BNY Expands AI to 400-Plus Uses but Keeps Spending Opaque

BNY says it now has more than 400 artificial-intelligence solutions in production, up from 117 a year earlier, yet the bank still will not say how much it spends on the technology. That leaves investors with a familiar mismatch: rapid deployment and confident claims about returns, but no standalone cost or productivity measure.

The spending sits inside a large technology budget

Chief Executive Robin Vince told analysts that BNY spends about $4 billion of its roughly $13 billion annual expense base on technology. AI spending is folded into costs across the company rather than reported as a separate line.

Management's argument is that isolating the input would encourage investors and employees to focus on the wrong number. Vince said the bank is generating returns and wants to invest more. In the prior quarter, Chief Financial Officer Dermot McDonogh described AI spending as modest within the technology budget.

BNY presents AI mainly as a capacity tool. The goal is to let employees handle more client work without requiring expenses to grow at the same rate as revenue, rather than to announce a headcount-reduction target. That framing makes the most useful evidence operational: unit costs, processing times, error rates and the gap between revenue and expense growth.

Strong results reduce pressure for disclosure

Third-quarter revenue reached $5.7 billion, up 12% year over year, while net income rose 16% to $1.7 billion in the source report. BNY's SEC-filed earnings release also showed strong profit growth. Net interest income benefited from reinvestment at higher yields and balance-sheet growth, partly offset by deposit-margin pressure.

Noninterest expense increased 8% to $3.5 billion, slower than reported revenue growth. Return on tangible common equity reached 29%, above the bank's 28% medium-term target. BNY raised its full-year outlook to 12% to 13% revenue growth excluding notable items and 15% net-interest-income growth.

Those figures do not prove AI caused the improvement. Higher markets, interest income and business growth also lifted results. They do show that the bank is funding AI expansion while maintaining operating leverage, which gives management little incentive to break out a small, fast-changing cost category.

The next proof point

The number of production use cases is an adoption measure, not a return-on-investment calculation. Four hundred tools can be valuable, duplicative or experimental in different proportions.

Investors should look for evidence that BNY can keep expense growth below revenue growth after market and rate tailwinds fade. If the bank begins quantifying lower unit costs or faster client processing, the AI story will become more measurable. Until then, the earnings record supports management's confidence, but not a separate valuation for the technology program.