Visa bet its future on AI — then cut 2,600 of the engineers who build it

On 28 July 2026 Visa confirmed it would eliminate about 2,600 positions, roughly 7% of its global workforce, cutting headcount from about 34,100 to around 31,500. The news arrived the same day the payments giant was due to report fiscal third-quarter results, and it came not from a press release but from a staff memo by chief executive Ryan McInerney, first reported by Bloomberg. Visa was not in trouble: McInerney described a business with 'real momentum,' and its headcount had grown about 8% the previous year.

The cuts were concentrated in Visa's technology and product organisations — the engineering corps that designs, builds, tests and maintains the payment network — while customer-service and back-office roles were largely spared. McInerney framed the move as part of an evolution in how work gets done, saying AI was 'helping to accelerate this evolution and shape the way work gets done at Visa.' People familiar with the decision said AI was a significant factor, though not the only one; Visa said it was redirecting payroll toward AI infrastructure, stablecoin settlement and cross-border growth.

The result was a striking bet: a company cutting the people who build and run its core product while declaring that the future would be built on AI, stablecoins and new payment rails — all of which require deep engineering to deliver. Visa's argument was that AI could now automate many of those engineering tasks, so headcount could shrink even as ambition grew. Whether that trade works is the open question the layoff poses: a firm can reorganise around AI, but it cannot build the AI era with fewer builders.

Visa redirected payroll from engineering to AI, stablecoins, and cross-border growth, betting AI could replace the builders it still needed.

The cuts fell on the technology and product teams — the engineers who build and maintain the network — not the back-office or service roles an efficiency drive would more usually target.

Visa framed AI as able to automate much of that engineering work, so it treated the headcount as a cost to cut rather than the capacity it would need to deliver the AI strategy.

Visa was financially strong — record results, a growing workforce — so the cut was a choice about where to spend, not a response to distress; the bet was judgment, not necessity.

Redirecting payroll to AI, stablecoins and cross-border growth means the same engineering talent is needed to build those new rails even as the teams that run the existing network shrink.

You cannot cut the builders and still bet the company on building. Visa cut the engineers who run its network and moved their payroll to AI — assuming machines could replace people it still needed.

Visa called the reductions a capital-allocation decision, not financial distress, and said it would keep investing in AI infrastructure, stablecoin settlement and cross-border expansion. The question the case leaves is whether a company can shrink its engineering corps and still execute an AI-led transformation — or whether, by cutting the people who build and run the network, it has traded away the capacity its strategy depends on. Other employers will read Visa as an early test of how far AI can really substitute for the builders.