Apollo's agent bank-run note
Business · Heat 1.2 / 10 · 5 shows · 19m airtime
On September 27, 2026, Apollo chief economist Torsten Slock published a research note warning that household AI agents optimizing cash returns could drain the cheap deposits banks lend against, risking a systemic banking crisis. Slock has since said he does not expect an agentic bank run, though on-chain deposit migration keeps the risk live. The FT estimates AI agents could cost banks $500 billion by winning better savings rates. What remains open is whether agent adoption would move deposits fast enough to matter, and whether the inertia the note depends on survives contact with the tools.
How 5 podcasts covered it
- The AI Daily Brief, 2026-09-28 (21:30 to 22:22)
Adds Slock's warning on agent-driven bank runs
Picks up Apollo chief economist Torsten Slock's research note warning that household AI agents optimizing cash returns could drain cheap bank deposits and trigger a systemic banking crisis, and sets it beside Gary Gensler's concern about agentic advisors clustering into the same trades.
New angle: Apollo's Torsten Slock warns household AI agents optimizing cash could drain cheap bank deposits and trigger a systemic banking crisis. - The AI Daily Brief, 2026-09-28 (22:22 to 25:24)
Extends the note into a friction debate
Extends the Apollo note into a broader debate on agents removing economic friction: Mollick on systems that only work because of friction, Levy on switching costs falling in healthcare and travel, and pushback that banks will block agents. - Closing Bell, 2026-09-28 (20:33 to 22:09)
Recaps the Apollo note on agent bank runs
Recaps the Apollo note between other stories: Slok's warning that agents optimizing checking and savings could hurt banks, with the fintech-versus-bank yield gap as the illustration. Stays inside the framing already on the record. - WSJ What's News, 2026-09-29 (5:01 to 8:11)
Recaps the Apollo note and yield gap
Recaps the Apollo note in the middle of the episode: Slok's weekend post, the 3.5 to 4.5 percent yield gap, and the squeeze on banks that lend cheap deposits. It stays inside the framing already on the record. - Bloomberg Surveillance, 2026-10-03 (21:45 to 23:38)
Walks through the agent bank-run thesis again
Torsten Slok walks through the agent bank-run thesis again, with Muse as the example: swiping cash out of a 0.01% checking account while walking into Whole Foods. It stays inside the note's frame, adding only the observation that faster money movement could make the economy more volatile. - Bloomberg Surveillance, 2026-10-12 (13:49 to 20:43)
Walks back the agentic bank-run headline
Slock himself, in studio, walks back the note's headline: he does not think an agentic bank run will happen, but says the deposit-swipe risk keeps growing as more deposits move on-chain. He also puts a number on the AI boom's contribution to growth.
New angle: Slock says he does not expect an agentic bank run, though on-chain deposit migration keeps the risk live. - Animal Spirits Podcast, 2026-10-21 (39:55 to 41:50)
Adds the FT's $500 billion agent estimate
Picks up the agentic bank-run thread with the FT's $500 billion estimate of what agents could cost banks by winning better savings rates, and notes JPMorgan, Citigroup and Bank of America moving to block or restrict outside agents. Reads it as the real-world barrier the tech prognosticators miss.
New angle: The FT estimates AI agents could cost banks $500 billion by winning better savings rates.