Morgan Stanley's Pick calls AI financing existential as Goldman's Solomon admits demand is unknown

Six banks are quoted in the report ^[9], and they make two claims that can both be true in the same week. Citigroup's Jane Fraser speaks in the present tense: "We're not seeing a slowdown in investment in the broader AI ecosystem. There's simply not enough compute," she said ^[11]. Goldman's David Solomon speaks in the future tense. He told analysts that "no one knows how the demands for capital and compute will evolve as the broader AI ecosystem matures" ^[10].

Morgan Stanley gives the clearest view of how a bank handles that gap. Its CEO, Ted Pick, told analysts that involvement in AI financing is "existential" ^[6]. Its CFO, Sharon Yeshaya, told Reuters the worry is that "peers and other firms allow for a loosening of credit standards, and that's kind of what creates the issue" ^[8]. A bank that cannot stay out of a market it calls existential has one filter left, which is declining particular deals, and Yeshaya said that is the intent: "We're very focused on making sure that we're not involved in those types of transactions" ^[8]. Each declined deal is a fee Morgan Stanley does not collect. If a peer takes it, the paper still gets written.

Dave Ellison, a portfolio manager at Hennessy Funds, said banks are facilitating AI-related financing but often as arrangers that distribute risk ^[14].

We think the uncertainty Solomon describes is paid for in fees now and carried later by whoever holds the debt, to the extent Ellison is right about distribution. The counter-thesis is his own: that banks have kept the lessons of the financial crisis and are more careful than investors fear ^[14]. Wells Fargo's Charlie Scharf gave the macro version, saying "the strength of the economy and the positive forces from these investments have thus far outweighed the potential downside" ^[4]. We would be wrong if the paper stays on bank balance sheets instead of being sold, or if the speculative data-center builds that Macrae Sykes of Gabelli Funds singled out ^[13] are repaid on schedule through a stretch of rates that stay higher for longer ^[5].

The second front is the banks' own use of AI. The report says they are deploying it to cut costs and boost productivity, but that the payoff remains difficult to quantify ^[15]. JPMorgan CFO Jeremy Barnum said it "takes a long time to really do things that are visible at an institution of our size" ^[16]. Executives also discussed the risks of AI agents acting on customers' behalf ^[1], though the report's text does not include what any of them said about those risks.