HSBC Resumes $1 Billion Buyback, Raises Cost-Cutting Target to $2 Billion and Credits AI for Efficiency Gains
The British bank reported $10.1 billion in pre-tax profit in the second quarter, raised its cost-saving target to $2 billion by 2027, and resumed share buybacks after a pause related to the acquisition of Hang Seng.
On August 4, HSBC reported pre-tax profit of $10.1 billion for the second quarter of 2026, exceeding the consensus estimate of $9.5 billion compiled by the bank itself. Total revenue reached $19 billion, and the annualized return on tangible equity was 19.5%, with all four divisions exceeding the internal minimum target of 17%.
Georges Elhedery, CEO since September 2024, used the earnings report to make two significant announcements: the resumption of the share buyback program, this time for $1 billion, and the increase of the annual cost-saving target from $1.5 billion to $2 billion by the end of 2027. The buyback had been suspended since the closure of the acquisition of Hang Seng Bank in Hong Kong.
The AI Aspect the CEO Chose to Highlight
Elhedery stated in an interview with Bloomberg TV that the bank is already "seeing benefits from AI" in its core operations and has decommissioned approximately 50% of the applications it plans to retire by 2028. This formulation is becoming established in the discourse of global banks: fewer bets on customer-facing AI products, more direct savings on payroll and legacy application portfolios. The bank did not disclose how much of the overachievement of the cost-cutting target is attributable to AI versus other programs, and this is the black box that analysts from UBS and Barclays requested clarification on during the call.
The underlying reading of the number is the same as that used by Morgan Stanley in its recent announcement of the elimination of 2,500 positions and echoed by JPMorgan throughout its earnings calls since the fourth quarter of 2025: the productivity wave driven by AI in the back and middle office is real, but it is being captured first in banks with a history of ongoing streamlining, and not generating additional headcount expansion in growth areas.
The Counterpoint That the Earnings Report Does Not Hide
A return on equity of 19.5% is good, but supported by still high interest rates and gains in wealth management in Asia. The projection for net interest income for the year has been raised to at least $46 billion, and HSBC itself acknowledged in the report that the sensitivity to interest rate cuts by the Fed and the Bank of England remains the greatest top risk. If the cycle turns, a significant portion of the margin that supports the two announcements will evaporate, AI or no AI.
Another point of tension: Asia accounted for more than half of the quarterly profit, and mainland China has once again pulled up provisions for real estate credit. The bank did not provide sufficient segregated numbers to quantify the problem, and this is where the market tends to reprice HSBC when caught off guard negatively.
Implications for the Rest of the Sector
Among global peers, HSBC is part of the second wave of banks bringing AI into the efficiency chapter of their earnings release, alongside Morgan Stanley, JPMorgan, and Goldman Sachs. UBS, Deutsche Bank, and BNP Paribas are still in the previous stage, treating AI as a product initiative rather than an accounting driver. Sell-side analysts have begun to mark this transition as a sign of operational maturity, which is a green flag for multiples.
In Japan, MUFG and Mizuho indicated equivalent programs in January 2026, but with less ambitious decommissioning goals and no firm timeline. HSBC's competitive pressure in Hong Kong is likely to prompt movement there, especially in the competition for wealthy Asian private banking clients. In Brazil, Itaú and Bradesco have been communicating efficiency gains through AI in their quarterly results since 2025, but still without quantifying the avoided FTE share or retired applications. HSBC's standard, with the figure of 50% of applications, is now being used as a disclosure benchmark by local analysts.
What Elhedery has delivered is the model that C-level executives of any global bank will need to emulate to some degree over the next three quarters: quantify the cuts, set deadlines, and publicly acknowledge the proportion of the delta attributable to AI. Those who continue to state only that "AI is a strategic priority" without accompanying numbers will start to face pressure on their valuations.