Featuring
Episode 120 tries to clean up the wreckage after SVB, then discovers that everyone brought a different broom. The best fights ask whether VCs were innocent depositors or conflicted accelerants, whether a bank account should behave like a paid vault, and whether venture's missing distributions were a normal J-curve or a warning flare. Friedberg has the best all-around episode: his full-balance-sheet rebuttal punctures the vault fantasy even though one of his deposit figures needs a very large correction.
Spice rack
Were venture capitalists merely SVB depositors, or did their conflicts and network behavior make them partly responsible for the bank's collapse?
Original point: Chamath says the critique of VCs is partly warranted because SVB could simultaneously invest in funds, lend to their managers, and receive portfolio-company deposits, creating relationships that deserved disclosure and scrutiny.
What everyone argued
Chamath Palihapitiya
Chamath distinguishes systemic bank stress from the narrower SVB ecosystem. He argues that VCs acting as the financial adults on startup boards should have disclosed SVB incentives and exercised more independent treasury judgment.
David Sacks
Sacks says depositors cannot be the root cause of a banking system whose balance sheets were already impaired. He accepts disclosure for specific VC conflicts but rejects scapegoating VCs for a failure produced by poor risk management, rapid rate increases, and weak supervision.
Winner circle
Sacks is right about primary causation: SVB management and supervision failed before depositors ran. Chamath is right that this does not absolve every VC relationship or treasury recommendation. The official record supports a layered verdict, but Sacks better answers the central question of what made the bank fail; the network explains speed more than underlying insolvency risk.
Commentary
Chamath Palihapitiya
Assumptions and fact checks
SVB's LP, lending, and banking relationships created material conflicts for some venture firms directing portfolio-company deposits.
Why it mattersThe relationship can create a genuine disclosure and fiduciary issue. The episode does not establish its prevalence, so it supports scrutiny rather than a blanket verdict.
David Sacks
Sacks wins the causal hierarchy but overstates the innocence of the trigger. Bad management loaded the gun; the networked withdrawal rush still helped pull the trigger.
Assumptions and fact checks
Because other banks were also stressed, VC behavior at SVB was immaterial.
Why it mattersSystemic stress and an institution-specific run mechanism can both be true. Similar vulnerabilities elsewhere do not erase the role of SVB's unusually concentrated depositor network.
SVB failed principally because management did not manage interest-rate and liquidity risk and supervisors did not respond forcefully enough.
CheckThe Federal Reserve's official review identifies management failure, concentrated uninsured deposits, rising-rate exposure, and inadequate supervision as central causes.
VC and technology networks did not contribute to the speed of SVB's run.
CheckThe Federal Reserve later concluded that social media and SVB's concentrated network of venture investors and technology firms reinforced coordinated withdrawals at unprecedented speed.
Should ordinary deposits be kept in paid, risk-free vaults instead of funding bank loans?
Original point: Jason argues that customers want payments and custody, not an undisclosed risky loan to a bank, and says many startups would gladly pay explicit fees for a true vault product.
What everyone argued
Chamath Palihapitiya
Chamath calls the vault tangent incomplete and argues for real-time, code-based regulatory oversight of bank balance sheets instead. Banks have costly infrastructure, so pure custody will be expensive; supervisors should monitor duration and liquidity continuously.
Jason Calacanis
Jason proposes fee-funded vault accounts that cannot lend deposits, leaving risk-taking to separately chosen investments. He argues that startup operators should not need to audit bank balance sheets merely to make payroll.
David Sacks
Sacks supports separating payment services from investment risk. He suggests liquid money-market holdings for deposits and securitized lending assets for investors who knowingly choose exposure.
David Friedberg
Friedberg argues that deposits fund mortgages and small-business loans; removing trillions from banks would shrink or reprice credit. Money-market funds simply lend through another manager, so the risk is shifted rather than removed.
Winner circle
Friedberg wins the architecture debate because he follows the money through the whole system. Jason identifies a legitimate product gap, Sacks improves it with marketable assets, and Chamath offers the most practical supervisory reform. But none defeats Friedberg's central point that separating deposits from lending moves risk and reprices credit rather than making either disappear.
Commentary
Chamath Palihapitiya
Chamath focuses on the operational bottleneck: supervisors must see and act on mismatches. His answer is less revolutionary than Jason's and more implementable.
Assumptions and fact checks
Real-time supervisory data and automated risk alerts would materially reduce the chance of another SVB-style failure.
Why it mattersBetter data would help, but only if supervisors have authority, expertise, and willingness to act. SVB showed that information without escalation is insufficient.
Jason Calacanis
Jason correctly notices that the legal economics of a deposit do not match ordinary customer intuition. He offers a valuable product category, but not yet a replacement architecture for the credit system.
Assumptions and fact checks
A large market of startups would pay explicit custody fees for deposits that cannot be lent or invested.
Why it mattersSVB demonstrated real demand for operational cash safety. Whether the fees would cover full-reserve custody at scale remains uncertain.
David Sacks
Sacks gives Jason's intuition a workable financial form, then concedes he is brainstorming. That modesty is warranted because the proposal moves intermediation rather than abolishing it.
Assumptions and fact checks
Turning deposits into money-market holdings and packaging loans as securities would make losses land more transparently on investors.
Why it mattersThe ownership chain can be clearer and more marked to market, though liquidity and run risk can still migrate into funds and securities markets.
David Friedberg
Friedberg wins by asking where the loans and risk go next. His $7 trillion figure is badly low, but correcting it makes the scale objection stronger rather than weaker.
Assumptions and fact checks
A large move to non-lending vault accounts would materially raise the cost or reduce the availability of mortgages and business credit.
Why it mattersUnless replacement funding arrived at comparable cost, removing a major stable funding base would reprice lending. The magnitude depends on adoption and transition design.
U.S. banks held about $7 trillion in deposits in 2023.
CheckFDIC data reported $18.6 trillion in total deposits in the second quarter of 2023, so the figure understated the system by more than half.
Did weak distributions from recent Sequoia funds show impairment, or were critics mistaking the normal venture J-curve for failure?
Original point: Jason says the report was misleading because recent venture funds naturally sit in the J-curve before portfolio companies mature and distribute cash.
What everyone argued
Chamath Palihapitiya
Chamath says limited partners must judge both capital-calling speed and time to return one times paid-in capital. After roughly five years with little distribution, an LP can reasonably see impairment and lose capacity for new commitments.
Jason Calacanis
Jason argues that early fees and immature holdings make recent funds look weak before exits arrive. He says judging 2018-2021 commitments mainly by DPI ignores the normal venture J-curve.
Winner circle
Jason wins the exact question because low DPI across young vintages was insufficient to prove impairment. Chamath wins an adjacent point: the same cash drought can still put an LP effectively out of business for new commitments. The clean ruling separates fund performance from LP liquidity instead of treating either as a proxy for the other.
Commentary
Chamath Palihapitiya
Chamath argues from the LP's real constraint—cash back, not flattering marks. His case is strongest as a liquidity warning and weaker as a final performance judgment.
Assumptions and fact checks
Little DPI after five years is strong enough evidence for an LP to treat a venture commitment as impaired.
Why it mattersIt is a serious liquidity signal, especially for an overallocated LP, but not a standalone verdict on ultimate value. Fund vintage, strategy, marks, and remaining assets matter.
Jason Calacanis
Jason wins the narrow measurement question, but 'normal J-curve' should not become a blanket excuse for bad marks or an LP cash squeeze.
Assumptions and fact checks
The majority of 2018-2021 venture funds were too young in March 2023 for low DPI to establish poor ultimate performance.
Why it mattersVenture funds commonly take many years to harvest. The later vintages were especially immature, although 2018 funds deserved closer scrutiny.
Chamath improves the debate by separating the systemic cause from ecosystem governance. He would be stronger with evidence showing how many firms had these overlapping ties and whether those ties actually drove treasury decisions.