E123: Trump indictment, de-dollarization, should VCs back Chinese AI? RIP Bob Lee

Evaluated as of November 30, 2025 (about 2.6 years after the prediction), the forecast of a three‑pronged U.S. financial crisis has not fully materialized:

1. Prong 1 – Long‑dated bond losses hitting regional/community banks: This clearly occurred. The 2023 U.S. banking turmoil saw Silicon Valley Bank, Signature Bank, and First Republic fail after steep unrealized losses on long‑maturity Treasuries and MBS, triggered by rapid Fed rate hikes; regulators and analysts explicitly tied these failures to interest‑rate‑driven losses on long‑dated securities at regional banks. (en.wikipedia.org)

2. Prong 2 – Commercial real estate (CRE) downturn metastasizing into a banking crisis: CRE has become a serious stress point, especially for regional banks with large office and multifamily books. Research estimated $80–160 billion in potential bank losses from CRE defaults, and strategists have repeatedly described a "commercial real estate crisis" risk as a major threat. (forbes.com) New York Community Bancorp, for example, suffered large losses, credit‑rating downgrades to junk, and a stock collapse in 2024 largely due to CRE exposure, which briefly hammered the KBW regional‑bank index and revived fears of another banking crisis. (markets.businessinsider.com) However, so far this has produced ongoing strain and isolated institutional problems, not a broad systemic banking crisis: regulators emphasize that CRE risks are significant but generally manageable, with banks boosting reserves and gradually recognizing losses. (reuters.com)

3. Prong 3 – U.S. government debt crisis: U.S. public debt has surged past $38 trillion, and all three major agencies have now downgraded U.S. sovereign credit from the pre‑2011 AAA status, citing chronic deficits and a deteriorating fiscal trajectory. (en.wikipedia.org) There have been repeated episodes of fiscal brinkmanship, including the 2023 debt‑ceiling fight and a 43‑day government shutdown in 2025, but no actual default, loss of market access, or acute funding crisis: Treasury auctions continue to clear, the dollar remains the primary reserve currency, and officials still project ongoing (if slower) growth rather than an immediate solvency event. (en.wikipedia.org) Most mainstream descriptions frame this as a worsening long‑term debt problem, not a present‑tense government debt crisis.

Because only the first prong has clearly occurred, the second has produced stress but not a full‑blown banking crisis, and the third has not yet manifested as an actual debt crisis (despite growing risks and downgrades), the specific prediction that the U.S. would soon face a three‑pronged financial crisis is best judged wrong overall rather than fully or mostly correct.

The prediction is about conditions in 2030:

By 2030, annual net interest expense on U.S. federal government debt will exceed $1 trillion per year and will constitute more than 25% of total federal budget outlays.

As of November 30, 2025, 2030 has not occurred, so the prediction cannot yet be directly evaluated.

However, existing projections are relevant for context:

• The U.S. Congressional Budget Office (CBO) projects that net interest outlays will exceed $1 trillion annually by around 2026–2027 and continue rising thereafter.
• For 2030 specifically, recent CBO long-term or 10‑year baseline projections show net interest as a significant but still smaller share of total federal outlays—generally on the order of the mid‑teens as a percent of total spending, not as high as 25% in their central estimates. (Exact percentages vary slightly across projection vintages, but they are well below one-quarter.)

Because:

• The target year (2030) is still in the future, and
• Projections, even from CBO, are not certainties, especially about interest rates and deficits,

we cannot yet say whether net interest will actually exceed $1T and reach more than 25% of the federal budget in 2030. Therefore, the proper classification is “inconclusive (too early)”, even though current projections suggest the dollar-amount part (> $1T) is plausible while the “>25% of total outlays” part looks aggressive relative to baseline forecasts.

Key elements of the normalized prediction were:

1. Timing and drama of the apex – The podcast (Apr 7, 2023) implied the standoff would reach a dramatic climax ~60 days later (early June 2023). The U.S. hit the debt ceiling on January 19, 2023 and used extraordinary measures until Treasury Secretary Janet Yellen warned on May 1 that funds could run out as early as June 1 (later revised to June 5). Congress then passed the Fiscal Responsibility Act: House on May 31, Senate on June 1, signed June 3, ending the crisis right in that early‑June window. This was widely described as a last‑minute resolution to avoid default, i.e., a dramatic apex close to the predicted timing. (en.wikipedia.org)

2. Serious market questions about possible U.S. default – During April–May 2023, market pricing and commentary clearly reflected non‑trivial default fears:

• 1‑year U.S. CDS spreads surged to an all‑time high of about 172 bps on May 10, as negotiations were deadlocked. (y94.com)
• Analysis by the Chicago Fed and MSCI estimated market‑implied default probabilities around 4% in April–early May 2023, far above normal levels, driven specifically by the debt‑ceiling episode. (chicagofed.org)
• Fitch put the U.S. on negative watch on May 24, 2023, explicitly warning that failure to resolve the ceiling could lead to missed payments and downgrades of affected Treasuries to default‑grade ratings. (en.wikipedia.org) These indicators show that investors and rating agencies were actively and seriously questioning the possibility of a U.S. default on Treasury obligations during roughly the forecasted window.

3. Questions about Treasuries as the global risk‑free benchmark – While Treasuries ultimately retained their de facto role as the world’s main risk‑free asset, the debt‑ceiling episode did prompt explicit, high‑level questioning of that status:

• A May 3, 2023 White House Council of Economic Advisers note, citing Moody’s, warned that even a short breach could mean that if Treasury securities are no longer perceived as risk‑free by global investors, future generations would face permanently higher borrowing costs. (bidenwhitehouse.archives.gov)
• A May 2023 research note from Lombard Odier reminded clients that U.S. Treasury bills provide the risk‑free benchmark rate for financial assets worldwide and cautioned that any threat to that status would force a repricing of essentially all other assets, highlighting that the debt‑ceiling brinkmanship was testing this benchmark role. (lombardodier.com) Together, these show that mainstream policy and market analysis in that period was explicitly raising the question of whether Treasuries might cease to be viewed as perfectly risk‑free if the standoff led to default.

Nuance: The original spoken quote went further, suggesting a lasting “real shift away” from Treasuries as the baseline risk‑free asset. That stronger long‑term shift did not occur: the crisis was resolved in early June 2023 without default, and Treasuries remain the primary global risk‑free benchmark. However, the normalized prediction you provided only claims that, around June 2023, the standoff would (a) reach an intense, dramatic apex and (b) generate serious market questions about both default risk and Treasuries’ risk‑free benchmark status. On those narrower points, the events of May–early June 2023 match closely.

Given the timing, severity of market stress, and the documented debate over both default risk and the risk‑free status of Treasuries, the normalized prediction is best classified as right.

The prediction matches what actually happened in the 2023 debt‑ceiling standoff:

• Last‑minute deal: Negotiations went down to the wire. Treasury moved the projected default "X‑date" to June 5, 2023, giving Congress only a few extra days. Congress then passed the Fiscal Responsibility Act of 2023: the House on May 31 and the Senate on June 1, just days before that X‑date, with the bill signed shortly thereafter—widely described as a last‑minute resolution that avoided default. (cnbc.com)
• Debt ceiling extended/suspended: The Fiscal Responsibility Act suspended (functionally, extended) the debt limit until January 1, 2025, explicitly to avert a default in June 2023. (nlc.org)
• Spending‑cut concessions for Republicans: The law included Republican‑sought fiscal restraints: caps on growth of discretionary spending for two years, clawbacks of unspent COVID funds, reductions in IRS funding, and added work requirements for some safety‑net programs—exactly the kind of spending concessions House Republicans had demanded and then claimed as a victory. (nlc.org)
• No default occurred: Because the deal was enacted before the June 5 X‑date, the U.S. did not default on its debt in 2023. Analyses of the episode explicitly describe the act as having avoided default by suspending the ceiling through 2025. (aha.org)

All key elements of Friedberg’s forecast—last‑minute timing, extension/suspension of the ceiling, Republican‑friendly spending concessions, and avoidance of default—were borne out, so the prediction is right.

As of the 2024–2026 tax years, the top U.S. federal individual income tax rate remains 37%, with seven brackets from 10% to 37%; there has been no move to a ~70% top marginal rate on high‑income individuals.

• The IRS’s official 2024 brackets show the top marginal rate is 37% for single filers above $609,350 and married joint filers above $731,200, with no higher bracket enacted. (irs.gov)
• Inflation‑adjusted brackets released for 2026 still keep the same seven rates—10%, 12%, 22%, 24%, 32%, 35%, and 37%—with the 37% rate applying above roughly $640,600 for single filers, again with no 70% bracket. (taxfoundation.org)
• A major 2025 tax law (“One Big Beautiful Bill”) described in tax‑policy coverage permanently extends the current rate structure rather than raising top marginal rates, confirming that U.S. policy is not currently trending toward a 70% top federal bracket. (theweek.com)

The prediction also included an attitude component (that such very high rates would be broadly popular among non‑wealthy voters). Because no such 70% top rate has been implemented, that part of the prediction cannot yet be tested.

However, Friedberg’s wording in the original quote is open‑ended (“we’ll end up seeing 70% tax rates on the wealthiest people”) and does not specify a clear deadline; the user’s paraphrase (“within the coming years”) is somewhat vague and could reasonably cover a span longer than the ~2½ years that have elapsed since April 2023. Since the prediction concerns an eventual tax regime change that could still occur in the future, and there is no hard time horizon that has clearly expired, it is too early to say it is definitively wrong.

Sacks’ forecast has two parts: (1) a major repricing of big‑city U.S. office towers, and (2) that this would significantly reduce the value of pension‑owned properties and worsen the funding status of those pensions.

1. Major repricing of U.S. office towers
This clearly happened in the “near future” after April 2023:

• CommercialEdge/CommercialSearch data show the average U.S. office sale price fell 37% between 2019 and the end of 2024, with CBD office buildings down about 60% from 2019 levels, and many buildings selling at 20–50%+ discounts to prior prices. (commercialedge.com)
• Green Street’s index and related analyses put office values roughly 35–37% below their 2022 peak, the steepest drop of any major property type. (creanalyst.com)
• A Washington Post analysis estimates U.S. office buildings lost about $557 billion in value between 2019 and 2023, with big-city tax rolls (e.g., New York, D.C., Boston) now reflecting large downward reassessments. (washingtonpost.com)
• Individual towers in major markets have seen drastic markdowns—for example, One Worldwide Plaza in Manhattan was appraised at about 20% of its 2018 value by 2025, implying very large losses to its lenders and equity owners. (en.wikipedia.org)
  This strongly supports the “huge reckoning” / major repricing piece of the prediction.

2. Impact on pension funds and their funding status
Pension funds do own material office portfolios, and those portfolios have been hit:

• A CRE Daily summary of public pension results notes that large U.S. public pension funds recorded about a –6% real estate return for 2023, with CalSTRS specifically posting about a –9% real estate return, and office properties highlighted as a continuing drag. (credaily.com)
• Other large pension investors (e.g., CPPIB, PSP Investments) reported mid‑single‑ to mid‑teens real estate losses, with office the most impacted sector, and are actively cutting office exposure. (kelownarealestate.com)
  So the part about pension‑owned office assets being hit in value is supported.

However, the funding-status piece is much less clear and, at the aggregate level, appears to have moved in the opposite direction of what Sacks implied:

• Equable Institute and Reason Foundation both find the national funded ratio for U.S. state and local pension plans improved from the mid‑70% range in 2022 to roughly 78–80% by 2023–24, while total unfunded liabilities fell (e.g., from about $1.6T to around $1.4–1.5T). (prnewswire.com)
• Wilshire’s funding-status reports similarly show the aggregate funded ratio for U.S. state plans rising through 2023 and again in 2024, driven by strong overall investment returns, even though real estate was a weak sleeve. (wilshire.com)
• Large individual systems like CalSTRS and CalPERS report higher funded ratios over this period (CalSTRS at ~76.7% funded by mid‑2024; CalPERS around 79% by mid‑2025), despite negative real estate/office results, because equities and other assets performed strongly. (calstrs.com)

In other words, office repricing clearly hurt pension real-estate portfolios and forced reallocations, but did not, on net, produce a broad worsening of pension funding ratios; aggregate funded status actually improved modestly.

Because one core element of the prediction (a major office-tower repricing, including in pension-owned assets) is strongly supported by the data, while the other (that this would worsen pension funding status overall) is contradicted by observed funded-ratio trends, the forecast is only partially validated. The causal impact on funding is nuanced and depends on what counterfactual you assume (funding is better than before, but arguably worse than it would have been without office losses), so the overall verdict is ambiguous rather than cleanly right or wrong.

Friedberg’s claim is that when U.S. pension systems face funding shortfalls and asset losses, the government will intervene so that pensioners keep getting paid—perhaps with reduced benefits, but not with payments going to zero.

Since the episode aired on April 7, 2023, U.S. policy and practice have indeed continued to follow a pattern of intervention rather than allowing full pension wipeouts:

• Under the American Rescue Plan’s Special Financial Assistance (SFA) program, the Pension Benefit Guaranty Corporation (PBGC) is providing an estimated ~$97 billion to more than 250 severely underfunded multiemployer plans so they can pay all benefits due through 2051, including reinstating and making up for some previously reduced benefits. (pbgc.gov)
• Department of Labor and PBGC guidance in 2023–2025 explicitly urges eligible, distressed plans to apply for SFA so they can continue paying benefits, reinforcing the goal of preventing insolvency from cutting off pension payments. (dol.gov)
• Existing legal frameworks such as the Kline–Miller Multiemployer Pension Reform Act and the Butch Lewis Act are designed either to reduce benefits in troubled plans or to inject federal aid, rather than permitting total loss of benefits—consistent with his description that payments might be reduced but not go to zero. (en.wikipedia.org)

Over the 2023–2025 period, there are no widely reported cases of major U.S. pension systems fully defaulting with pensioners permanently receiving no payments; distressed plans have generally been supported via PBGC guarantees or special federal assistance, with at most partial benefit reductions. (This is an inference from the above policy and assistance landscape.)

However, Friedberg’s statement is effectively a long‑term structural prediction—that the U.S. will not allow pensions to go to zero in future crises. That kind of "we will never allow X" claim can only be conclusively evaluated over decades. The fact that government behavior up to late 2025 is consistent with his view does not prove it will always hold in future.

Because the prediction is about an ongoing, long‑horizon policy pattern and not a specific event by a specific date, and no contrary event has yet occurred but the future remains open, the status is best classified as inconclusive (too early to tell).

As of November 30, 2025, U.S. Social Security has not yet reached the projected insolvency date window (roughly 2030–2035), so Friedberg’s conditional prediction cannot be evaluated.

Key points:

• The Old-Age and Survivors Insurance (OASI) and Disability Insurance (DI) trust funds are still paying full scheduled benefits as of 2025.
• The 2024 Social Security Trustees Report projects the combined OASDI trust fund will be depleted around 2035, after which incoming payroll taxes would cover only about 83% of scheduled benefits if no changes are made (tax increases, benefit cuts, general-revenue transfers, or money creation).
• The prediction specifically concerns what the federal government will do at the point of insolvency (e.g., creating/issuing additional dollars to cover the shortfall vs. allowing benefit cuts). That situation has not yet occurred, and no binding policy decision has been forced by actual trust-fund exhaustion.

Because the triggering event (trust fund insolvency in the 2030–2035 window) lies in the future relative to today, it is too early to say whether the government will in fact monetize the shortfall or allow benefit reductions. Therefore the correct status for this prediction is “inconclusive (too early)”.

As of November 30, 2025, it is too early to definitively judge a multi‑decade prediction about “our lifetimes.” We can, however, check the early parts of it:

1. Debt-to-GDP continuing to rise in the U.S. and many peers

   • U.S. federal debt held by the public has remained very high post‑COVID and is on an upward trajectory as a share of GDP, with CBO projections (updated regularly) continuing to show rising debt-to-GDP over the coming decades absent major policy change.
   • Many other advanced economies (e.g., Japan, Italy, the U.K., France) also have high and generally elevated or rising gross debt-to-GDP ratios according to IMF and OECD data.
     These observations support the direction of his claim so far, but they concern a trend that is meant to unfold over decades.

2. No major systemic U.S. sovereign debt crisis so far

   • As of late 2025, the U.S. has not defaulted on its Treasury obligations, has not experienced hyperinflation, and has not faced an IMF‑style forced austerity crisis. While there have been recurring political standoffs over the debt ceiling and concerns about fiscal sustainability, none have yet produced the kind of systemic sovereign debt event that Chamath is explicitly saying will not happen in the hosts’ lifetimes.

3. Time horizon makes the core assertion untestable yet

   • The key part of the prediction is that rising debt “will not really be an issue in our lifetimes.” Because the hosts are middle‑aged and the horizon is “several decades,” the absence of a crisis by 2025 is not strong evidence either way; a crisis (or its continued absence) could still materialize many years from now.
   • Similarly, the claim that the U.S. will remain relatively better positioned than other major economies is a structural, long‑run comparative statement. While the U.S. remains a benchmark safe asset issuer and the dollar remains the dominant reserve currency as of 2025, those facts alone don’t conclusively prove the prediction for the rest of the century.

Given the explicitly long‑duration nature of the claim, the correct status as of November 2025 is “inconclusive (too early)”, even though early data (rising debt, no crisis yet, U.S. still relatively strong) are directionally consistent with his view.

Evidence from 2023–2025 shows the U.S. government and private sector have in fact accelerated AI development and deployment, with no nationwide moratorium or sustained slowdown, even as leaders publicly acknowledge major risks.

On the private side, the U.S. has decisively dominated global AI investment: from 2013–2023 U.S. private-sector AI spending was more than three times that of any other country. (axios.com) In 2024 alone, private AI investment in the U.S. reached a record ~$109 billion, far exceeding all other countries and driven heavily by generative AI infrastructure and startups. (humansareobsolete.com) By 2025, AI-related spending by major tech companies was so large that it accounted for a substantial share of U.S. GDP growth, with AI investments projected at hundreds of billions of dollars per year and described as making the U.S. economy “hooked on AI spending.” (wsj.com) This is consistent with “breakneck speed” investment rather than caution or pause.

On the public side, federal policy has focused on adding safeguards while explicitly encouraging rapid AI development for productivity and competitiveness, not halting it. Biden’s October 2023 Executive Order 14110 frames AI as having potential to make society “more prosperous, productive, innovative, and secure,” and sets a national approach to govern AI risks while emphasizing U.S. global competitiveness—there is no moratorium language. (bidenwhitehouse.archives.gov) The Commerce Department’s one‑year review of the order in late 2024 describes the administration’s mandate as to “pull every lever to keep pace with rapid advancements in AI” and to “continue charging ahead” to spur safe development and deployment. (commerce.gov) Follow‑on actions in 2024 explicitly aim to “deepen the U.S. lead in AI innovation,” including launching the National AI Research Resource pilot to give dozens of research teams significant compute and resources. (bidenwhitehouse.archives.gov)

In 2025, the subsequent Trump administration doubled down on acceleration: it rescinded Biden’s AI safety executive order and launched the large‑scale "Genesis Mission" to exploit federal data and supercomputing for AI foundation models, explicitly framed as an effort to secure U.S. dominance in AI and boost economic growth. (en.wikipedia.org) This shift further underscores that, despite ongoing concerns about AI risks, national policy and market behavior have continued to prioritize rapid AI advancement and productivity gains, without any broad, sustained pause.

Given this trajectory—rapid and growing AI investment; federal policy focused on governance and competitiveness rather than moratoria; and an explicit push under both administrations to harness AI for productivity and economic growth—the prediction that the U.S. would "pursue AI at breakneck speed" with no real slowdown or moratorium, despite acknowledged risks and in pursuit of productivity gains, has been borne out by events so far.

Evidence since April 2023 matches Chamath’s directional call on both sides of his claim.

1. U.S. LP capital into VC/PE has been constrained

• NVCA data show U.S. VC fundraising fell sharply from the 2021–22 boom: firms raised about $66.9B in 2023 and $76.8B in 2024, well below 2022 levels, with cautious LP behavior and large amounts of undeployed capital.(nvca.org)
• Commentary on 2024 fundraising notes it was one of the weakest years in a decade (only 508 funds raised capital), blaming LPs’ reluctance to commit more money because exits and distributions have been poor.(eisneramper.com)
• The Financial Times reports the number of active U.S. VC firms has fallen by over 25% since 2021, and that many LPs—especially for smaller and newer funds—have withdrawn or reduced support due to weak exits.(ft.com)
• A report summarized by The Information (drawing on PitchBook/NVCA data) highlights that U.S. venture firms raised about $67B from pensions, endowments and other LPs in 2023, down roughly 60% from $173B in 2022, confirming that traditional U.S. dollar LP flows have become much harder to secure.(theinformation.com)

2. Gulf SWFs—especially PIF—have become major sources of capital for top-tier global funds

• Global SWF and multiple press reports show GCC sovereign wealth funds (PIF, ADIA, Mubadala, QIA, etc.) have dominated global sovereign dealmaking since 2022, accounting for around 40% of global SWF transactions and managing roughly $4.9T in AUM in 2024, projected to rise significantly by 2030.(arabnews.com)
• Saudi Arabia’s PIF was the world’s most active sovereign investor in 2023, deploying about $31.6B across 49 deals, an increase of ~33% year-on-year while most other state investors cut spending.(gulfnews.com)
• A 2025 legal/market memo from Skadden notes that GCC SWFs’ capital is expected to roughly double by 2030 and that they are regularly sought out as co‑investment partners and limited partners in private market funds, confirming their role as core LPs for global PE/VC managers.(skadden.com)
• Reporting on PIF’s Sanabil (a PIF-backed investment arm) shows that by 2023 it publicly listed investments or LP relationships with dozens of top U.S. VC and PE firms, including Andreessen Horowitz, Coatue, Founders Fund, KKR, General Atlantic and Bain Capital, and that it receives heavy inbound interest from U.S. founders and VCs seeking capital.(observer.com)
• Reuters and other outlets describe PIF anchoring new fund platforms with major global managers like BlackRock and Goldman Sachs Asset Management, committing billions as the cornerstone LP for Gulf-focused private credit and equity funds—classic “replacement capital” behavior from a sovereign fund stepping into an LP role.(reuters.com)
• In parallel, Abu Dhabi’s Mubadala has surged to become the most active SWF by deal volume in 2024, with a large tilt toward private equity, AI and private credit, and has taken big stakes or partnerships with global asset managers (e.g., Silver Rock, Fortress), again underscoring Gulf funds’ centrality as capital providers to leading alternative managers.(ft.com)

Putting this together: since April 2023, (a) traditional U.S. LPs have indeed been more constrained and slower to re-up VC/PE commitments, while (b) Saudi PIF and other Gulf SWFs have aggressively expanded as LPs and strategic capital partners to top-tier global venture and private funds, and are now widely described as key sources of capital for those managers. That matches the direction and essence of Chamath’s prediction, so it is best judged right, even though the exact degree of “replacement” is qualitative rather than precisely measurable.

The prediction described Bob Lee’s killing as a random street stabbing by a psychotic homeless person with a substantial prior arrest record, repeatedly released under lenient decarceration policies.

Public reporting and the subsequent trial show a very different picture:

• Killer’s identity and status: Police arrested Nima Momeni, a Bay Area IT/tech consultant and business owner, not a homeless person. (en.wikipedia.org)
• Relationship to Lee: Authorities stated that Lee and Momeni knew each other, and the killing was not a random street crime. (washingtonpost.com)
• Motive and circumstances: Prosecutors argued that Momeni stabbed Lee after a personal dispute involving Momeni’s sister and her alleged drug-related assault, luring Lee to a secluded area and attacking him with a kitchen knife. This was framed as a planned, personal confrontation, not an unprovoked attack by a stranger. (fortune.com)
• Criminal history: Reported records show only a 2011 misdemeanor switchblade charge for Momeni, not a long “revolving door” history of arrests and releases characteristic of the prediction. (cbsnews.com)
• Political narrative rejected: Multiple outlets explicitly noted that the initial narrative of a random, homelessness-linked street killing in a “crime-ridden” San Francisco was dispelled once Momeni was arrested and his connection to Lee became clear. (washingtonpost.com)

Because the actual facts contradict every key element of the prediction (randomness, homelessness, psychosis, extensive prior record, decarceration-driven release), the prediction is wrong.