E146: Did the Fed break the VC model? Plus IPOs, M&A, revaluing unicorns & more
2023-09-22 spoken.md · speaker-labeled ▶ watch ← E144 all episodes E147 →
Every number here is replayed from score_events — the same ledger the pool ranks on. Decay is what the 103 ideas nobody mentioned gave up this week; it applies only when an episode is processed.
Tier crossings
conviction thresholds crossed by this episode — 65 / 45 / 15 · ideas born here show where they landed. why 65 isn't always green
Kill dates that landed since E144
1 hit · 0 partial · 4 miss — windows that closed after 2023-09-01 and up to 2023-09-22, auto-scored against price data and never hand-set. verdict · R · α
| idea | verdict | R | α | closed |
|---|---|---|---|---|
| 📈 Adobe/Figma at $20B is a good deal - the 15-20% selloff is an overreaction | HIT | +78.6% | +62.7 | 2023-09-17 |
| 📈 Ag retail is the chokepoint for farm productivity - Friedberg SPACs Lavoro | MISS | -26.6% | -42.5 | 2023-09-17 |
| 📈 FedEx's warning is competitive, not macro - Amazon is eating the parcel business | MISS | -59.6% | -75.5 | 2023-09-17 |
| 🏛️ US antitrust is punitive politics, not competition policy - hot-button platforms carry the overhang | MISS | -102.9% | -118.8 | 2023-09-17 |
| 📈 Pediatric ADHD/antidepressant over-prescription becomes an opioid-style reckoning | MISS | -2.8% | -12.9 | 2023-09-10 |
Who moved the board
each voice's force on conviction this episode — supports and opposes, weighted exactly as the replay applied them · share = % of this episode's movement
What got argued (22 ideas)
ordered by how hard each idea moved · quotes are verbatim from the transcript, timestamps deep-link into the episode
Chamath's numeracy argument on the 2023 UAW strike: sensitize Ford, GM and Stellantis P&Ls to the terms the union has actually proposed and the legacy Detroit cost base stops working — his Ford pro forma swings from ~$30B of profits to ~$17B of losses, a $47B swing. The only offset is more expensive cars, which cannot be sold into 7-9% consumer lending rates, so the structural gap to non-unionized, highly automated builders (Tesla, Rivian) becomes unbridgeable and the eventual pension/bailout tail lands on the taxpayer. Bearish the unionized OEMs specifically, not the auto market.
I think that the labor deal is an existential risk to the unionized auto industry in America. ... If the deal as announced happens and you sensitize Stellantis, GM and Ford's P&L to these new terms ... it's going to be very difficult for the established auto industry to survive
The Fed broke the VC model. Run the Instacart-vs-S&P chart at every entry point and venture's excess return over the index is thin (the 2015 Series C generated ~32% alpha over eight years, the 2018 Series F was negative), and that thin alpha now has to beat 5-6% T-bills and 11-13% structured credit that are liquid and far less volatile. So LPs do the arithmetic and cut: fewer managers, smaller commitments, and materially less total capital into venture in aggregate. The listed read-through is the alt-asset managers whose fee streams depend on LP appetite for illiquid alternatives, plus the venture-backed cohort that gets starved of follow-on capital.
when people do the calculations on what the true alpha of venture capital is, they're going to come back with answers like this, which is it's not that great ... because the alternative universe is more liquid, it's less volatile, and it has roughly the same amount of return
Chamath's sum-of-parts read off the Klaviyo IPO: Shopify quietly owns ~11-12% of Klaviyo, ~13-14% of Flexport after selling Deliverr into it, and probably a non-trivial share of Stripe — an ecosystem-building strategy that converts partner support into balance-sheet equity. Against only ~$2-3B of cash and equivalents in the market's valuation, those stakes are not being priced, so there is upside in SHOP 'just for free'. Disclosed as an unowned, un-modelled observation rather than a position.
Chamath's underwriting frame for the newly listed Instacart: businesses that intermediate large gross transaction values have never historically defended their take rate, because either competition or supplier pricing power erodes it. In grocery specifically Walmart and Amazon will do the same job for much cheaper, so the honest model is a declining take rate against growing share. Layer on contracted advertising multiples — the market discounts ad revenue as too levered to rates and the economy, and Facebook/Google take an increasing share of it — and the ad-led revenue mix does not rescue the valuation. Airbnb is the same shape one step further along, with regulatory capture (New York) doing the take-rate damage.
for these businesses that drive huge GTVs, gross transaction values, is I think most people, when they try to find what they're worth, are very sensitive to the take rate ... take rate tends to decay. So said in, you know, in grocery land, I think it's because Walmart and Amazon will try to do it for much, much cheaper ... I think that you have to model the health of the business with a declining take rate and growing share.
Meaning everyone's talking about a soft landing. Everyone's banking on a soft landing. But soft landings are actually exceedingly rare. ... When you have very fast rate tightening cycles, it generally has a very predictable effect on the economy. There's a lag, but the effect is very predictable, which is it causes recessions. ... But I think the consumer hasn't been hit yet.
I told my CEOs, guys, let's get enough cash to last through the middle of 25 Remember, I was pretty clear about that to folks. ... I think that that was wrong. ... I think now you got to be Q1 of 26 and maybe even mid 26 ... So I was wrong by at least a year, Jason, because of this. I think I got the words right, but I got the timing wrong.
the forcing function will be the debt service costs, which has just crossed a trillion dollars a year just to pay the interest and it's mounting. 30% of our debt, I think, is coming up for refinancing in the next 12 months. And it's going to refinance at a 5% rate
I asked him about hiring and the size of the company said, I'm not hiring anybody in the next year or two because all my developers are 30 or 40% better ... I got senior developers who are turning into 10x developers. We're not hiring. We're just going to have increased margin.
I think Jeremy Ball's substacks showed that the average growth forecast for SaaS companies for the next 12 months had been cut roughly in half. ... That meant they were buying a lot less software on a per-seat basis. So I think we've been through, call it a B2B or enterprise recession.
New options are issued and a round is done at a significant discount. And there's a huge recap and a pay to play and all this other sort of stuff starts to play out that the original founders in the company get wiped out. ... I think that's the scary scenario that's likely going to play out
Episode digest
written during extraction and stored in data/extractions/ep146.json — the auditable source of truth, including everything market-adjacent that did not earn a capture
Post-All-In-Summit episode dominated by the Instacart/Klaviyo/Arm IPO week and the Fed's hawkish hold. Chamath calls the reopening a failure on mechanics — sub-10% floats, no anchor buyers, no lockups — so 'the grand reopening was a grand closing', while Sacks reads the same tape as a green shoot at a permanently lower valuation level. The core new thesis is the episode's title: run venture's alpha over the S&P at every Instacart entry point and it does not clear 5-6% T-bills or 11-13% structured credit, so LPs cut managers and total venture capital shrinks. Chamath takes his 'higher for longer' victory lap and in the same breath reverses his own March-2023 call that the Fed would be forced into cuts — he now tells portfolio CEOs to fund to Q1/mid-2026 and says he was 'wrong by at least a year'. New tradeable coinages: the UAW deal as existential for Ford/GM/Stellantis against non-union Tesla/Rivian (his Ford pro forma is a $47B swing), Shopify's unpriced Klaviyo/Flexport/Stripe stakes, and take-rate decay as the right way to underwrite Instacart. Friedberg's Science Corner on glycosylated-antigen 'inverse vaccines' for autoimmune disease was captured as no idea — novel modality, but academic-stage with no listed instrument. ATTRIBUTION FLAG: Friedberg has zero labelled turns and is merged into Chamath's label; his two captures here (11:10, 1:21:57) are each confirmed by another host naming him immediately before or after, but the long unicorn-recap monologue at 1:10:29 is attributed to Chamath on receipts (he is the one doing recap rounds this week) and is worth an audio spot-check.