AI revenue grows while the company stands still (Tech Economics)

Tata Consultancy Services now books US$3.1 billion a year in AI revenue, more than 10 per cent of the company. Total revenue grew 0.2 per cent in the quarter.

That is not proof enterprise AI is fake. Tech Economics reads it as evidence that AI revenue is arriving at the same pace as automation and pricing pressure erode the traditional work it was sold to improve. Every migration that now needs fewer engineers is a fee that shrinks, whatever label goes on the invoice.

The services model was built on hours, headcount and long transformations. AI rewards fewer manual steps, reusable software and measured outcomes. Those two economies do not stack neatly. One eats the other.

The counter-case deserves respect. TCS signed US$9.6 billion of new contracts in the quarter and held its operating margin at 24 per cent. Enterprises genuinely need help connecting models to messy data, controls and legacy workflows, and the big firms own those relationships. The obituary is premature.

But the scorecard matters more than the headline. "How much AI revenue do you have?" is the wrong question for an investor or a buyer. The better ones: is total growth accelerating, is revenue per employee rising, and what share of fees is tied to results rather than effort? A vendor can grow AI revenue by re-labelling migration and consulting work. The test is whether it more than replaces the projects that now need fewer people.

I expect a barbell. Scaled incumbents who can carry liability and navigate procurement at one end; narrow AI-native specialists charging for a finished outcome at the other. The labour-heavy middle, without distribution or proprietary workflow, is where the pain concentrates.

If your advisers bill by the hour for work their own tools are compressing, whose interests does the next statement of work serve?

#ProfessionalServices #BusinessModels #AusBiz

Source: Tech Economics, 14 October 2026: https://lnkd.in/gvfE_diM