Wall Street’s largest banks are financing AI infrastructure while adopting AI in their own operations, and executives this week described uncertainty about returns, credit exposure and security risks. Goldman Sachs CEO David Solomon said no one knows how demand for capital and computing power will evolve as the AI ecosystem matures, and said the bank must remain vigilant about what it finances and how it manages risk. Citigroup CEO Jane Fraser said higher interest rates had not slowed investment in the broader AI ecosystem and that computing capacity remained insufficient.

Big Tech has committed billions to data centers, which provide capacity to train and run generative AI models. Bank of America CEO Brian Moynihan said AI investment matters to the bank’s investment-grade debt and leveraged finance businesses. But banks face uncertainty over whether demand for computing power will match expansion plans, what returns data center projects will produce, and whether local opposition will arise. Wells Fargo CEO Charlie Scharf called the AI investment cycle a macroeconomic force and cited opportunities alongside risks from the leverage used to fund projects and upward pressure on prices in some industries and supply chains. He said economic strength and investment benefits had so far outweighed potential downside. Higher-for-longer or rising interest rates could raise project financing costs, squeeze returns and make debt service more expensive.

Morgan Stanley CFO Sharon Yeshaya said the bank was selective about risk and wary that looser credit standards at other firms could create problems; CEO Ted Pick called the bank’s involvement in AI financing “existential.” Investors are weighing economic benefits against possible losses on loans to projects that miss expected returns. Macrae Sykes of Gabelli Funds said speculative data-center projects could cause problems if their credit exposure fails, while Dave Ellison of Hennessy Funds said banks were being more careful than some investors feared and often distribute financing risk as arrangers.

Banks are also using AI to cut costs and raise productivity, though JPMorgan CFO Jeremy Barnum said benefits take time to show up in quarterly earnings at an institution of the bank’s size. Fraser said Citi is developing products at lower cost and up to 40% faster, while coding tools have reduced development effort and technology vendor expenses by 30%. She said financial benefits from AI and automation had significantly exceeded year-to-date investment, and the bank is using AI to reduce manual processes and client-service response time. Investors expect productivity gains across US companies; Brian Mulberry of Zacks Investment Management said they may question AI hype if gains and lower costs have not arrived by the end of next year. AI agents that act autonomously add security challenges alongside external cyber threats. JPMorgan CEO Jamie Dimon said the bank is developing tools to navigate agentic AI risk.

Wall Street’s largest banks are financing AI infrastructure while adopting AI in their own operations, and executives this week described uncertainty about returns, credit exposure and security risks. Goldman Sachs CEO David Solomon said no one knows how demand for capital and computing power will evolve as the AI ecosystem matures, and said the bank must remain vigilant about what it finances and how it manages risk. Citigroup CEO Jane Fraser said higher interest rates had not slowed investment in the broader AI ecosystem and that computing capacity remained insufficient.

Big Tech has committed billions to data centers, which provide capacity to train and run generative AI models. Bank of America CEO Brian Moynihan said AI investment matters to the bank’s investment-grade debt and leveraged finance businesses. But banks face uncertainty over whether demand for computing power will match expansion plans, what returns data center projects will produce, and whether local opposition will arise. Wells Fargo CEO Charlie Scharf called the AI investment cycle a macroeconomic force and cited opportunities alongside risks from the leverage used to fund projects and upward pressure on prices in some industries and supply chains. He said economic strength and investment benefits had so far outweighed potential downside. Higher-for-longer or rising interest rates could raise project financing costs, squeeze returns and make debt service more expensive.

Morgan Stanley CFO Sharon Yeshaya said the bank was selective about risk and wary that looser credit standards at other firms could create problems; CEO Ted Pick called the bank’s involvement in AI financing “existential.” Investors are weighing economic benefits against possible losses on loans to projects that miss expected returns. Macrae Sykes of Gabelli Funds said speculative data-center projects could cause problems if their credit exposure fails, while Dave Ellison of Hennessy Funds said banks were being more careful than some investors feared and often distribute financing risk as arrangers.

Banks are also using AI to cut costs and raise productivity, though JPMorgan CFO Jeremy Barnum said benefits take time to show up in quarterly earnings at an institution of the bank’s size. Fraser said Citi is developing products at lower cost and up to 40% faster, while coding tools have reduced development effort and technology vendor expenses by 30%. She said financial benefits from AI and automation had significantly exceeded year-to-date investment, and the bank is using AI to reduce manual processes and client-service response time. Investors expect productivity gains across US companies; Brian Mulberry of Zacks Investment Management said they may question AI hype if gains and lower costs have not arrived by the end of next year. AI agents that act autonomously add security challenges alongside external cyber threats. JPMorgan CEO Jamie Dimon said the bank is developing tools to navigate agentic AI risk.