E120: Banking crisis and the great VC reset

Evidence from 2023–2025 shows that Chamath’s time‑stamped crisis call tied to the Bank Term Funding Program (BTFP) did not materialize.

• The facility’s structure didn’t create a single "March 2024 cliff." The Fed’s own description of the BTFP made clear that banks could take loans of up to one year any time until March 11, 2024, and could prepay without penalty. Advances were available until March 11, 2024, meaning the last loans can run off into March 2025, not all in March 2024. (stlouisfed.org) This undercuts the premise that “all those folks that took money” would hit a single hard maturity wall around March 15, 2024.

• When BTFP stopped making new loans in March 2024, there was no renewed systemic banking crisis. On Jan. 24, 2024, the Fed announced it would let the BTFP expire on schedule on March 11, 2024, citing that it had stabilized the system and that banks would continue to have ready access to the discount window. (federalreserve.gov) On March 11, 2024, Reuters reported that as the facility ended new lending, deposits had stabilized, bank loan books were growing, and “no bank of meaningful size” had failed in 10 months, describing BTFP as having “helped turn the tide of trouble.” (reuters.com) That is the opposite of a renewed March‑2024 panic.

• Rates were not cut massively before March 15, 2024 — yet no calendar‑driven crisis followed. The Fed kept the federal funds target range at 5.25–5.50% at its March 20, 2024 FOMC meeting, the same level it had maintained since July 2023, and merely projected gradual cuts later in 2024. (federalreserve.gov) Chamath’s conditional framing was essentially: if rates aren’t slashed, there will be a problem when the one‑year loans come due. Rates weren’t slashed, but a discrete March‑2024 banking crisis tied to BTFP maturities did not occur.

• There were localized bank stresses, but not the systemic "problem" he forecast or one clearly linked to BTFP roll‑off. New York Community Bancorp experienced serious losses and governance issues in early 2024, but coverage emphasized that broader regional‑bank indices were only modestly affected and analysts did not see its problems as systemic. (ft.com) The first FDIC‑insured bank failure of 2024, Republic First Bank in Philadelphia, came later on April 26, 2024, was small (~$6 billion in assets), and was handled routinely via sale to Fulton Bank; it highlighted ongoing rate‑ and CRE‑related pressures on some regionals, not a sudden March‑15 BTFP maturity shock. (reuters.com)

• Ex‑post Fed research describes BTFP as successfully preserving liquidity, not merely “kicking the can one year.” A 2024 Federal Reserve staff paper finds that BTFP borrowing played an outsized role in funding outflows at vulnerable banks and helped them build cash buffers and preserve liquidity during and after the 2023 turmoil. (federalreserve.gov) That record does not show a sharp, program‑driven relapse in March 2024.

Given that (1) there was no identifiable, system‑wide banking or financial crisis around March 15, 2024 associated with BTFP loans coming due, and (2) interest rates were not massively cut yet the feared event did not occur, Chamath’s prediction as stated is best classified as wrong.

Sacks argued in March 2023 that the banking turmoil would not end with Silvergate, SVB, Signature, First Republic, and Credit Suisse, and that there were likely “more shoes to drop.” Subsequent events support this view:

• Additional U.S. bank failures after the initial five (2023): FDIC data show that, beyond SVB, Signature, and First Republic, two more U.S. banks failed later in 2023—Heartland Tri‑State Bank on July 28 and Citizens Bank of Sac City on November 3—indicating further stress in the system rather than a clean resolution in March. (forbes.com)
• A sizable new failure in 2024: On April 26, 2024, regulators closed Republic First Bank (about $6 billion in assets). The FDIC was appointed receiver, and S&P Global described it as the sixth‑largest U.S. bank failure since 2010, well after the March 2023 crisis peak. (fdic.gov)
• Ongoing chain of smaller failures into 2024–2025: FDIC and secondary summaries list further bank closures—First National Bank of Lindsay in October 2024, Republic First in April 2024, Pulaski Savings Bank in January 2025, and Santa Anna National Bank in June 2025—demonstrating that bank failures continued beyond the original set of five institutions. (forbes.com)
• Serious distress episodes short of outright failure: PacWest Bancorp suffered large deposit outflows and steep share‑price declines in spring 2023, explored “all options” including a sale, and ultimately needed a rescue merger with Banc of California, reflecting continued regional‑bank fragility linked to the same rate‑shock dynamics. (cnbc.com) New York Community Bancorp then became a fresh flashpoint in early 2024, reporting a surprise quarterly loss, taking a huge provision for credit losses, slashing its dividend twice (down 80% in total), raising over $1 billion in emergency capital, suffering a Moody’s downgrade of its debt to junk, and triggering the biggest one‑day drop in the KBW Regional Banking Index since March 2023. (cnbc.com)
• Sector‑wide stress persisting a year later: A March 2024 Reuters analysis concluded that, a year after SVB’s collapse, U.S. regional banks still faced significant headwinds from high deposit costs and commercial‑real‑estate exposure, with expectations of further reserve builds and ratings downgrades—evidence that systemic pressures had not fully resolved. (reuters.com)

Because multiple meaningful bank failures and high‑profile distress episodes occurred after March 2023, and regulators and markets continued to grapple with elevated regional‑bank risks, Sacks’s prediction that there would be further significant banking‑sector problems beyond the initially affected five institutions has been borne out.

It is too early to evaluate this prediction.

The claim is that “the only stopgap in the next decade is going to be significantly higher tax rates in the United States” on corporations and high‑net‑worth individuals, as the primary way to deal with fiscal pressure from debt and entitlements, with a horizon of roughly the next decade from 2023 (i.e., around 2033).

As of November 30, 2025, several points hold:

• The core elements of the U.S. federal tax code affecting corporations and high‑income individuals (e.g., the 21% statutory corporate tax rate set by the Tax Cuts and Jobs Act, top individual brackets) have not undergone a large, clearly debt‑driven hike of the sort described. Proposals by the Biden administration to raise corporate and high‑earner taxes (e.g., higher corporate rate, minimum tax on billionaires) have either been scaled back or stalled in Congress, and have not produced a sweeping, clearly dominant fiscal response to debt/entitlement pressures.
• U.S. fiscal stress from debt and entitlement spending is well documented and growing, but there has not yet been a decisive policy turn in which significantly higher taxes on corporations and high‑net‑worth individuals are adopted and recognized as the primary stopgap for these pressures.

However, the prediction’s time window extends out to ~2033, and tax policy can change substantially in election years or around scheduled expirations (e.g., the 2017 tax cuts that are due to sunset after 2025). Until that full period has elapsed, we cannot say whether the U.S. will or will not ultimately adopt such significantly higher tax rates as its main response to fiscal pressures.

Because the forecast is explicitly about what will happen over the coming decade, and we are only about 2.5 years into that period, the correct status classification is “inconclusive (too early).”

Evidence since March 2023 runs opposite to the prediction that central banks would "take on the whole balance sheet" and function as one giant primary bank for the economy.

1. U.S.: emergency backstops were temporary and are being unwound, not expanded into a permanent mega‑bank.

   • The Fed’s Bank Term Funding Program (BTFP), created in March 2023 after the regional‑bank stress, was explicitly temporary and stopped making new loans on March 11, 2024. (en.wikipedia.org)
   • The Federal Reserve has been shrinking its own balance sheet via quantitative tightening: total assets fell from about $7.1T (Sept 25, 2024) to about $6.7T (Mar 26, 2025). (federalreserve.gov)
     These moves reduce, rather than expand, the share of the financial system’s assets held or directly backstopped on the Fed’s own balance sheet.

2. Private banks still dominate balance sheets; there is no de facto “single bank.”

   • As of early 2025 there are roughly 3,900–4,500 FDIC‑insured commercial banks in the U.S., with about $24.5 trillion in assets. (statista.com)
   • By contrast, the Fed’s balance sheet is about $6.7 trillion, and consists largely of Treasury and MBS holdings, not the direct assumption of commercial‑bank loan books. (federalreserve.gov)
     The structure of the system is still many private banks funded by deposits and wholesale markets, with a central bank as lender of last resort—not one giant operational bank providing most credit directly.

3. Deposit guarantees have not been turned into a blanket, permanent backstop of all deposits.

   • The standard FDIC insurance limit remains $250,000 per depositor, per bank, per ownership category, even after rule changes in 2024 that mainly simplified trust‑account treatment. (fdic.gov)
   • After Silicon Valley Bank and Signature Bank failed, uninsured deposits were protected via one‑off “systemic risk” exceptions, but the FDIC later recommended keeping the general $250k limit and only considering targeted higher coverage for certain business accounts, which has not been implemented system‑wide. (forbes.com)
     That is not equivalent to central banks or deposit insurers permanently absorbing “the whole balance sheet” of the banking system.

4. Globally, major central banks are also normalizing and shrinking their balance sheets.

   • The ECB has been reducing its balance sheet since 2022 by letting TLTRO bank loans mature and winding down APP/PEPP bond holdings; its consolidated balance sheet fell to about €6.42T in 2024 and is projected to keep shrinking through 2027. (eligher.zentral-bank.eu)
   • The Bank of England is likewise running quantitative tightening, actively selling or allowing gilts to mature so that its balance sheet declines from pandemic‑era peaks. (reuters.com)
     These trends mean central banks are reducing their footprint relative to the banking system, not absorbing a “very large share” of its assets.

Given that (a) the number and role of private banks remain substantial, (b) central‑bank balance sheets are shrinking rather than expanding to encompass most bank assets, and (c) deposit insurance and backstops have not been turned into a universal, permanent guarantee of essentially all bank liabilities, the world today does not resemble a system where “one big central bank” effectively runs the bulk of banking activity. The observable trajectory since 2023 is the opposite of the predicted consolidation of balance sheets into the central bank. Therefore, as of November 30, 2025, this prediction is best classified as wrong rather than merely “too early” or “ambiguous.”

There is not yet enough lifecycle data on 2023‑and‑later VC vintages to know whether they will ultimately outperform the 2021 vintage.

Venture funds typically have 10–15+ year lives, with DPI (actual cash returned) and final performance only becoming clear many years in. Industry analyses of VC and PE J‑curves show that most DPI is generated roughly years 6–16, and that typical VC funds may not even reach 1x DPI within a 15‑year window. Early‑life TVPI/IRR are known to be noisy and not reliable predictors of final outcomes.(medium.com) Since the 2021 vintage is only about four years old and 2023+ vintages are around two years old as of late 2025, both cohorts are still in the early, illiquid phase.

Available data mainly shows that the 2021 vintage has been weak so far: Carta and Wall Street Journal reporting indicate that more than 90% of 2021 US VC funds have made no distributions yet and that their median IRR is slightly negative three years in, underperforming earlier vintages.(wsj.com) Early signs for the 2023 vintage are somewhat more encouraging at the top end (e.g., high 90th‑percentile IRR), but median TVPI for 2021–2023 vintages still hovers around 1x and DPI across 2018–2023 vintages remains near zero, meaning very little has actually been realized.(sourcery.vc) Because the prediction concerns ultimate average performance of vintages 2023+ versus 2021, and we are still far from the period when final DPI and total returns can be observed, the prediction cannot yet be definitively scored.

War outcome:

• As of late November 2025, large-scale fighting in Ukraine is ongoing (e.g., Russian offensives around Pokrovsk, Kupiansk, Sumy/Dnipropetrovsk fronts and repeated missile/drone attacks). (aljazeera.com)
• There is no peace settlement or clear “end of the war,” so the claim that the war “is going to end” on this timeline has not materialized.

U.S. funding levels ("we’re not funding this" / “tens of billions” in years 2–3):

• Congress has appropriated about $174.2 billion in Ukraine-related emergency funding across FY2022–FY2024. (congress.gov)
• A major new package, the Ukraine Security Supplemental Appropriations Act, 2024 (part of Public Law 118‑50), provided about $60.8 billion more for Ukraine and related regional support in April 2024, well into “year 3” of the war. (appropriations.house.gov)
• Tracking of donors shows the U.S. remained Ukraine’s largest supporter through June 30, 2025, having allocated roughly €114.6 billion (~$134B) in military, financial, and humanitarian aid since 2022—sustained “tens of billions” over multiple years rather than a collapse. (aljazeera.com)
• While there was a months‑long political deadlock in late 2023–early 2024 that temporarily slowed new U.S. aid, it ended with the large April 2024 supplemental, and significant funding has continued rather than stopping. (reuters.com)

U.S. public and political support:

• Polling in 2024–2025 shows growing partisan division and more Americans saying the U.S. is giving “too much” support, especially among Republicans, but not a wholesale collapse of support: views are split among “too much,” “about right,” and “not enough.” (pewresearch.org)
• By early–mid 2025, substantial portions of the public still favor continuing or even increasing aid, and many believe the U.S. has a responsibility to help Ukraine, indicating sustained—if contested—support rather than the near‑total rejection implied in the prediction. (pewresearch.org)

Causal claim (“war is going to end there because we’re not funding this”):

• The war has not ended, and U.S. funding has not ceased or even dropped below “tens of billions” in the 2024–2025 window; instead, a large new package was enacted in 2024 and total commitments through mid‑2025 remain very high. (congress.gov)

Because (1) the war has not ended, and (2) U.S. aid did not collapse but continued at tens‑of‑billions‑per‑year scale into years 2–3 and beyond, Jason’s prediction—both on mechanism and outcome—is wrong.