Barclays Cuts 5,000 Jobs as AI Takes the Cost Base, and the CEO Calls AI the Icing, Not the Cake

Barclays will cut about 5,000 jobs this year, roughly 5 percent of its workforce, and it is unusually explicit about why: artificial intelligence and the technology change it enables are letting the bank shrink its cost base rather than simply trim it. The reductions were disclosed inside third-quarter results on Thursday, alongside a £500 million share buyback and a pretax profit that rose 12 percent but still fell short of analyst expectations.

The shape of the cuts is the part worth reading twice. About 1,000 roles went in the first half of 2026, with the remaining 4,000 due before the end of the year, and chief financial officer Anna Cross said roughly 60 percent fall in support functions including developers. Barclays employed about 93,000 people globally at the end of 2025, so this is a material change in the composition of the bank, not a rounding error. Cross added that cost reductions year to date are roughly double the level of the previous two years, and that the bank is structurally reducing the cost base to generate capacity for investment.

The financial print was mixed. Pretax profit came in at £2.32 billion against £2.08 billion a year earlier, below the £2.42 billion company-compiled consensus. Total income rose 3 percent to £7.37 billion against expectations of £7.62 billion. Fixed income, currencies and commodities income fell 19 percent to £1.01 billion where analysts had forecast a small rise, a steeper decline than the 5 percent average drop among the largest US banks, while equities income jumped 47 percent to £1.01 billion, beating the 37 percent average gain at US peers and marking the desk's best year since at least 2009. Investment bank income of £3.13 billion missed the £3.3 billion consensus. Return on tangible equity was 11.7 percent, and the common equity tier 1 ratio stood at 14.2 percent.

Chief executive C.S. Venkatakrishnan declined to dress the quarter up. He called the fixed income performance somewhat disappointing and said it would get his attention, blaming lower client activity in rates and credit plus further adversity in September from the macroeconomic environment. On the role of AI he was notably measured: the bank will employ AI at scale, he said, but AI is the icing on the cake, not the cake. That framing sits awkwardly beside a headcount reduction justified by technology efficiency, and it is a useful corrective to the more breathless claims coming from vendors this year.

Cost actions booked in the quarter were £201 million, with a further £300 million planned for the fourth quarter. Barclays guided 2026 costs to about £18.5 billion, easing to £18.4 billion in 2027, and spent £65 million in the quarter changing how it pays material risk takers, skewing senior pay further toward bonuses. The buyback of £500 million came in below the £579 million consensus, bringing year-to-date distributions to £2.8 billion. Shares have fallen more than 12 percent this year and dipped about 0.2 percent in early London trading.

The broader context is a European bank earnings season that opens with Barclays and runs into a UK budget on October 28, where finance minister John Healey is expected to raise levies on lenders. Barclays is also one of eight firms selected by the Financial Conduct Authority for the second cohort of its AI Live Testing programme, trialling AI applications with the regulator and its technical partner Advai from April through the end of the year. Elsewhere in Europe, Commerzbank and Norway DNB have announced AI-linked job reductions, and Standard Chartered chief executive Bill Winters walked back comments in May that AI would let the bank replace lower-value human capital.

What makes this quarter a marker is not the size of the cut but where it lands. Support functions and developers are exactly the roles that banks spent the past three years telling investors would be protected by reskilling, and they are also the roles where coding assistants and agentic tooling have the most measurable effect on throughput. If the savings show up as forecast, other lenders will face the same arithmetic. The honest caveat is that Barclays is cutting into weakness in its investment bank as much as into an AI dividend, and a chief executive who calls AI the icing is implicitly telling investors the cake is still trading revenue.