E62: Elizabeth Holmes verdict, fraud origins & takeaways, navigating "The Great Markdown" & more

As of late 2025, Jason’s long‑range prediction ("maybe in our lifetime…we’re talking about decades") can’t be definitively scored.

On the Theranos-style, point-of-care vision he was reacting to — a small consumer device that takes a single finger‑prick and returns hundreds of routine clinical lab results (CBC, chemistries, hormones, etc.) with central‑lab accuracy — there is still no FDA‑cleared, widely deployed system. Theranos itself failed, and its later miniLab work never delivered hundreds of tests from a finger stick.(forbes.com) Other finger‑prick platforms like Sight Diagnostics’ OLO and similar devices focus on narrow panels (e.g., a complete blood count) rather than hundreds of different assays from a single tiny capillary sample.(forbes.com) Even ambitious microfluidic systems such as rHEALTH emphasize that their devices remain investigational and are currently being advanced around specific assays (e.g., factor VIII levels, platelet counts, COVID antigen tests, 15‑parameter blood counts), not a fully validated menu of hundreds of routine tests from one drop.(rhealth.com) Laboratory experts also continue to flag intrinsic issues with capillary/finger‑stick blood (limited volume, contamination with interstitial fluid), which make large numbers of highly precise quantitative tests from a single prick technically challenging for point‑of‑care use.(wired.com) In that practical, clinical sense, nothing since 2022 has disproved his skepticism.

However, his statement is broader: he framed such high‑throughput diagnostics from a single small sample as not technologically feasible for many decades and “maybe in our lifetime,” absent a major breakthrough. In research and central‑lab settings, the picture is more optimistic than that:

• Multiplex proteomic platforms (e.g., Olink and SomaLogic) can already quantify hundreds to thousands of proteins from microliter‑scale blood samples. Olink’s panels measure up to 3,072 proteins from dried blood spots (which are typically obtained from small capillary pricks), and prior work has shown 96‑plex assays from 1 µL of sample.(olink.com)
• Recent studies have successfully performed high‑throughput proteome profiling on home‑sampled finger‑prick dried blood spots, analyzing hundreds of proteins per sample to derive clinically relevant information (e.g., about SARS‑CoV‑2 infection status).(pubmed.ncbi.nlm.nih.gov)
• Devices like rHEALTH and its NASA‑backed variants are explicitly designed to run diverse assay classes from a single ~8–20 µL sample and have demonstrated broad multiplexing in research and spaceflight contexts, even if they are not yet commercial general‑purpose analyzers.(rhealth.com)

Taken together, these developments show that in principle high‑throughput, multi‑analyte diagnostics from tiny blood volumes are already technically achievable in controlled settings, which weakens the claim that such capabilities are fundamentally “many decades” away at the technology level.

Because Jason’s prediction is explicitly about a time horizon spanning "decades" and "maybe" our entire lifetime, and we are only a few years past his 2022 statement, there has not yet been enough time to know whether he is ultimately right or wrong about that long‑term forecast. Current evidence simultaneously supports his skepticism about near‑term, Theranos‑style point‑of‑care devices and contradicts his implication that the underlying technology is nowhere close. That combination means we cannot yet give a definitive verdict, so the fairest classification is inconclusive (too early to tell).

Market data for calendar year 2022 shows a large, sustained correction in high‑growth and tech‑related asset prices relative to their 2020–2021 peaks:

• The tech‑heavy Nasdaq Composite fell about 33% in 2022, its worst annual performance since 2008, driven largely by sharp selloffs in high‑growth technology names.【turn0calculator0】
• Popular high‑growth/"innovation" ETFs such as ARK Innovation (ARKK) suffered even steeper drawdowns: ARKK was down roughly 67% in 2022 and over 80% from its February 2021 peak, reflecting a severe repricing of unprofitable and high‑multiple tech stocks (widely reported in financial press and fund performance summaries).
• Many individual high‑growth tech stocks (e.g., pandemic beneficiaries in software, e‑commerce, fintech) declined 60–90% from 2020–2021 highs during 2022, as documented in year‑end 2022 market reviews and drawdown tables across major financial outlets.

Taken together, 2022 was clearly characterized by a broad, deep, and sustained correction in high‑growth and tech‑related asset prices from the elevated valuations of 2020–2021. This matches the substance of Sacks’s prediction that 2022 would be a “year of the correction” in those assets. Therefore, the prediction is best judged as right.

The prediction tied a specific near‑term path to bond volatility, inflation, rates, and risk assets:

1. Condition on 10‑year Treasury volatility: Chamath’s scenario required Treasury yield volatility to remain subdued (“if that continues to hold…”). In reality, bond‑market volatility picked up: by late March 2022, the MOVE index (a standard measure of Treasury volatility) had jumped roughly 25%, reflecting renewed turbulence in yields rather than continued calm.(reddit.com) So even the stated precondition didn’t really hold.

2. “100 bps of hikes will wash most inflation through the system”: Instead of a modest 100 bp hiking cycle, the Fed raised rates from near zero in March 2022 to 4.25–4.50% by December 2022—about 425 basis points of tightening in 2022 alone, continuing into 2023.(orangebybk.com) Inflation did not quickly fade after the first 100 bps: CPI rose from 7.5% year‑over‑year in January 2022 to a 9.1% peak in June 2022 and stayed above 6% through late 2022, only gradually easing in 2023.(bls.gov) That is the opposite of “most of it washed through” by a small initial hike.

3. Market path: brief Q1 2022 pullback then “back to the races”: Q1 2022 itself ended down: the S&P 500 returned about ‑5.2% for the quarter, and the Nasdaq Composite about ‑9.6%.(statmuse.com) Rather than a quick dip then renewed bull market, selling continued: for full‑year 2022 the S&P 500 fell about ‑19.4%, its worst year since 2008,(spglobal.com) and the Nasdaq had deeply negative returns in Q2 and remained down each quarter of 2022.(statmuse.com) Risk assets did not simply “get back in” after a short Q1 wobble; they went through a sustained bear market before recovering in 2023.

Because (a) the low‑volatility condition he described didn’t persist, (b) inflation clearly was not largely resolved by ~100 bps of hikes, and (c) markets did not experience only a brief Q1 pullback followed by an immediate, durable rebound in risk assets, the prediction does not match what actually happened.

Available data through late 2025 strongly support Friedberg's macro call that hundreds of unicorns valued at 2020–21 peaks would not be able to float at comparable valuations in public markets over the following 3–5 years. By 2021 there were over 800 global unicorns, and by mid‑2022 more than 1,000 were piled up in the private markets; TechCrunch described a ‘unicorn glut’ where many 2021-era unicorns were mispriced relative to public comps and noted that even if the IPO window reopened, a large portion of these companies would still be unable to exit at their old valuations. (en.wikipedia.org) A 2024 Economist summary reported that 344 unicorns were minted in 2021 and that about 700 unicorns with a combined $2.4 trillion valuation still had not IPO’d or been acquired, with investors reluctant to fund them at lofty valuations and instead ‘mulling how to sell their stakes’—evidence of a large backlog of companies whose private prices the public markets will not currently bear. (edwardconard.com) Bloomberg reporter Katie Roof’s 2025 analysis of the 2021 cohort found that of 354 VC-backed firms newly valued over $1 billion in 2021, only 6 had gone public via IPO and 4 via SPAC, fewer than 30% had raised any new capital in three years, and almost half of those did so in down rounds; she described this as an ‘era of the zombie unicorn’ where many former unicorns had become ‘unicorpses’. (edwardconard.com) Among the unicorns that have listed since 2022, IPO valuations are typically far below their peak private marks: Instacart went public at about $9.9 billion versus a roughly $39 billion private valuation in 2021, and Klarna listed around $15 billion after having been valued at about $45–46 billion at the height of the BNPL boom. (fortune.com) PitchBook data summarized in 2025 show that almost every major IPO in Q2 2025 came to market below its prior peak valuation, with examples like MNTN, Circle, Hinge, and Chime all going public at steep discounts to their earlier private highs; Investopedia likewise notes that most of the tech unicorns that IPO’d in 2025 did so below their peak private valuations. (fortune.com) At the same time, industry data from Carta and PitchBook show elevated down-round rates at late-stage and unicorn levels, confirming that many startups have had to accept lower valuations rather than sustain 2021 pricing. (carta.com) Taken together, these facts show that (a) there are indeed many hundreds of unicorns still stuck in the private markets at valuations public investors won’t currently validate, and (b) those that have tested the public markets in 2022–25 almost always did so at large haircuts versus their boom-time valuations. Although we cannot observe every hypothetical IPO, the realized IPO discounts, pervasive down rounds, and large backlog of ‘zombie’ unicorns align closely with Friedberg’s claim that, over the 3–5 years after early 2022, hundreds of highly valued private companies would be unable to achieve their peak private valuations in the public markets, so his prediction is best judged as right.

Evidence since 2022 shows that a large share of unicorn and late‑stage startup exits have indeed been flat or down relative to their last private valuations, leaving late‑stage investors with little or negative return – consistent with Jason’s “lots of pushes” prediction.

1. Macro: many exits now price below the last private round

• Silicon Valley Bank’s 2025 State of the Markets data (summarized by SaaStr/Jason Lemkin) reports that 57% of 2024 VC‑backed tech IPOs priced below their last private valuation, i.e., they were down‑round IPOs rather than step‑ups. That is explicitly framed as a normal but now very common outcome, contradicting the 2020–21 mentality that every IPO would be an up‑round. (linkedin.com)
• Earlier research on unicorn exits by Stanford’s Ilya Strebulaev shows that average exit valuations dropped sharply after the 2021 peak, with 2022 exits averaging about $1.5B vs much higher averages in 2019–21, reflecting a more muted exit environment for unicorns that reached peak valuations during the bubble. (linkedin.com)

2. Concrete unicorn examples from the 2020–21 bubble cohort
Several flagship 2021–22 unicorns have exited at or well below their last private valuations:

• Instacart raised at a $39B private valuation in 2021, but IPO’d in 2023 around $8B, leaving many late‑stage investors underwater on that round. (notescast.com)
• Reddit last raised at $10B in 2021 and went public in 2024 at up to $6.4B, which the Financial Times explicitly describes as part of a “growing trend of ‘down round’ IPOs” among Silicon Valley startups. (ft.com)
• Klaviyo IPO’d at about $9B, a small but real dip from its 2021 private valuation, and is cited alongside Instacart and Reddit as an example where IPO prices were at or below peak private marks, leaving late‑stage equity with little upside. (linkedin.com)
• Navan (TripActions) reached a $9.2B valuation in a 2022 round but IPO’d in 2025 at roughly $6.2B, another prominent unicorn whose public valuation came in materially below its last private round. (businessinsider.com)
• Klarna peaked at $45.6B in 2021 but eventually listed at about $15.1B, a massive down exit from its bubble valuation. (businessinsider.com)

These are all part of the early‑2020s unicorn cohort Jason was talking about; in each case the exit valuation is at or far below the last private valuation, matching his notion that many exits would be “pushes” or worse for late‑stage capital.

3. Late‑stage investors often just get capital back or lose money

• The FT notes that these down‑round IPOs “could lead to significant losses for late‑stage investors who fueled the 2021 bubble”, explicitly saying those investors gave up protections and are now taking hits when companies list below their last private valuations. (ft.com)
• Commentary from the All‑In crew and others on Instacart and Klaviyo emphasizes that late‑stage 2021 investors are underwater at IPO pricing, whereas early investors still make strong multiples – exactly the distribution Jason described, where late‑stage rounds mainly see capital back (or less) while earlier rounds capture most of the gains. (notescast.com)
• SVB’s and SaaStr’s analysis goes further, arguing that 80–90% of large 2020–22 late‑stage SaaS rounds done at 15–20× ARR are “structurally impaired” and will never grow into those valuations, implying that many eventual exits will cluster around or below those peak prices. (saastr.com)

4. Broader valuation reset for the ~2021–22 unicorn herd

• TechCrunch/Secfi data show that in 2022 about 24% of startups on Secfi’s platform cut their valuations, and that late‑stage startups were already seeing far more flat and down rounds than before—an early sign that many unicorn valuations from 2021 could only be realized, at best, as pushes. (techcrunch.com)
• European data from PitchBook and the FT estimate that the continent’s unicorns were overvalued by roughly €100B (over 20% of aggregate value) post‑2021, with many now failing to meet the $1B threshold and a high share of new rounds being raised at lower valuations. (ft.com) This reinforces the idea that a large slice of the global ~900–1,000‑company unicorn pool is facing exits or follow‑on financings near or below prior marks, rather than big step‑ups.

Assessment

• Jason’s claim was qualitative (“a lot of pushes”), not a precise percentage. By late 2025 we have: (i) multiple marquee unicorns from the 2021–22 cohort exiting below their last private valuations; (ii) system‑level data showing that a majority of recent VC‑backed tech IPOs are down rounds; and (iii) broad evidence that late‑stage 2020–22 vintages are impaired, with many investors getting at best modest or zero upside on those peak rounds.
• While most of the ~900 early‑2022 unicorns have not yet exited, the exits we do see, plus aggregate IPO statistics and valuation data, are exactly in the direction he described. In many cases the reality has been harsher than a simple “push,” with clear capital loss for late‑stage rounds.

Given that, the prediction that many of the unicorn exits from that cohort would be pushes (or worse) for late‑stage investors has clearly materialized in the observable data to date, so classifying the forecast as right is justified, even though the story is still playing out for the remaining private companies.

There isn’t enough realized exit data yet on the 2020–2022 late‑stage/$1B+ cohort to say whether “a significant share” have ultimately produced only capital-back outcomes for their last-round investors.

Key points:

• The unicorn herd that existed around early 2022 is still largely un-exited. A TechCrunch/Aileen Lee update in Jan 2024 notes that about 93% of current unicorns are still “papercorns” (privately valued with no exit yet), and only about 7% had exited versus ~66% at a similar stage a decade earlier.(techcrunch.com) That means we simply don’t observe final exit prices for most of the companies Jason was talking about.
• Stanford’s Venture Capital Initiative (Ilya Strebulaev) finds that, across a large sample of U.S. unicorns over multiple cycles, the average unicorn exit still returns a large multiple of capital raised (around 31x), and only about 6% of unicorns in their dataset exited below the total capital they had raised.(linkedin.com) That historical pattern does not yet show a broad regime where most late-stage investors merely get their money back.
• However, more recent data do show serious stress in the 2021–2022 vintages: Carta and WSJ report that more than 90% of U.S. venture funds raised in 2021 have made no distributions three years in, with median IRRs negative, largely because exits collapsed after the 2021 boom.(wsj.com) This supports Jason’s directional concern (late-stage deals from that period are underperforming expectations) but doesn’t yet quantify realized exit multiples.
• Post‑COVID unicorn exit multiples have compressed sharply (Stanford data show exit‑to‑capital ratios falling from ~15x in 2021 to roughly 4–6x in recent years), but even those lower figures are still above mere return-of-capital levels.(linkedin.com) Again, this is at the whole‑company level and not specifically for the last private round.
• Secondary-market and private-mark data indicate that a large share of unicorns are now valued below both $1B and their peak private valuations (e.g., ~40% of U.S. VC‑backed unicorns trading below $1B in secondaries; over 80% trading below their 2021–2022 highs), showing that many late-stage positions are underwater on paper.(economictimes.indiatimes.com) But that still does not tell us how many ultimate exits will clear at or just above the last round’s price versus recovering to provide 3–4x returns.

Because:

• the majority of the specific 2020–2022 $1B+ cohort Jason referred to have not yet exited,
• the available research aggregates over many vintages and does not break out realized multiples for that narrow late‑stage slice, and
• current evidence is dominated by interim markdowns rather than final cash outcomes,

we cannot reliably say what share of those late‑stage unicorn deals will end as “pushes” (capital back only) versus 3–4x+ winners. The environment is consistent with his warning being plausible, but the necessary outcome data are not yet observable, so the prediction is best classified as inconclusive (too early to call) rather than clearly right or wrong.

The prediction was that, after the Fed’s early‑2022 hawkish shift, the risk (probability) of a U.S. recession during calendar year 2022 was now “much greater” than before. This is a statement about changes in probability, not a direct forecast that a recession would definitely occur.

What we can check:

• The Federal Reserve did, in fact, pivot to aggressive tightening starting in March 2022, with rapid rate hikes through the year, which many economists and market analysts argued raised the risk of recession.
• However, the U.S. did not experience an officially dated NBER recession in 2022; while real GDP was negative in Q1 and Q2 2022, the National Bureau of Economic Research has not classified 2022 as a recession year, citing continued job growth and other indicators.

Whether the probability of recession in 2022 became “much greater” is not directly observable ex post; it depends on models or expectations that are not uniquely defined. The fact that no official 2022 recession occurred doesn’t by itself show that the risk didn’t rise; high‑probability events can fail to occur, and low‑probability events sometimes do occur. Because we cannot objectively verify the claimed change in probability against a clear standard, the statement cannot be cleanly scored as right or wrong.

Therefore the appropriate classification is ambiguous: the prediction concerns a shift in recession risk rather than an outcome that can be definitively checked against realized data.

Friedberg’s claim was that the post‑2021 startup downturn would not resemble a 2000/2008‑style crash where roughly a third of startups disappear and a large swath of marquee VC‑backed companies are wiped out (a ‘Sequoia graveyard’).

What actually happened 2022–2025:

• PitchBook data cited by multiple outlets shows that in 2023 about 3,200 VC‑backed startups failed, together having raised ~$27B. At the same time, there were over 55,000 VC‑backed companies operating in the U.S., implying failures on the order of only a few percent of the base in that year, not anywhere near one‑third, even allowing for undercounting. (linkedin.com)
• Carta and the Financial Times report that U.S. startup failures did spike sharply (e.g., 254 Carta‑tracked venture‑backed shutdowns in Q1 2024, with failure rates ~7x 2019 and a 60% year‑over‑year rise), but this is an increase in flow, not evidence that a third of the ecosystem has vanished. (ft.com)
• Sector‑specific data (e.g., travel & mobility tech) finds annual shutdown rates in the ~1.3–2.2% range over the past few years, with a noticeable 2024 spike but still low‑single‑digit percentages, again far below dot‑com‑bust‑type attrition. (tnmt.com)
• Funding conditions deteriorated substantially: U.S. startup funding fell about 30% to $171B in 2023 versus 2022, and global VC activity dropped sharply from 2021 peaks. (cryptopolitan.com) But by 2024, total VC investment rebounded nearly 30% to around $209B, with AI startups capturing a record share of that capital, signaling a difficult reset rather than a sustained systemic collapse. (reuters.com)
• In contrast, the dot‑com bust saw a majority of dot‑com startups disappear; only about 48% were still around by 2004, implying that more than half failed—a vastly higher wipeout rate than what we see in the 2022–2025 downturn. (en.wikipedia.org)
• As of January 2025 there are roughly 1,523 unicorns globally and 758 in the U.S., underscoring that a very large population of high‑profile, heavily VC‑backed startups remains; while some notable names (e.g., Convoy, Zume, WeWork’s bankruptcy and restructuring) failed, there has been no broad ‘graveyard’ of top‑tier VC franchises. (en.wikipedia.org)

Taken together, the data show a painful but selective shake‑out: elevated shutdowns, down‑rounds, and layoffs, yet far from a 2000‑ or 2008‑style extinction where a third or more of startups and a huge portion of marquee VC‑backed companies are wiped out. On that core point, Friedberg’s prediction that we would not see a crash of that severity has, so far, been borne out.

Available data around early 2024 shows increased concentration in both venture capital and broader private markets, but nowhere near the level Friedberg predicted, and certainly not concentrated in only 3–4 firms.

In U.S. venture capital, VC Cafe’s analysis (reported by Forbes) found that in 2024 the top 30 VC firms captured about 75% of all U.S. VC fundraising, while just nine firms captured 50%; Andreessen Horowitz alone represented a bit over 11% of total VC funds raised. (forbes.com) Similarly, SVB’s 2025 State-of-the-Markets analysis of Preqin and PitchBook data shows that over the 2023–2024 period, the top 10 VC firms captured only about 22% of global venture fundraising, indicating that capital is spread across dozens of managers, not dominated by 3–4. (slideshare.net) Other VC industry summaries likewise describe top 5 firms at ~40% and top 10 at just above 50% of global VC funding, again far from 80% for three or four firms. (wifitalents.com)

In broader private markets (private equity, credit, real estate, infrastructure, etc.), the pattern is similar: concentrated but still dispersed across many firms. Private Equity International’s PEI 300 data shows the top 10 private‑equity firms account for only about a quarter of total capital raised by the largest 300 managers over recent five‑year windows, not a dominant majority. (privateequityinternational.com) A 2025 analysis of global private markets notes that the top 10 PE firms raised roughly 22% of fundraising among the top 300 managers, and that even the six very largest PE houses together raised about 60% of capital in early 2024—still far short of 80% and involving more than 3–4 firms. (linkedin.com) An industry CEO interviewed by the Financial Times even forecasts that in future “50 to 100” large diversified private‑capital firms might capture 90% of capital—implicitly acknowledging that, as of the mid‑2020s, the market is far from being dominated by only three or four players. (ft.com)

Given these figures, by early 2024 neither venture capital nor the broader private‑markets universe had anything close to 80% of aggregate deployed capital coming from just three or four firms. The best evidence instead shows that dozens of managers share meaningful market share, with the top 3–4 responsible for only a small fraction of global deployment. Therefore Friedberg’s prediction did not come true.

As of November 30, 2025, Andreessen Horowitz (a16z) remains a privately held venture capital firm and has not gone public via an IPO. (en.wikipedia.org) Major U.S. VC peers commonly discussed alongside a16z—such as Sequoia Capital and Founders Fund—are also still structured as private firms rather than listed asset managers. (en.wikipedia.org)

There have been reports that General Catalyst is exploring an IPO and could become the first U.S. VC firm to go public, but as of now this remains exploratory and no IPO has occurred. (axios.com) Meanwhile, several venture-capital-style investment firms are publicly traded (e.g., Molten Ventures in the UK, JAFCO in Japan, HBM Healthcare Investments in Switzerland), but these were listed long before the 2022 podcast episode and are generally not the U.S. Sand Hill Road "peer group" under discussion. (en.wikipedia.org)

Because (1) Andreessen Horowitz has not yet IPO’d and (2) none of its obvious U.S. VC peers have gone public ahead of it either, the claim that a16z is "very likely" to be the first major VC firm to IPO has neither been confirmed nor falsified. It will only be clearly wrong if another peer actually lists before a16z, or clearly right if a16z is the first among that group to go public. Therefore the prediction’s status is inconclusive (too early to tell).