The FDE Wars: In the space of one week, two of the world’s largest technology companies committed $3.5 billion to embedding their own engineers inside enterprise customers. This is not a product launch. It is a strategic land grab for the most valuable layer in the AI stack – deployment – and Indian IT services is directly in the line of fire.
In the space of one week, two of the world’s largest technology companies committed $3.5 billion to embedding their own engineers inside enterprise customers. This is not a product launch. It is a strategic land grab for the most valuable layer in the AI stack – deployment – and Indian IT services is directly in the line of fire.
$3.5 Billion in One Week
On June 30, AWS announced a $1 billion investment to create a dedicated Forward Deployed Engineering unit – thousands of engineers organised in pods of five or six, sent directly into customer operations to build and deploy agentic AI systems. Engagements are structured around 45-day cycles, measured on business outcomes, not billable hours.
Two days later, Microsoft went bigger. Its new operating business – Microsoft Frontier Company – launches with $2.5 billion and 6,000 embedded industry and engineering experts. Microsoft’s commercial CEO Judson Althoff was unambiguous: this will be the largest, most capable, outcome-driven engineering organisation in the industry. Initial clients include Unilever and Novo Nordisk. The Frontier Company also lists Accenture, Capgemini, EY, KPMG, and PwC as SI partners – making this both a competitive and a collaborative move.
The Model Builders Got There First
The hyperscalers are not the pioneers here. In May, Anthropic formed a $1.5 billion joint venture with Blackstone, Goldman Sachs, and Hellman & Friedman to embed AI engineers inside mid-sized businesses. Days later, OpenAI launched its Deployment Company with over $4 billion from a 19-firm consortium led by TPG – alongside an acquisition of London-based consulting firm Tomoro and its 150 forward deployed engineers. The investor roster included Advent, Bain Capital, Brookfield, McKinsey, and Capgemini – some of whom are effectively funding their own future competition.
The pattern is now unmistakable. Model builders moved first. Hyperscalers followed. Every layer of the AI stack is converging on the same conclusion: the deployment gap is the real product now. Enterprises are not struggling to access AI models – they are struggling to make them work inside their organisations.
Must Read: Why India’s Large-Cap Private Banks Are About to Have a Very Good Quarter
The Karp Doctrine: Ownership Over Tokens
Palantir CEO Alex Karp – whose company pioneered the FDE model over a decade ago – crystallised the enterprise frustration on CNBC last week. His argument was blunt: enterprises are paying for tokens that create no value while surrendering their intellectual property, competitive alpha, and data sovereignty to frontier AI labs.
His sharpest provocation was economic. If AI models could genuinely generate a billion dollars in value for a client, why are labs charging for tokens instead of taking a share of the upside? The implication is uncomfortable for the frontier labs: the token-based pricing model looks extractive, not value-aligned, and enterprises are beginning to see through it.
What technical customers actually want, Karp argued, is control – over their compute, their models, their data stack, and their proprietary advantage. That means custom models built on open-source foundations. Proprietary data stacks that never leave the customer’s perimeter. Deployment architectures designed for sovereignty, not subscription.
This is not a fringe view. Microsoft CEO Satya Nadella made a remarkably similar argument in June, warning against a world where every company cedes value to a few models that consume everything they see. When two of the most powerful voices in enterprise technology are saying the same thing, the demand signal is clear: enterprises want to own their AI, not rent it.
Where Indian IT Fits – And Where It Doesn’t
Infosys flagged this shift at its Investor AI Day in February. The company identified forward deployed engineering as a critical growth vector, noting that talent demand is pivoting sharply from legacy roles – front-end web developers, QA testers, IT support specialists, blockchain developers – toward AI engineers, data annotators, AI forensic analysts, and forward deployed engineers. The numbers framing this transition are stark: 92 million traditional IT jobs globally at risk of displacement, against 170 million new roles being created.
Infosys is already working on 4,800 AI projects, has built over 600 agents, and is scaling its FDE team. The broader AI services market is estimated at $300- 400 billion through 2030. The demand is real and growing.
But so is the competition – and it is qualitatively different from anything Indian IT has faced before. An OpenAI FDE embedded inside a customer has direct access to model roadmaps, feature previews, and integration depth that no third-party system integrator can match. When model providers and hyperscalers own the deployment relationship, they control the customer’s AI architecture.
The traditional IT services playbook – labour arbitrage, linear headcount scaling, offshore delivery – does not compete in that world. The companies that will win are the ones that combine deep enterprise context, domain expertise, and the ability to deploy AI on the customer’s own terms – helping clients build sovereign AI stacks using open-source models and proprietary data, rather than locking them into vendor ecosystems.
Where This Leaves Investors
What is unfolding is not a cyclical demand shift. It is a structural redrawing of who captures value in enterprise AI deployment. The winners will be companies that transition from selling AI-discounted labour to selling AI-driven outcomes, platforms, and sovereignty. The losers will be squeezed between hyperscalers from above and AI-native startups from below.
For equity investors, this is exactly the kind of inflection point where the gap between perception and reality creates opportunity. Understanding which companies are positioning as deployment partners rather than displacement targets – and sizing that exposure before the market fully prices it in – is where long-term compounding begins.
If you want to understand how these structural shifts in enterprise AI affect the Indian IT services names in your portfolio, connect with us. This is the kind of analysis we bring to every investment conversation.
This article is for informational purposes only and does not constitute investment advice or a solicitation to buy or sell any security. Investors should conduct their own due diligence.