A 1.14 billion dollar vote that AI grows revenue. The question is who owns the operating model.

HCLTech landed the first AI-operating-model mega-deal in years. It is a vendor-owned operating model; the client-owned one is the opening.

HCLTech ended a roughly three-year drought in mega-deals this week with a 1.14 billion dollar contract to build and run a client's AI operating model. It matters as a data point well beyond one firm: it is proof that AI can be an engine of large enterprise contracts rather than a threat to services revenue, which is the fear that has hung over the whole sector. C Vijayakumar used it to position HCLTech as one of the best AI solutions companies in the world.

The framing around these deals is where it gets interesting, because a competitor's chief executive has already described the product better than any of them can sell it. Cognizant's Ravi Kumar told TIME that agentic AI is not something an enterprise can buy off the shelf and configure. It is a system built around the particulars of each enterprise, its reimagined workflows, its institutional knowledge, its data, its risk tolerance, and its regulatory reality. He is right, and that is precisely the argument for owning it.

Because the split is not whether the system is bespoke. Everyone now agrees it is. The split is what the client is left holding when the work is done. A billion-dollar engagement that builds the operating model inside the vendor's people and platform leaves the client renting the thing that runs the business. When the contract renews or the vendor's terms change, the leverage sits on the far side of the table.

The opposite of that is not a smaller deal. It is a different deliverable: the same bespoke operating model, shipped as code the client owns and can audit, running on whatever model is available underneath. If agentic AI has to be built for each enterprise, the enterprise should own the thing that gets built. The mega-deal proves the demand. Ownership is the open position.