E1: US Response to COVID-19 & Impact on Startups, Venture Capital & Public Markets with David Friedberg

Major U.S. equity indices never reached anything close to a ~50% peak‑to‑trough decline in the COVID‑19 crash.

• The S&P 500’s pre‑COVID closing peak was 3,386.15 on February 19, 2020; at the March 23, 2020 bottom it closed at 2,237.40, a drawdown of about 34% from that peak, not 50%. (justapedia.org)
• On the same day, the Dow Jones Industrial Average closed at 18,591.93, roughly 37% below its February 2020 high, again far short of a 50% correction. (grokipedia.com)
• After March 23, 2020, U.S. markets rebounded and the S&P 500 made new all‑time highs by August 18, 2020, without ever having been down ~50% from the February 19, 2020 peak. (justapedia.org)

Since the COVID‑era stock market decline from the pre‑crash peak bottomed around 34–37% and then recovered, Jason’s expectation of a roughly 50% correction did not come to pass.

Steven Mnuchin warned senators in mid-March 2020 that U.S. unemployment could reach about 20% without swift economic rescue measures in response to COVID-19 shutdowns, a scenario widely reported as a worst‑case outcome. (cnbc.com)

In reality, the official U.S. unemployment rate (U‑3) peaked at 14.7% in April 2020 and then declined to 13.3% in May, 11.1% in June, 10.2% in July, and 8.4% in August—well below the 20–30% range “by the summer of 2020.” (jobmarketmonitor.com) A broader “real” unemployment measure (U‑6) did briefly reach about 22.8% in April 2020, but it too had fallen below 20% by June and July, so even that measure did not sit in the 20–30% range during the summer months. (cnbc.com)

However, Friedberg explicitly framed the 20–30% figure as conditional: it would occur “if we keep this up,” i.e., if the then‑current level of economic shutdown persisted. In practice, that condition did not hold. All U.S. states began phased reopenings in late April and May, and by June 2020 every state had at least partially lifted stay‑at‑home orders and business closures, substantially relaxing the economic shutdown compared with March. (pmc.ncbi.nlm.nih.gov) Because the severe nationwide lockdown did not continue through the summer, we never observe the counterfactual world Friedberg was talking about.

So, as a prediction about what actually happened, unemployment did not reach 20–30% by summer 2020; but as a strictly conditional “if the shutdown at March levels continues” forecast, the antecedent never occurred, making its truth value inherently indeterminate. For that reason, the fairest grading is ambiguous rather than clearly right or wrong.

Evidence on global food and grain stocks around 2020 shows that the world had far more than 30 days of caloric supply in storage.

• FAO cereal stock data (2019/20–2020/21). FAO’s Cereal Supply and Demand Brief projected world cereal stocks at about 863 million tonnes for 2019/20, with a global cereal stock‑to‑use ratio of ~31%, characterized as a “comfortable” supply situation.(farms.com) For 2020/21, FAO projected cereal stocks of ~866 million tonnes and a stock‑to‑use ratio of 30.7%.(feedstuffs.com)
• Converting stock‑to‑use ratios to days of consumption. IFPRI and other analysts routinely translate stock‑to‑use ratios into “days of use” by multiplying the ratio by 365 (e.g., SUR × 365).(ifpri.org) A 30–31% stock‑to‑use ratio therefore implies on the order of 110–115 days of global cereal consumption in storage, not 30 days.
• Historical “tight” years are still well above 30 days. Analyses of earlier tight markets found world grain stocks at 57 days of consumption in 2006(search.issuelab.org) and about 68 days in 2012(countercurrents.org)—levels widely described as dangerously low, yet still more than double Friedberg’s 30‑day figure.

Friedberg’s statement—“The world lives on a 30 day food supply. So if you stop food production today, there would only be 30 days of food for the whole world to eat based on our calorie consumption per day”—rests on the factual claim that existing global food stocks equaled roughly 30 days of caloric consumption at that time. Contemporary FAO and related data instead show roughly 3–4 months of cereal consumption in storage, even before counting non‑cereal foods and retail/household stocks.

While the hypothetical “if you stopped production today” scenario was never literally tested, the empirical prerequisite of his prediction (only ~30 days of food in stock) is inconsistent with the best available global stock data. On that basis, the prediction is wrong.

Multiple retrospective estimates of the United States’ COVID-19 infection fatality rate (IFR) fall comfortably inside David Friedberg’s predicted range of 0.15%–1.8%.

Key evidence:

• A 2025 analysis using New York City data, adjusted to the U.S. age structure, estimates an overall U.S. COVID-19 IFR of 0.86%. The same paper reviews other work, including Sorensen et al. (2022), which estimated a U.S. IFR of 1.28% (95% CI 0.771–1.877%) before April 2022, based on seroprevalence surveys. Both central estimates, 0.86% and 1.28%, lie within 0.15%–1.8%, and Sorensen’s confidence interval essentially matches Friedberg’s upper bound. (pmc.ncbi.nlm.nih.gov)

• The COVID-19 pandemic statistics page (summarizing CDC and other data) reports that by November 2022, an estimated >313,686,000 Americans (about 94.2% of the population) had been infected at least once according to serosurveys, and that total U.S. COVID-19 deaths were about 1,231,440. Using these numbers, a crude cumulative IFR is ~0.39% (1,231,440 / 313,686,000), again squarely in Friedberg’s range. (en.wikipedia.org)

• A CDC-linked seroprevalence-based analysis notes that by February 2024 the U.S. had experienced roughly 1.2 million COVID-19 deaths, consistent with the mortality levels used above and with IFR estimates on the order of a few‐tenths of a percent to about 1%. (academic.oup.com)

• Meta-analyses and national-level modeling (summarized in the Wilson 2025 paper) find typical pre-vaccine IFR estimates around 0.4%–1% in high‑income settings, with a commonly cited early-pandemic global/meta-analytic average near 0.68%—again entirely within 0.15%–1.8%. (pmc.ncbi.nlm.nih.gov)

• Later in the pandemic, the variant-period IFR dropped sharply (e.g., Omicron-era IFR around 0.05% in Italy), but this is for late phases only; the cumulative U.S. IFR over the whole pandemic, dominated by more lethal 2020–2021 waves, remains well above 0.15% and below 1.8%. (pubmed.ncbi.nlm.nih.gov)

Given that best-available U.S. estimates cluster roughly between 0.3% and 1.3% for the population-wide IFR over the pandemic, Friedberg’s forecast interval of 0.15%–1.8% successfully brackets the realized values. Therefore, the prediction is essentially correct as stated.

Why this prediction is hard to score

1. What Chamath actually predicted
   In the episode’s closing segment, Chamath lays out an optimistic scenario where cheap IgG/IgM antibody tests come online at scale in ~2 weeks, cities are split into “green zones” based on test results, and people can re‑enter restaurants and public life with some form of “papers.” He then says: “We will establish demark zones … green zones … where people who are either negative or who have already gotten it and have tested positive for the antibodies will be allowed to interact. So I am telling you that it's within six weeks from now. Six weeks from now we'll be in restaurants.” (podscripts.co)
   Earlier in the episode, David Friedberg similarly argues that with large‑scale antibody testing of the general population, “we can really … nip this thing in the bud in a … four, five, six week time frame” if the U.S. mobilizes voting infrastructure, the National Guard, etc. (podscripts.co)

   The normalized version you gave captures this as a conditional: if the U.S. rapidly deploys massive antibody testing and zoning as described, COVID spread could be brought under control in ~4–6 weeks.

2. The condition was never met
   The scenario required nationwide, cheap, high‑volume antibody testing used to gate access to “green zones” (workplaces, restaurants, etc.). In reality:

   • Antibody (serology) tests did start rolling out in spring 2020, but their accuracy and interpretation were highly uncertain; the WHO explicitly warned in April 2020 that there was “no evidence” that antibodies conferred reliable immunity and cautioned against using them as the basis for “immunity passports” or risk‑free certificates. (who.int)
   • Civil‑liberties and technical analyses (e.g., EFF) likewise argued that immunity passports based on antibody tests were scientifically weak and logistically problematic, and described such systems as something governments were considering, not implementing. (eff.org)
   • The U.S. never adopted a federal antibody‑based passport / zoning system of the kind Chamath and Friedberg describe; later access controls (where they existed) were largely based on vaccination status, not serology. Contemporary coverage of “immunity passport” debates focuses on proposals and ethical objections, not on any rollout in the U.S. (cnbc.com)

3. What actually happened to COVID in the U.S.
   Irrespective of that hypothetical policy, COVID in the U.S. was not “nipped in the bud” within a few weeks of March 2020:

   • The first major U.S. wave peaked in April 2020, but was followed by substantial summer 2020, winter 2020–21, Delta (2021), and Omicron (late 2021–22) waves, with the national public‑health emergency only ending on May 11, 2023. (en.wikipedia.org)
   • By 2022, over 94% of Americans were estimated to have been infected at least once, and the cumulative death toll exceeded 1.2 million—hardly a situation where spread was brought under control within 4–6 weeks of spring 2020. (en.wikipedia.org)

4. Why the prediction is scored “ambiguous”
   The factual, unconditional hopes voiced on the show—e.g., “six weeks from now we’ll be in restaurants” in a generally safe scenario—were plainly wrong about how the U.S. pandemic actually unfolded.

However, the normalized prediction you supplied is explicitly counterfactual: it asserts that if the U.S. had implemented a very specific, aggressive antibody‑testing‑and‑zoning regime, then spread could have been brought under control in 4–6 weeks. Since that policy mix was never actually tried, we have no direct empirical test of whether it would have worked. Given that:

• The key precondition (nationwide antibody‑based zoning) never occurred, and
• Expert bodies at the time questioned whether antibody status even corresponded to reliable immunity, undermining but not definitively disproving the scenario,

we cannot definitively say whether the conditional “could be brought under control in 4–6 weeks” is true or false. It remains a hypothetical claim about an unrealized policy.

Bottom line:

• The unconditional timing speculation about being safely back in restaurants in ~6 weeks was wrong in practice.
• The conditional, policy‑dependent claim you asked about—what could have happened if a massive antibody‑testing/zoning program had been implemented—cannot be empirically verified or falsified, so it is best labeled ambiguous rather than right or wrong.

Why this prediction is rated wrong

Jason said founders should assume that “no money is going to get raised for the next 3 or 6 months at a minimum” from March 2020.

Evidence from venture data and sector reports shows that, although deal counts fell and investors became more selective, substantial new fundraising continued throughout that 3–6 month window:

1. Overall US & global VC in Q2 2020 (roughly April–June)

   • PitchBook–NVCA’s Venture Monitor reports $34.3B invested across 2,197 US VC deals in Q2 2020, noting that deal activity slowed but fundraising and investment dollars "remained healthy." (pitchbook.com)
   • KPMG’s Global Venture Pulse finds global VC investment of $62.9B across 4,502 deals in Q2 2020, almost equal to Q1 2020 and only slightly below Q2 2019. The US alone accounts for more than half of that ($34.3B across 2,197 deals). (prnewswire.com)
   • The same KPMG report notes that first‑time venture financings in 1H 2020 totaled $10.2B across 2,439 deals—well below 2019’s pace, but clearly far from “no money.” (prnewswire.com)

2. Sector example: digital health and telehealth

   • Mercom Capital reports global VC funding for digital health companies hit a record $6.3B in 1H 2020, up 24% vs. 1H 2019, with $2.8B in 161 deals in Q2 alone and $690M of that in early (seed/Series A) rounds. (mercomcapital.com)
   • Telehealth specifically saw record investment: Mercom notes nearly $2B into telehealth in 1H 2020, including roughly $962M in 50 deals in Q2 2020. (mercomcapital.com)
     These are exactly the months Jason said to expect “no money” to be raised.

3. Rebound by Q3 2020 (within the 6‑month horizon)

   • The PwC/CB Insights MoneyTree report shows US VC investments reached $36.5B in Q3 2020, a seven‑quarter high and the second‑strongest quarter ever for US VC‑backed companies, up 30% from Q2 2020. (cbinsights.com)
   • Global Q3 2020 funding across North America, Asia, and Europe also hit a historical record, indicating that by months 4–6 after March 2020, fundraising was not only happening but very strong. (cbinsights.com)

4. Concrete deal activity even in March 2020

   • A compiled list of Pacific Northwest venture deals shows 20 startups raising over $200M in March 2020 alone, including multiple Series B and C rounds—evidence that even in the initial shock, new rounds were still closing. (reddit.com)

Conclusion
While COVID‑19 clearly slowed new deals—especially seed and first‑time financings—data from Q2 and Q3 2020 show tens of billions of dollars invested and thousands of deals, including many new rounds. Fundraising became harder and more selective, but it did not effectively cease for 3–6 months. Therefore Jason’s prediction that “no money is going to get raised” in that period is wrong.

Assessment of the prediction
Friedberg’s advice (March 19, 2020) was that startups should do everything possible to extend cash so they had roughly 18 months of runway, implying that the COVID‑19 downturn and funding constraints could last long enough that companies needing to raise sooner would face serious risk.

What actually happened in venture markets

• There was a clear but relatively brief slowdown in venture activity in early 2020, especially at seed and early stage. Seed deals to VC‑backed companies in Q1 2020 dropped 27% vs. Q4 2019 and 43% vs. Q1 2019.(cbinsights.com)
• In Q2 2020, US venture deal count fell materially (e.g., about a 23% decline in total US deal count vs. Q1; North American seed/early/late‑stage deals and dollars were all down vs. Q2 2019), confirming a temporary funding crunch.(prnewswire.com)
• However, by late 2020 overall VC funding was very strong: global 2020 VC investment (~$300.5B) was actually higher than 2019 and the second‑largest year on record; the US hit a record annual VC total.(businesswire.com)
• In 2021 the market moved into an outright boom: global venture funding hit a record ~$621B, with six consecutive quarters of growth and unprecedented numbers of new unicorns.(siliconangle.com)

Runway guidance and risk

• Early in the pandemic, many investors and advisors independently urged startups to extend runway to 18–24+ months because raising in the near term was expected to be very difficult; Gartner, for example, recommended that tech startups with more than six months of cash should extend their runway to at least 18 months to ensure survival.(gartner.com)
• This shows Friedberg’s advice was mainstream risk‑management guidance at the time, but it doesn’t by itself prove that startups without 18 months’ runway were in fact doomed or even generally faced catastrophic failure rates.

Why the outcome is hard to score cleanly

• On one hand, there was a non‑trivial funding crunch in Q2 2020; startups that needed to raise in mid‑2020 with only a short runway did face elevated risk. On the other hand, venture markets recovered quickly and then entered a historic boom in 2021, which undercuts the idea of a prolonged, 18‑month‑plus funding freeze.(ey.com)
• The prediction is phrased in terms of “serious risk of failure” for startups with less than ~18 months of runway—a qualitative condition that is not directly measurable in aggregate data, and for which we lack clear statistics linking 18‑month runway shortfalls in early 2020 to actual failure outcomes.

Because (a) there was a meaningful but relatively short funding shock, (b) capital then became plentiful well before 18 months had elapsed, and (c) the claim hinges on a vague, hard‑to‑verify threshold of “serious risk,” the evidence does not allow a clean true/false judgment. Hence the prediction is best scored as ambiguous rather than clearly right or clearly wrong.

Assessment of the prediction

Chamath’s prediction (made around March 19, 2020) was that over roughly the next six months, LPs and auditors would force significant markdowns in VC portfolios, causing venture capital returns to look terrible.

What actually happened (March–September 2020 and full-year 2020)

1. Only moderate early markdowns, not a later forced wave.

   • Adams Street Partners reported that Q1 2020 venture fund valuations were marked down about 10–15% on average vs. year‑end 2019, with expectations of some additional pressure in Q2, but not catastrophic writedowns. (adamsstreetpartners.com)
   • HarbourVest’s H1 2020 commentary notes “significant markdowns for some portfolio company valuations at the end of Q1,” i.e., immediately after the shock, not a distinct, larger wave six months later driven by LPs and auditors. (investegate.co.uk)
   • A later analysis of private-market valuation behavior finds that from Q2 2020 onward, private equity and venture valuations reacted quickly on the way up, not with extended markdowns, as public markets rallied. (linkedin.com)

2. By Q2–Q3 2020, venture performance was at record highs, not “terrible.”

   • eFront’s global VC performance report shows that in Q2 2020, aggregated TVPI for active VC funds hit an all‑time high of 1.64x, and remained very high (1.63x) in Q3 2020, concluding that it was “difficult to find any impact of the Covid‑19 pandemic on the performance of active VC funds.” (financialinvestigator.nl) These quarters fall squarely within the ~six‑month window after March 2020.
   • An Institutional Investor piece summarizing the same eFront data described 2020 venture performance through Q3 as “exceptional,” with returns at record levels despite the pandemic. (institutionalinvestor.com)

3. Full‑year 2020 VC returns were extraordinarily strong.

   • Cambridge Associates reports that the US Venture Capital Index returned 50.1% in calendar year 2020, its second‑best year ever, and that VC handily outperformed public markets. The same commentary notes record fundraising, strong investment pace, and plentiful exits in 2020. (cambridgeassociates.com) This is the opposite of “returns will look terrible.”

4. Survey evidence: VCs and LPs expected only mild performance impact.

   • A large VC survey (over 900 institutional VCs) analyzed in a Harvard Law School Forum article (Sept 1, 2020) found that VCs expected only a small negative effect on fund IRRs (~–1.6 percentage points) and MOICs (~–0.07), and that the vast majority still believed VC would outperform public markets. The authors conclude that “the most dire predictions of the impact of COVID‑19 on venture capital have not materialized.” (corpgov.law.harvard.edu)

Conclusion

Within six months of March 2020, there were some initial, modest markdowns, but no broad, LP/auditor‑driven wave of forced writedowns. Instead, venture valuations and performance rebounded quickly, and by both mid‑2020 and year‑end 2020, venture capital returns were among the strongest on record, not “terrible.”

Given the aggregate data on valuations and index returns, Chamath’s prediction that LPs and auditors would force large markdowns that made VC returns look very poor did not come true.

Evidence from public markets and venture funding shows that the COVID‑19 shock and related capital‑market disruption in early 2020 were sharp but relatively short‑lived, and that fundraising conditions for startups were very strong well before 36 months had elapsed.

Key points:

1. Public markets recovered in months, not years.

   • The S&P 500 fell ~34% between Feb 19 and Mar 23, 2020, but then rallied quickly; by August 2020 it had already returned to its pre‑crash highs, and by the end of 2020 it posted a positive annual return of about 18%. (nasdaq.com)
   • This indicates that the capital‑market disruption tied directly to the initial COVID shock lasted on the order of months, not a multi‑year freeze.

2. Venture funding rebounded and then hit record highs by 2021.

   • Global VC investment in 2020 was on the order of $335–347 billion, only modestly changed from 2019 and still robust. (news.crunchbase.com)
   • In 2021, global VC funding almost doubled to around $643–671 billion, setting all‑time records; the U.S. alone saw about $330 billion, also a record. (news.crunchbase.com)
   • Reuters and other reports at the time describe 2021 as a record year for venture capital and IPO/exits, not a period of continued severe disruption. (news.crunchbase.com)

   For a startup that had 12–18 months of runway as of March 2020, its next raise would have fallen in mid‑ to late‑2021—precisely when funding and valuations were at or near historic peaks, contradicting the idea that they needed 36 months of cash just to survive a prolonged capital‑markets drought.

3. Contemporaneous VC advice centered on 18–24+ months, not 36 months.

   • Startup‑finance advisors summarizing VC guidance during spring 2020 noted that the pre‑COVID rule of thumb was 18–24 months of runway, and that in response to COVID many VCs began recommending ~24 months or a bit more, not 36 months, to bridge a tough year for new financing. (burklandassociates.com)
   • Sequoia’s famous “Black Swan of 2020” memo (March 2020) urged founders to extend runway and prepare for potentially difficult 2020–2021 fundraising, but did not suggest anything like a 36‑month minimum; the memo spoke of turbulence lasting “several quarters,” not three years. (cnbc.com)

4. Later downturns were driven by different factors.

   • There was another sharp deterioration in late‑stage tech and growth equity markets starting in 2022, driven largely by rising interest rates and a valuation reset, not by renewed COVID‑related disruption. VC funding fell from 2021’s peak but remained substantial in 2022–2024. (barrons.com)
   • Chamath’s prediction, however, was specifically about the duration of disruption stemming from the early 2020 COVID shock and implied that that same disruption would persist long enough that startups needed 36 months’ cash from that point to be safe. The data instead show a V‑shaped recovery in both public markets and venture funding by late 2020–2021.

Putting this together: a startup that entered COVID with an 18–24 month runway in early 2020 generally had access to exceptionally strong fundraising windows in 2021, and did not need to have 36 months of pre‑existing cash just to safely outlast COVID‑era capital‑market disruption. The prediction that “18 months is not sufficient” and that companies would “need at least 36 months” of runway due to prolonged COVID‑driven capital‑market stress is therefore not borne out by what actually happened.

On that basis, the prediction is best classified as wrong.

Why this prediction is evaluated as wrong

1. What Chamath said and in what context In the March 19, 2020 episode, Jason Calacanis asks whether, if the virus is contained quickly ("in the next 60 days") and mitigation drugs work, the markets could "rebound" and whether there is "a chance it could come roaring back". Chamath responds flatly: "It cannot come roaring back" and then explains structural problems in the capital markets. (podscripts.co)

   The normalized prediction you provided – that even if COVID-19 were contained quickly, the U.S. stock market would not see an immediate sharp V‑shaped “roaring back” recovery to prior levels – accurately captures the spirit of his claim: he assigns essentially zero probability to a rapid, V‑shaped recovery in equities from the COVID crash.

2. What actually happened to the U.S. stock market

   • The S&P 500 peaked at a then‑record close of about 3,386 on February 19, 2020, then crashed roughly 34% to a low on March 23, 2020 amid COVID fears. (spglobal.com)
   • From that March 23 low, the index launched an extremely rapid rally, gaining around 50–55% by mid‑August 2020. (benzinga.com)
   • On August 18, 2020, the S&P 500 closed at a new all‑time high (around 3,389–3,395), fully erasing all COVID‑related losses just five months after the bottom, and roughly six months after the pre‑COVID peak. Major outlets and analysts explicitly described this as one of the fastest recoveries from a bear market in history and a “furious” rally. (benzinga.com)

   That pattern – a deep but very short bear market followed by a swift surge back to (and then beyond) prior highs – is exactly what is commonly referred to as a V‑shaped, "roaring back" recovery.

3. About the "even if contained quickly" condition The U.S. did not in fact “contain” COVID‑19 within ~60 days; cases continued to rise through spring 2020 and the pandemic persisted for years. (cdc.gov)

   However, Chamath’s emphatic statement "It cannot come roaring back" was given as a general assessment of the market’s ability to rebound, not narrowly limited to a specific, precisely defined containment scenario. In the real world that unfolded, the outcome he said essentially could not happen – a rapid V‑shaped return to prior highs – did occur.

Conclusion Because U.S. equities in fact staged a historically fast V‑shaped recovery from the COVID‑19 crash, contradicting his claim that the market “cannot come roaring back,” this prediction is best classified as wrong.

Summary

David Friedberg made a compound prediction for the U.S.:

1. RT‑PCR testing backlogs and scaling problems would be “over the first hump” in roughly 7–10 days from March 18–19, 2020 (i.e., by about March 25–29).
2. Broad, general‑population antibody testing would begin roughly 30–45 days after that, i.e., about 45 days from the recording date (late April–early May 2020).

Outcome: the PCR‑testing part was clearly wrong, while the antibody‑testing timeline was roughly right. Because the prediction was framed as a single combined forecast (“so call it 45 days out”), the overall result is best classified as wrong, though with an important correct subcomponent.

1. RT‑PCR backlog & scaling in the U.S. (7–10 days)

What he predicted: that the acute backlog and scaling issues around RT‑PCR testing would be largely resolved within 7–10 days of March 18–19, 2020.

What actually happened:

• A national survey of 323 U.S. hospitals conducted March 23–27, 2020 (right in that 7–10‑day window) by the HHS Office of Inspector General found:
  • “Severe shortages of testing supplies and extended waits for test results” that limited hospitals’ ability to monitor patients and staff.
  • Hospitals frequently waited 7 days or longer for test results, and shortages of swabs, media, and reagents meant hospitals “were unable to keep up with testing demands.” (oig.hhs.gov)
• The same report summarized that hospitals’ “most significant challenges” included testing shortages and long turnaround times, not a resolved situation. (medicaleconomics.com)
• Large commercial labs illustrate the ongoing backlog:
  • Quest Diagnostics reported that as of March 25, 2020 it had a backlog of ~160,000 tests awaiting processing. (orthospinenews.com)
  • In an April 1 media statement, Quest said it ended March with capacity >30,000 tests/day but still had a backlog of 115,000 tests, with average turnaround times of 4–5 days (2–3 days for prioritized patients). (marketscreener.com)
  • An NPR report on April 3, 2020 described Quest’s backlog being cut from 160,000 to 115,000 in late March but still characterized national testing as plagued by backlogs and 4–5‑day result times. (news.wjct.org)

In other words, during and even after the 7–10‑day window, U.S. RT‑PCR testing was still experiencing severe shortages, long delays, and large lab backlogs. Backlogs at major labs were only reported as eliminated around mid‑April, not late March. (marketscreener.com)

Assessment for part (1): The U.S. had not “gotten over the first hump” of PCR testing by March 25–29, 2020. Backlogs and bottlenecks remained severe. This part of the prediction was wrong on timing and severity.

2. Broad general‑population antibody testing (≈45 days out)

What he predicted: that after the PCR “first hump,” it would take another 30–45 days to get antibody tests “more broadly distributed” for general‑population testing, i.e., roughly late April to early May 2020.

What actually happened:

• The first FDA Emergency Use Authorization (EUA) for a COVID‑19 antibody (serology) test was issued to Cellex on April 1, 2020. (fda.gov) This enabled clinical labs to begin serology testing but did not yet constitute broad consumer availability.
• Through April, multiple additional antibody tests (e.g., from Abbott, Ortho, others) received EUAs, creating a growing supply of lab‑based serology tests in the U.S. (en.wikipedia.org)
• Broad, walk‑in consumer access began in late April:
  • On April 28, 2020, Quest Diagnostics launched a consumer‑initiated COVID‑19 antibody test through QuestDirect, explicitly marketed as allowing individuals to purchase an antibody test without visiting a doctor’s office, with blood draws at 2,200 patient service centers nationwide. The company’s release states that this service “broadens access to COVID-19 antibody testing in the United States” and that “individuals can purchase COVID‑19 antibody testing for themselves online.” (newsroom.questdiagnostics.com)
  • Contemporary coverage summarized this as: “For about $120, anyone can now get a COVID-19 antibody test from Quest Diagnostics,” emphasizing no doctor referral and nationwide access via Quest’s centers. (fiercehealthcare.com)
  • User reports from late April and early May 2020 show ordinary people scheduling and obtaining these Quest antibody tests directly, supporting that general‑population access was underway by then. (reddit.com)

Counting from March 19, April 28, 2020 is about 40 days later, squarely inside Friedberg’s 30–45‑day window for the start of broad general‑population antibody testing. The timing and nature of Quest’s national rollout, plus multiple EUA‑backed serology platforms in April, match his description well: not perfect saturation, but the beginning of broad, consumer‑level distribution.

Assessment for part (2): His forecast that broad antibody testing for the general population would start roughly 30–45 days after mid‑March was substantially correct on timing and general character.

Overall judgment

Because Friedberg’s statement bundled these together—first, PCR backlogs clearing in ~7–10 days, then broad antibody testing ~30–45 days later—the PCR component was decisively wrong, while the antibody‑testing timing was quite accurate.

Given that the acute PCR backlog/scaling problem did not resolve within his 7–10‑day window and remained a major national constraint into April, the combined prediction did not come true as stated.

Overall result: wrong, with the important nuance that his antibody‑testing timeline was basically right, but his near‑term optimism on PCR backlogs was not.

Timeframe of the prediction
Chamath made the prediction around March 18–19, 2020, saying that in a worst case the U.S. would have mass testing within 30–60 days. That window runs roughly from April 18 to May 18, 2020.

What U.S. testing actually looked like in that window

• In the week of March 25–April 1, the U.S. was performing about 110,000 tests per day, but hospitals and clinics were still denying tests to people with milder symptoms and often had week‑long result delays, indicating testing was far from broadly accessible. (en.wikipedia.org)
• During the weeks of April 6 and 13, the U.S. conducted about 150,000 tests per day, while experts recommended at least 500,000 tests per day before easing social distancing. A New York Times report on April 26 concluded the U.S. still had not reached an adequate level of testing capacity due to persistent shortages of reagents, swabs, PPE, staff, and lab equipment. (en.wikipedia.org)
• A Harvard‑linked analysis cited in April 2020 noted the U.S. was averaging about 146,000 tests per day and estimated the country needed roughly 500,000–700,000 tests per day by mid‑May just to reopen safely, showing that even by the end of the 60‑day window the U.S. remained well below what experts considered "mass" or sufficient testing. (planetizen.com)
• By early May 2020, U.S. testing had risen to around 240,000–260,000 tests per day, but this level was still described in contemporaneous analyses as inadequate to contain the outbreak. (en.wikipedia.org)
• Per‑capita comparisons reinforce this: as of April 28, FactCheck.org, summarizing Our World in Data and other sources, reported that numerous countries had done more tests per capita than the U.S., underscoring that American testing remained relatively limited despite large absolute numbers. (factcheck.org) And as of May 16, the U.S. had performed about 32.4 tests per 1,000 people, still behind many peers and criticized for having been slow and insufficient in ramping up “crucial testing to contain the spread.” (statista.com)

Interpretation relative to the prediction
Chamath’s claim wasn’t just that the U.S. would increase testing, but that it would reach “mass testing” capability within 30–60 days. By mid‑April to mid‑May 2020, the U.S. had indeed ramped up from almost no testing to hundreds of thousands of tests per day, but:

• Testing was still heavily rationed, with many symptomatic people unable to get tested promptly, and long delays in results. (en.wikipedia.org)
• Major analyses at the time consistently described U.S. testing as inadequate to monitor and contain outbreaks, and far below expert benchmarks for what would count as broad, population‑level testing. (en.wikipedia.org)

Given that contemporaneous public‑health and data sources viewed U.S. testing during April–mid‑May 2020 as insufficient and not yet at the level needed for widespread screening and control, the prediction that the U.S. would achieve mass testing capability within 30–60 days from March 18–19 did not come true.

So the prediction is best scored as: wrong.

Summary of the prediction
Chamath predicted in March 2020 that:

1. “this eurozone is going to collapse” – i.e., the eurozone would cease to function as a meaningful economic/political construct.
2. “Japan is finished” – i.e., Japan would effectively stop mattering as an independent major economic power.
3. Only China and the United States would be the two economies that meaningfully matter globally.

As of late 2025, these components are not borne out by observable facts.

1. Status of the eurozone

• The eurozone (countries using the euro and participating in the EU’s Economic and Monetary Union) is intact and has not collapsed. The European Central Bank (ECB) continues to operate, and the euro remains the world’s second most important reserve currency after the U.S. dollar.
• The eurozone collectively remains one of the world’s largest economic blocs by GDP, alongside the U.S. and China; standard rankings of economies and blocs still list the euro area/EU as a top-tier economic entity rather than a defunct or collapsed system.

While the eurozone has faced challenges (COVID-19 shock, energy crisis after Russia’s invasion of Ukraine, inflation, etc.), none of these have led to dissolution, widespread abandonment of the euro, or a loss of its role as a major global economic center. The core claim of a collapse is therefore false.

2. Status of Japan as a major economic power

• Japan remains one of the largest economies in the world by nominal GDP (typically ranked 3rd or 4th globally, depending on whether you look at euro area as a bloc and how you treat India’s growth). It is still regularly listed in global GDP rankings as a top economy.
• Japan remains central in key global industries and supply chains (automotive, high-end manufacturing, semiconductors/electronics components, robotics, and advanced materials). Major Japanese firms (Toyota, Sony, Mitsubishi, etc.) remain global leaders with significant market caps and global operations.
• Japan is also a core member of the G7, a major holder of U.S. Treasury securities, and a key player in international economic and security arrangements in the Indo‑Pacific region.

Japan faces well-known structural issues (aging population, high public debt, long periods of low inflation/low growth), but it has not ceased to matter as a major independent economic power.

3. Are only China and the U.S. the economies that “meaningfully matter”?

• Global economic influence today is still distributed among several major actors: the U.S., China, the EU/euro area, Japan, and increasingly India, among others.
• International economic policy fora (G7, G20, IMF, World Bank, WTO) and global market behavior all treat the eurozone/EU and Japan as major players whose policies and conditions significantly affect global growth, trade, capital flows, and financial stability.

Thus, it is not accurate to say that only the U.S. and China “meaningfully matter” in the global economic system as of 2025.

Conclusion

Because:

• The eurozone has not collapsed and continues to function as a major economic and political bloc; and
• Japan remains a top-tier independent economic power with substantial global influence; and
• Other economies beyond just the U.S. and China still clearly matter in global economics,

the prediction from March 2020 is not borne out by events up to November 29, 2025.

Result: The prediction is wrong.

Summary Chamath’s statement is essentially counterfactual: that in the years after March 2020, the U.S. could run federal deficits of $5–10 trillion per year without investors losing confidence in the dollar as the world’s primary safety asset. The U.S. never actually tested that range, and the dollar remains dominant (though somewhat eroding) as a reserve and safe‑haven currency. Because the key part of the claim is about capacity under an unobserved scenario, it cannot be cleanly judged right or wrong.

What actually happened to the deficit? Annual federal deficits since the onset of COVID were far below $5–10T:

• FY 2020: ~$3.13T deficit. (home.treasury.gov)
• FY 2021: ~$2.77T. (en.wikipedia.org)
• FY 2022: ~$1.38T. (en.wikipedia.org)
• FY 2023: ~$1.70T. (en.wikipedia.org)
• FY 2024: ~$1.83T. (bidenwhitehouse.archives.gov)
• FY 2025: ~$1.78T (Reuters). (reuters.com)

So while the U.S. ran very large deficits (around $3T at peak), it never approached the $5–10T per year range described in the quote. That means we have no direct empirical test of whether deficits at that scale would have preserved or undermined confidence in the dollar.

What happened to confidence in the dollar? On the other side of the claim, we can check whether confidence in the dollar as the primary reserve/safe asset has collapsed. Evidence shows:

• The U.S. dollar’s share of global foreign‑exchange reserves has drifted down but remains dominant—about 57–58% of allocated FX reserves in 2024–2025, more than double the euro and vastly above the renminbi. (reuters.com)
• A 2025 Federal Reserve note describes “widespread confidence in the U.S. dollar as a store of value,” with the dollar at roughly 58% of global reserves in 2024, far ahead of all competitors. (federalreserve.gov)
• IMF COFER data and related commentary likewise show the dollar still clearly top reserve currency and backbone of international banking liquidity, despite gradual diversification and rising gold holdings. (tomorrowsaffairs.com)
• U.S. Treasuries and dollar assets continue to attract large foreign inflows, and official analyses still frame the U.S. as providing the world’s principal safe asset, even as concerns about fiscal sustainability grow. (investopedia.com)

At the same time, there is visible debate and some erosion:

• Major banks and media highlight an ongoing but gradual “de‑dollarization” trend in central‑bank reserves and commodity pricing, even while emphasizing that the dollar remains dominant. (jpmorgan.com)
• Commentators increasingly warn that erratic U.S. policy, high debt, and tariffs are putting the dollar’s haven status at risk, with episodes where the dollar failed to rally or even fell during geopolitical shocks. (ft.com)

So, confidence has not “broken”—the dollar is still the primary global reserve and safety asset—but there is a gradual, noisy erosion and more open questioning of long‑run dominance.

Why this is ultimately ambiguous Chamath’s normalized prediction is about a hypothetical capacity: that the U.S. will be able to run $5–10T annual deficits without losing the dollar’s safe‑asset status. What we observed instead is:

1. The U.S. did not actually run $5–10T annual deficits—the peak was just over $3T, then deficits settled in the $1.4–1.9T range. (home.treasury.gov)

   • Because the policy scenario he posited never occurred, we do not know whether markets would tolerate sustained $5–10T shortfalls without a sharp confidence shock.

2. Under much smaller (but still huge) deficits, confidence in the dollar largely held.

   • The dollar’s reserve share stayed above ~57%, and it remained the world’s main safe asset and invoicing currency, even after record deficits around $3T during COVID. (home.treasury.gov)
   • This is consistent with his intuition that, at least in crisis, the U.S. could run very large deficits without immediate loss of dollar primacy—but it does not tell us where the tolerance limit lies.

3. There is growing—but not decisive—evidence of slow erosion in dominance.

   • The dollar’s share of reserves edges down, and more trade/commodities are invoiced in other currencies, suggesting some marginal loss of confidence or at least greater diversification. (jpmorgan.com)
   • Yet no alternative currency has displaced the dollar, and the core thesis that the U.S. provides the world’s primary safe asset is still true as of 2025.

Because:

• The crucial $5–10T deficit scenario never occurred, and
• Available data show continued, though somewhat eroding, dominance of the dollar rather than a clear loss of its role as the primary global safety asset,

we cannot definitively say his claim about what the U.S. could do under that untested scenario is either correct or incorrect.

Hence the verdict: ambiguous, not because the time horizon is too short, but because the prediction is counterfactual and not directly testable with realized fiscal and currency data.

Summary: Chamath correctly anticipated a directional shift in rhetoric and policy focus—especially in large economies—toward resilience, industrial policy, and some de‑risking from China. However, the stronger claim that most countries would swing “back towards nationalized economies and away from global economies,” with broad domesticization of food/critical goods and sharply reduced China-dependence, is not borne out by global data as of late 2025. The outcome is mixed rather than clearly right or wrong.

1. Globalization has not broadly reversed

• Global trade as a share of world GDP (“trade openness”) rebounded after the 2020 COVID dip and reached record highs by 2022 (about 63% of global GDP), exceeding pre‑pandemic levels. This indicates continued or even deepened integration rather than a sustained move away from global trade. (ourworldindata.org)
• Across 161 countries, average trade openness in 2023 was ~95% of GDP; high‑income economies averaged over 126%, showing that economies remain highly intertwined internationally rather than “nationalized.” (theglobaleconomy.com)
• The WTO’s World Trade Report 2023 and subsequent WTO commentary state that while geopolitical tensions and some fragmentation exist, evidence of de‑globalization remains “scarce” and global trade “continues to thrive.” (wto.org)
• A 2024 academic review, “Challenging the deglobalization narrative,” using the DHL Global Connectedness Index, concludes that international flows of trade, capital, information, and people have not decreased relative to domestic activity and there is no general deglobalization trend—only targeted reorientations (e.g., Russia vs. the West, partial US–China decoupling). (link.springer.com)

Taken together, the data contradict a clear global swing “away from global economies” toward nationally closed systems.

2. Strong evidence of policy moves toward resilience and self‑reliance in major economies

Where Chamath’s prediction does align with reality is in policy direction in several large economies:

• United States: Since 2020 the US has pivoted to more interventionist, security‑driven industrial policy. A package including the Infrastructure Investment and Jobs Act, CHIPS and Science Act, and Inflation Reduction Act involves large subsidies, export controls, and investment screening, described in the literature as “interventionist techno‑nationalism” aimed at reshaping global value chains and protecting national autonomy. (link.springer.com)
• European Union: The EU Chips Act explicitly aims to reinforce Europe’s semiconductor capacity, reduce dependence on “a limited number of third country companies and geographies,” and build a resilient internal chip ecosystem. (europarl.europa.eu)
  • The Critical Raw Materials Act sets quantitative 2030 targets for domestic extraction, processing, and recycling of critical raw materials (10%, 40%, and 25% of EU consumption respectively) to strengthen “strategic autonomy” and reduce import dependence. (reneweuropegroup.eu)
• India: The Atmanirbhar Bharat (“Self‑reliant India”) framework, launched during COVID, explicitly pursues economic self‑sufficiency and reduced external reliance. Production‑Linked Incentive (PLI) schemes introduced from 2020 cover 14 sectors—electronics, pharmaceuticals, medical devices, autos, specialty steel, telecom, batteries, etc.—with large subsidies to boost domestic manufacturing and exports. (indbiz.gov.in)
  • Defense policy under Atmanirbhar Bharat has indigenised thousands of imported items, reducing reliance on foreign suppliers. (opindia.com)
• Japan and EU energy/tech: Japan’s subsidized push into next‑gen perovskite solar cells is framed as a way to challenge China’s dominance and improve energy and supply‑chain security. (ft.com) The EU’s broader “open strategic autonomy” agenda in industrial policy likewise emphasizes resilience and reduced vulnerability. (link.springer.com)
• Across advanced economies, “de‑risking,” reshoring, nearshoring, and friendshoring entered mainstream policy vocabulary as tools to enhance supply‑chain resilience and economic security post‑COVID and post‑Ukraine. (wealth-db.com)

So the direction of policy—more national focus, resilience, and selective domestic anchoring of critical sectors—is indeed visible, especially among major economies.

3. Shifts away from China are partial and slow, not clean breaks

Chamath also implied less dependence on single‑country suppliers like China. Here the evidence is mixed:

• U.S. trade data show some diversification: by 2024 Mexico had become the top exporter to the U.S., but China still accounted for about $438.9 billion of U.S. imports (13.3% of the total), making it a leading supplier despite very high tariffs and trade tensions. (wsj.com)
• A Federal Reserve analysis finds that U.S. outward FDI has shifted away from China/Hong Kong toward Mexico, India, and Europe, and there are “tentative signs” of reshoring/localization in high‑tech manufacturing “but not more broadly.” (federalreserve.gov)
• Trade composition data suggest U.S., EU, and Japan imports from China have declined in some categories, with rising imports from Mexico, ASEAN (especially Vietnam), and other partners, consistent with friendshoring and “China+1” strategies. (japanpolicyforum.jp)
• However, studies of supply‑chain reallocation show that China’s exports and participation in global value chains remain robust, increasingly through intermediate goods shipped to Asian and European partners, so that many “China+1” suppliers remain tightly tied to Chinese upstream inputs. (arxiv.org)
• Surveys and regional analyses emphasize that “just cutting off the supply chain with China is not easy – it’s almost impossible”; firms often keep production in China when it remains economically compelling, and de‑risking proceeds slowly. (asiasociety.org)
• Even Chinese and foreign commentary notes that while some low‑value production has moved, global production networks frequently still rely on China as a key link, given its cost and infrastructure advantages. (chinadaily.com.cn)

So, while there is a clear policy intent in many major economies to reduce single‑country risk—above all China—actual dependence has been reconfigured rather than sharply reduced, and remains substantial.

4. Food and critical‑goods production: some moves, but no global swing to self‑sufficiency

Chamath highlighted more domestic food and critical‑goods production:

• COVID‑19 and the Ukraine war triggered waves of export restrictions and spurred countries to reassess vulnerabilities in medical supplies, vaccines, fertilizers, energy, and some food staples. WTO and other analyses document expanded export controls on critical raw materials and strategic goods and a greater focus on supply‑chain resilience. (supplychainreport.org)
• India, for example, has linked Atmanirbhar Bharat to targets for fertilizer self‑reliance and broader domestic industrial capacity, and runs campaigns promoting “vocal for local” products. (en.wikipedia.org)
• Yet global food systems remain heavily trade‑dependent. FAO and UN food security reports highlight continued reliance on imports by many low‑income countries and focus on conflicts, climate, and economic shocks as primary drivers of hunger—not a systemic move to food autarky. (apnews.com)
• There is no strong evidence that “most countries” have significantly increased overall food self‑sufficiency; rather, many remain structurally reliant on global markets, as reflected in the persistent high levels of trade openness, especially for low‑income and small economies. (theglobaleconomy.com)

Thus, while some countries made targeted efforts in critical sectors, a broad global swing to domestically supplied food and essential goods has not occurred.

5. “Most countries” and the magnitude of the shift

Finally, the prediction uses strong language—“we’re going to swing the pendulum back towards nationalized economies and away from global economies” and emphasizes “most countries.”

• Academic and WTO evidence both stress that, at the global level, the world has not entered a deglobalization phase; instead, globalization remains high, with selective fragmentation around specific geopolitical fault lines. (link.springer.com)
• There is large cross‑country variation: a handful of big economies (US, EU, China, India, Japan) have adopted more assertive industrial and security‑oriented policies, but many middle‑ and low‑income countries lack the fiscal and institutional capacity to emulate extensive reshoring or “nationalized” strategies and remain deeply trade‑dependent. (link.springer.com)

So while some large economies have indeed shifted their policy stance toward resilience and partial re‑nationalization of key sectors, this falls short of a documented, broad global move in which “most countries” fundamentally pivot away from globalization.

Overall assessment

Correct in spirit but overstated in scope and magnitude.

Chamath accurately foresaw:

• The rise of resilience, de‑risking, and strategic autonomy as central policy themes after COVID;
• Expanded domestic and allied production in critical sectors (chips, energy, defense, some medical supplies);
• Growing concern over over‑reliance on China and other single‑country suppliers.

However, data and major analyses show that:

• Globalization has not reversed; trade openness is at or near record highs and international flows remain robust. (ourworldindata.org)
• Moves toward nationalized or self‑sufficient economies are partial and concentrated in a subset of large economies and sectors; they do not characterize “most countries.” (link.springer.com)
• Dependence on China has been reconfigured rather than sharply reduced, with enduring deep linkages in many supply chains. (arxiv.org)

Because there is substantiated movement in the forecasted direction but not at the global scale or intensity implied, the prediction cannot be cleanly labeled right or wrong. The most accurate rating is “ambiguous.”

What the prediction claimed (normalized)

• After the 2020 COVID shock, Apple would be pushed by policy/public pressure to:
  1. Relocate a significant portion of iPhone production away from China to the United States, or
  2. At least to a more diversified set of countries with a materially higher share in the U.S.
• Apple would accept lower margins in exchange for greater supply‑chain resiliency.

What actually happened by late 2025

1. iPhone production is still overwhelmingly in China, not in the U.S.

   • Multiple recent analyses estimate that around 80–90% of iPhones are still assembled in China, with India and other locations accounting for only the remaining 10–20%. (coincentral.com)
   • A 2025 policy analysis notes that Apple currently has “no meaningful iPhone production” in the United States and effectively 0% of iPhones for commercial sale made there, with roughly 80% of production in China and about 20% in India. (faf.ae)
   • Discussion of U.S. manufacturing repeatedly emphasizes that large‑scale iPhone production in the U.S. remains economically and logistically infeasible, not something Apple has begun doing. (forums.appleinsider.com)

2. Diversification has been to India and other Asian countries, not to the U.S., and China is still dominant.

   • Apple has been following a “China + 1” strategy: adding India (and some Vietnam) as secondary hubs while keeping China as the core manufacturing base. (archive.demodaenperu.com)
   • Apple announced and pursued plans to move a portion of iPhone 14 and later models’ production to India and has targets such as up to ~25% of iPhones from India by mid‑decade, but that still leaves the majority in China and does not materially raise the U.S. share, which remains near zero. (en.wikipedia.org)
   • In 2025 Apple’s strategy for the U.S. market is to shift U.S.-bound iPhone assembly from China to India by 2026—explicitly India, not the United States. (theguardian.com)

3. Margins went up, not down.

   • Apple’s overall gross margin has risen steadily from about 38% in FY 2019–2020 to roughly 46–47% by FY 2024–2025. (valueinvesting.io)
   • Net and operating margins likewise remained strong or improved. Apple did not accept structurally lower margins as a trade‑off for resilience; instead, it preserved or increased profitability while modestly diversifying production.

Assessment versus the normalized prediction

• Relocating a “significant portion” of iPhone production to the U.S. / materially raising the U.S. share: This clearly did not happen; U.S. iPhone manufacturing is effectively zero, and China still produces the vast majority of units. (faf.ae)
• Major diversification away from China with the U.S. as a key alternate base: Some diversification to India occurred, but China remains dominant, and the U.S. share is not materially higher—still near zero—so this condition is not met. (faf.ae)
• Accepting lower margins for resiliency: Apple’s margins increased meaningfully after 2020; they did not fall as the prediction suggested. (neyman.ai)

Because the core, testable components of the normalized prediction (substantial production relocation to or heavily involving the U.S., and lower margins as a trade‑off) did not occur by 2025, the prediction is wrong.

Bottom line: The predicted broad, long‑term deregulation did not materialize. There were some temporary COVID-era flexibilities and one short‑lived diagnostic testing deregulatory move, but these were reversed or time‑limited, and Congress did not pass the kind of sweeping reforms described.

(a) Right‑to‑try and experimental drugs

• Federal right‑to‑try already existed before COVID-19. A federal Right to Try Act was enacted in 2018, allowing certain terminally ill patients, under state law, to access investigational drugs that had completed Phase I trials without FDA approving each individual use. (en.wikipedia.org)
• Post‑2020, there was no major expansion or “new” federalization. After 2020, Congress introduced only narrow Right to Try Clarification Act bills (e.g., to clarify access to Schedule I substances under the existing law), but these bills stalled at the “introduced” stage and did not become law. (congress.gov)
• FDA’s role remains limited but not eliminated, and usage is tiny. FDA’s own guidance emphasizes that under the Right to Try Act its role is mostly to receive annual summaries; FDA does not review or approve individual Right‑to‑Try requests, but the pathway applies only to a narrow set of eligible patients and investigational drugs, and manufacturers are not required to provide drugs. (fda.gov) Available analyses note that very few patients have actually used the Right‑to‑Try pathway, with physicians and sponsors overwhelmingly preferring the existing FDA Expanded Access route. (en.wikipedia.org)

Net effect: there was no new, major, post‑COVID shift that “federalized” right‑to‑try in a dramatically more permissive way or broadly allowed doctors and patients to use experimental drugs “on their own discretion without a federal oversight body” beyond what the 2018 law had already done.

(b) Diagnostics and testing regulation

• COVID emergency measures did temporarily speed test deployment. Starting in early 2020, FDA used Emergency Use Authorizations (EUAs) to let COVID‑19 tests reach the market much faster than under normal review, a deliberate emergency‑only flexibility. (pew.org) Those EUA‑based flexibilities and special enforcement policies were explicitly tied to the COVID‑19 emergency and have since been wound down as the public health emergency expired in May 2023. (fda.gov)
• A short‑lived deregulatory move on lab‑developed tests (LDTs) was reversed. In August 2020, HHS under the Trump administration announced that FDA could not require premarket review of LDTs—including COVID‑19 tests—without formal notice‑and‑comment rulemaking, effectively rescinding prior informal FDA LDT guidance and allowing labs to use many tests without EUA. This was framed as part of a broader COVID‑related deregulatory push. (ropesgray.com)
  • In November 2021, the Biden administration withdrew that policy and explicitly restored FDA’s traditional approach, again expecting COVID‑19 LDT developers to seek EUAs or other authorization. (ascp.org)
• The medium‑term trajectory has actually been toward more oversight, not less. In 2024 FDA finalized a rule to increase federal oversight of LDTs, phasing out enforcement discretion and requiring FDA review for many new tests. (reuters.com) Although this rule was later vacated by a federal court and is being rescinded in 2025, the action merely blocks an increase in regulation; it does not create a durable, new pro‑innovation deregime beyond the pre‑2020 status quo. (reuters.com)
• Congress did not enact enduring diagnostic deregulation. Bills such as the VITAL Act, which would have shifted LDT oversight firmly away from FDA to CMS (and were explicitly justified in light of COVID‑19 testing), were introduced but never passed. Conversely, the VALID Act—aimed at strengthening and systematizing FDA oversight of diagnostics—came close to passage but ultimately failed in 2022. (congress.gov)

Overall, while emergency EUAs and the brief 2020–2021 HHS LDT policy did temporarily make COVID‑era testing easier to deploy, those changes were either emergency‑only or later reversed. The enduring regulatory picture by 2025 is not one of broadly and permanently “much faster deployment” via relaxed federal oversight.

Verdict

Because:

• No major new post‑2020 federal reform has broadly “federalized” right‑to‑try or removed FDA from oversight of experimental drugs beyond the 2018 law (whose pathway remains narrow and rarely used), and
• The net long‑run trajectory for diagnostics/testing after COVID includes only temporary flexibilities, a brief Trump‑era deregulatory step that was reversed in 2021, and an attempted tightening of oversight in 2024 (even though that rule was vacated),

the prediction that “within a few years after the COVID‑19 crisis” the U.S. would significantly and durably relax federal healthcare regulation along those lines is best classified as wrong.

Chamath’s prediction was that within six weeks of March 17–18, 2020 (by late April 2020), some U.S. cities or towns would set up “green zones” where people who could show a negative COVID‑19 test or positive IgG/IgM antibody test plus ID would be allowed to interact and dine in restaurants.

What actually happened in the U.S. March–April 2020

• From mid‑March 2020 onward, most U.S. states issued stay‑at‑home orders that closed dine‑in restaurants, allowing only takeout and delivery; this was the dominant pattern through April 2020. (en.wikipedia.org)
• State timelines show that stay‑at‑home orders were still in effect through most or all of April in many states (e.g., New York, Illinois, Michigan, etc.), with formal reopenings generally not beginning until late April or early May. (en.wikipedia.org)
• Early reopenings that did begin near the end of April (e.g., Georgia and Texas) allowed restaurants to resume dine‑in service under strict capacity limits, spacing, and mask/sanitation rules—not based on individual test or antibody status. Georgia’s April 27, 2020 guidelines capped patrons per square footage and mandated employee masking and distancing, with signage asking symptomatic people to stay away, but did not require showing test results or antibodies. (atlanta.eater.com) Texas’s May 1, 2020 reopening similarly limited restaurant capacity (25%, with possible 50% in some rural counties) and described general health protocols, again without any per‑person testing or antibody checks for entry. (cnbc.com)

Status of “immunity passports” and green‑zone ideas

• In spring 2020, “immunity passports” based on antibody tests were being proposed and debated in the U.S., not implemented. For example, a May 18, 2020 ACLU analysis discusses immunity passports only as a hypothetical proposal, warning of civil‑liberties and equity problems if such a system were adopted in the United States. (aclu.org)
• The World Health Organization, in April 2020, explicitly warned against issuing “immunity passports” or “risk‑free certificates” based on antibodies, stating there was insufficient evidence that antibodies conferred reliable immunity; this further discouraged governments from adopting such schemes at that time. (reviewjournal.com)
• Later in the pandemic, some U.S. jurisdictions did introduce vaccine or vaccine/negative test requirements for indoor dining (e.g., New York City, San Francisco, New Orleans), but these appeared in mid‑ to late‑2021, were based primarily on vaccination, and were not framed as localized April 2020 “green zones” built around antibody tests. (congress.gov)

Across contemporary reporting, government orders, and retrospective summaries of early‑2020 U.S. COVID policy, there is no evidence that any U.S. city or town, by late April 2020, implemented a system matching Chamath’s description: geographically bounded “green zones” where restaurant access was contingent on showing a recent negative viral test or positive antibody test plus ID. Instead, the reality was blanket closures followed by capacity‑limited reopenings without individual immunity‑based screening.

Conclusion: Since the specific scenario he forecast did not materialize within the six‑week window (or at all in that form in the U.S.), the prediction is wrong.

Evidence from state-level reopening orders shows that by late April 2020 (about six weeks after March 17–18, 2020), people in some parts of the United States were again allowed to eat inside restaurants, with health‑screening requirements in place.

Key points:

• Timeframe match: Six weeks after March 17–18, 2020 falls in late April 2020. Jason’s statement (“Six weeks from now will be in restaurants”) was made around March 19, 2020, so the target window is roughly the last week of April 2020.

• Alaska: indoor dining from April 24, 2020 – The U.S. COVID-19 timeline and Alaska’s pandemic chronology show that Alaska allowed restaurants to reopen for dine‑in service at 25% capacity with tables 10 feet apart on April 24, 2020. (en.wikipedia.org) This is within the six‑week window.

• Georgia: indoor dining from April 27, 2020 with health screening – Georgia’s detailed reopening order allowed restaurants and dining rooms to resume dine‑in services starting April 27, 2020, subject to multiple safety measures. These included a requirement that restaurants “screen and evaluate workers who exhibit signs of illness, such as a fever over 100.4°F, cough, or shortness of breath” and keep them from working, along with capacity limits and mask requirements for staff. (littler.com) These provisions amount to explicit health checks (symptom and temperature screening) associated with the reopening of indoor dining.

• Scope of prediction: The normalized prediction only requires that “people in at least some parts of the United States” be allowed to eat inside restaurants by that time, with health checks such as temperature or symptom screening. The Georgia and Alaska examples satisfy this: indoor dining was permitted again in those states in the late‑April window, and at least in Georgia it was tied to formal employee health screenings.

Because indoor restaurant dining did resume in some U.S. jurisdictions by late April 2020 and was accompanied by health‑screening measures, Jason’s six‑week prediction is best judged as right.

Evidence shows that the San Francisco Bay Area and the broader U.S. West Coast had not begun reopening by April 7–8, 2020, and strict shelter‑in‑place rules (including closure of restaurant dining rooms) were still in force.

San Francisco Bay Area:

• The original Bay Area shelter‑in‑place order issued March 16, 2020 was set to run until April 7.(sf.eater.com)
• On March 30–31, officials extended this order; San Francisco announced the stay‑home order would now last until May 3, explicitly rolling past the April 7 date.(sf.eater.com)
• Restaurants remained limited to takeout and delivery only; dining rooms stayed closed, and broader reopening (even of some outdoor businesses) didn’t begin until May and was further extended to at least May 31 while restaurants were still restricted.(sf.eater.com)

Broader U.S. West Coast:

• Washington State: The statewide “Stay Home – Stay Healthy” order closed non‑essential businesses and limited restaurants to takeout/delivery only. Initially through early April, it was then extended through May 4 and again through May 31, meaning no general reopening or dine‑in restaurant service by April 7–8.(capaa.wa.gov)
• Oregon: A statewide stay‑at‑home order on March 23, 2020 kept non‑essential businesses closed; the reopening framework allowed in‑person dining only starting with Phase 1, which counties did not begin entering until mid‑May 2020.(cnbc.com)

Because neither the Bay Area nor the broader West Coast had begun the kind of reopening described (people going back out to restaurants, end of strict shelter‑in‑place) by April 7–8, 2020, Friedberg’s prediction that this would occur by that date was wrong.

The prediction hinged on three concrete claims about early April 2020:

1. Dramatic decline in West Coast COVID metrics by early April 2020
   While the Seattle area’s growth rate of new cases had begun to slow by late March, it was still among the metros with the highest incidence, and King County hospitals were described as under “significant pressures” as of March 29, not in a state of dramatic relief. (cityobservatory.org) In California, large counties such as Los Angeles continued to see rising cumulative cases and deaths through the first half of April, indicating continued spread rather than a sharp drop in burden. (en.wikipedia.org) This contradicts the idea that ICU use, ER strain, and new cases had dramatically declined by April 7–8 across the U.S. West Coast.

2. West Coast “reopen for business” around April 7–8, 2020
   Policy actions directly oppose this.

   • Washington: The statewide “Stay Home – Stay Healthy” order issued March 23 initially ran to April 6, but on April 2 it was formally extended to May 4, and later to May 31 with a phased reopening plan. (housedemocrats.wa.gov)
   • California: A mandatory statewide stay‑at‑home order began March 19 and remained in effect with no early‑April end date; as of late April, counties were still under extended or indefinite stay‑home orders, with only very limited easing (e.g., curbside retail) starting in May. (en.wikipedia.org)
   • Oregon: Issued a stay‑at‑home order in late March and, together with California and Washington, formed the Western States Pact on April 13 explicitly to coordinate a gradual, data‑driven reopening later; Oregon’s order was extended into summer with phased county reopenings beginning mid‑May. (en.wikipedia.org)
     In reality, none of the major West Coast states “reopened for business” around April 7–8; all were still under stringent stay‑at‑home regimes and planning extended shutdowns.

3. Infection fatality rate (IFR) would prove materially lower than prevailing projections by that time
   Early public numbers in February–March 2020 often cited crude case‑fatality rates around 3–4% from China, which were clearly overestimates of IFR. (pmc.ncbi.nlm.nih.gov) However, more sophisticated models emerging in March 2020 (e.g., Imperial College) were already assuming an IFR around 0.9–1% for high‑income countries. (brownstone.org) Later meta‑analyses and modeling work converged on an overall IFR in the roughly 0.5–1% range in high‑income settings, consistent with those early model‑based projections rather than “materially lower” than them. (imperial.ac.uk) Moreover, by early April 2020 there simply wasn’t enough serological data from the U.S. West Coast to demonstrate a clearly lower IFR at that time.

4. New York “two weeks behind” and under strain longer
   New York state did become the U.S. epicenter, with daily new cases peaking in early April and enormously stressed hospitals and excess mortality through that month. The statewide NY PAUSE order, initially effective from March 22, was extended on April 6 to April 29 and again on April 16 to May 15, confirming severe, prolonged strain well beyond early April. (en.wikipedia.org) This relative comparison (NY worse and later) was broadly accurate, but it was tied to the incorrect premise that the West Coast would already be in a position to reopen by April 7–8.

Net evaluation
The core actionable part of the prediction—that West Coast metrics would dramatically improve by early April 2020, allowing the region to reopen around April 7–8 while New York lagged—conflicts with both epidemiological data and actual policy. Stay‑at‑home orders on the West Coast were extended well past that date, cases and deaths were still accumulating, and the IFR ultimately aligned with serious early model‑based projections rather than being clearly and materially below them.

Therefore, the prediction as stated is best judged: wrong.