Barclays Says AI Is Costing 5,000 Jobs, And Developers Are On The List

In this storyBarclays

Barclays said on 22 October it will cut around 5,000 jobs, about 5 per cent of its global workforce, by the end of 2026, attributing the reduction to efficiency from artificial intelligence and other technology.

The bank's third-quarter earnings presentation described "organisational simplification and technology change efficiency driven by AI" and said it was structurally reducing the cost base to generate capacity for investment. About 60 per cent of the reductions fall in support functions, and the bank stated that this includes fewer developers as AI makes its people more productive. The total includes roughly 1,000 positions already removed in the first half of the year.

The announcement came with results that were mixed. Pre-tax profit rose 12 per cent year on year to £2.3 billion, in line with forecasts compiled by LSEG, and profit attributable to shareholders rose to £1.65 billion from £1.46 billion. Total group income climbed 3 per cent to £7.37 billion but fell short of market expectations.

Within the investment bank, equities income jumped 47 per cent year on year while fixed income, currencies and commodities trading revenue declined sharply. The bank also announced a share buyback.

Barclays is led by chief executive C. S. Venkatakrishnan. Its shares were marginally higher following the announcements. AJ Bell investment director Russ Mould said the buyback had not entirely offset investor disappointment with the quarter's performance.

Two things in this announcement do not sit comfortably with the explanation attached to them.

The first is that a fifth of the cuts had already happened. Roughly 1,000 of the 5,000 roles went in the first half of the year, before this framing existed, and have now been gathered into a number presented as an AI efficiency programme. That is not dishonest, and consolidating a cost plan into one figure is ordinary practice. But it does mean the announcement is partly a relabelling of decisions already taken for older reasons.

The second is the quarter it arrived in. Pre-tax profit rose 12 per cent and equities income rose 47 per cent, so this is not a bank in difficulty. What it is, is a bank whose group income of £7.37 billion came in below what the market wanted and whose fixed income trading revenue fell sharply. Cost reduction was going to be announced either way. The question was what it would be called.

"Technology change efficiency driven by AI" is a considerably better thing to tell shareholders than "our bond desk had a poor quarter and we need the costs out." One describes a dividend from investment; the other describes a shortfall. Both can be true at once, and the first is what went in the presentation.

None of which means the underlying claim is false. The detail that suggests it is not merely a story is the one about developers, because it is the least convenient thing a bank could say. Automating call centres is a decade-old pitch. Saying you need fewer of the people who write the software is an admission that the technology has reached the function that was supposed to be safest, and no bank volunteers that for presentational advantage.

The line about developers is the one that should travel furthest, and it points at India.

Barclays runs substantial operations here through Barclays Execution Services, with sites in Pune, Chennai and Noida doing precisely the work the bank has described: support functions and software engineering. For two decades the standard response to a Western bank cutting costs has been that the work moves to India rather than disappearing. That assumption has held through every previous restructuring.

It holds less well when the stated saving is that fewer developers are needed. Offshoring arbitrages the price of an engineer. Automation reduces the number required. The first is good for India's global capability centres, which now employ well over a million people. The second is not, and this is a named employer saying the second out loud in an earnings presentation.

The caveat is that none of this tells us where the 5,000 actually fall. Banks cut expensive locations first, and London and New York are more expensive than Pune. India may yet gain headcount from this exercise while the global number falls.

What has changed is the sentence that gets said in public. Last year the justification for cost programmes was simplification. This year it is AI, and the roles being named are the ones that build it. Indian IT has spent a year arguing it is the beneficiary of that shift. Barclays has just described a version where the engineers are the line item.

For daily, sharp analysis of the biggest moves in the Indian business and startup ecosystem, follow StartupFox.