalist consultants. They are an explicit, named, services-tier force. Tomasz Tunguz put a number on the 12-month commitment in early July: roughly $9.75 billion in dedicated FDE spend across OpenAI, Microsoft, Anthropic, Amazon, and Google Cloud, plus Salesforce's 1,000 FDE roles. That figure is a billable-hour moment. It just doesn't look like the billable hour agencies ran for thirty years. The unit is now the embedded engineer, the deployment pod, the alliance at Cannes, the named FDE-rollup on the M&A tape.

The interesting move is not the deployment layer. It is the layer that sits above it.

The deployment layer is being claimed at extraordinary speed. TCS — India's largest IT services firm, $30B+ market cap, the named anchor of the offshore delivery model — is publicly retraining itself as an FDE shop. The OpenAI Deployment Company, which launched in May with $4B TPG-led capital, has now done two FDE M&As (Tomoro first, Northslope second), is buying Palantir-pedigree applied-AI teams, and is scaling to hundreds of FDEs. Microsoft Frontier Company is functionally a 6,000-person FDE business inside Microsoft, with $2.5B behind it and named clients Unilever and Novo Nordisk already on the books. AWS has $1B committed and is "seeding" thousands more. Salesforce, Google Cloud, and Anthropic are in the same race.

The alliance layer is being claimed at the same speed. At Cannes Lions 2026, Adobe, WPP, Microsoft, Anthropic, Stagwell's Code and Theory, Accenture, and Omnicom publicly stood up an agentic-AI coalition. WPP and AWS signed a multi-year partnership putting AWS's agentic tools in WPP Enterprise Solutions' hands. L'Oréal — a Fortune Global 500 brand — announced a strategic partnership with OpenAI at VivaTech 2026 covering product development, content creation, and product discovery inside ChatGPT across Lancôme, Kérastase, and Maybelline in the U.S. IAS — the ad verification and media quality company — just named Lidiane Jones, the ex-Bumble and ex-Slack product leader, as its new CEO, with an explicit mandate to make IAS the "trust layer" for an AI-shaped ad market. These are the seven different lanes the deployment alliance is now running down in parallel.

The brand-side strategic AI partnership is no longer hypothetical. It is named, public, and at the Fortune Global 500 tier. When L'Oréal names OpenAI as its strategic AI partner, the strategic AI partnership decision itself becomes the open, named, accountable layer above the technology partner. The deployment layer is solved by FDEs. The alliance layer is solved by consortiums. The brand-side partnership layer is solved by CMO judgment. The layer that runs above all three is the one no FDE rollup, no hyperscaler alliance, and no agentic coalition is currently positioned to staff. That is the strategy accountability layer — the place where the buyer names what the FDEs are being asked to deploy against, the alliance is being asked to enable, and the strategic partnership is being asked to advance.

Here is what that looks like in practice. TCS's 8,900 FDEs need a strategic objective or they are 8,900 expensive engineers. OpenAI's deployment bench needs a buyer-side decision framework or it is a $4B services practice with nothing to deploy. The Cannes 2026 agentic coalition needs a named strategic problem or it is a co-marketing press release. L'Oréal's OpenAI partnership needs an accountable brand-side strategy owner or it is a VivaTech announcement. IAS's new AI-credible CEO needs a measurement framework that ties to the CMO's decision calendar or it is a rebrand. Every lane above the deployment layer is the same lane, and the lane is open.

The trap for buyers is that the FDE boom is genuinely useful and genuinely seductive. The pitch is "we embed the engineers, you get the deployment." It is a real offer, and at $9.75B of named-vendor commitment in 12 months, it is now a category. But the embedded engineer still needs someone inside the buyer's organization to name the strategic question the FDE is being asked to answer. Without that, the FDE delivers competent execution against a vague brief, and the buyer pays $9.75B of supplier effort for outcomes that compound into more tool sprawl on top of the deployment team. The pattern looks like this: a CTO signs an FDE contract, the FDEs ship, the CMO still cannot name the strategic decision the deployment was supposed to advance, and the next quarter the buyer is back in the market for another round of deployment.

Autostrat is positioned to do the work the FDEs and the agentic coalitions are not staffed to do. The FDE is the new billable hour. The agency partner that delivers decisions — not access, not dashboards, not deployment, not infrastructure — is the layer the buyer still needs to name. One subscription, many outcomes, no tool sprawl. The strategy accountability layer is open. It is the one lane above the $9.75B that the named-services-tier, the named-FDE-rollup, the named-hyperscaler-alliance, the named-brand-side-partnership, and the named-AdTech-leadership-transition are not currently claiming.