The buyers named the deflation themselves.

Infosys conceded AI-driven price deflation, Accenture rationed its own AI on cost, Wipro's margin hit a fifteen-quarter low.

The buy side spent the week conceding the bill from the inside. Infosys's chief financial officer named AI-driven pricing deflation on the record, which is a provider admitting the technology compresses its own price faster than it can prove the offsetting value. Accenture disclosed that it is curbing routine internal use of generative AI to manage token cost, a firm rationing its own consumption of the thing it sells to everyone else. Wipro posted a fifteen-quarter-low operating margin with AI investment named as the driver.

Infosys used the same window to name a successor to Salil Parekh, the three-decade insider Ashiss Kumar Dash, effective April 2027, alongside a second cut to its growth outlook. Read together it is one admission repeated across the tier: the people paying for AI are pricing its cost with precision and its return not at all, because the cost is metered and the return is unmeasured.

This is the reckoning the labor-based model has been circling all year. When productivity compresses the price and no one can measure the value that is supposed to offset it, the way through is not another discount. It is to capture the productivity as an asset the client owns, so the provider is paid for a proven outcome rather than the hours the technology just erased.