E143: Nvidia smashes earnings, Arm walks the plank, M&A market, Vivek dominates GOP debate & more
2023-08-25 spoken.md · speaker-labeled ▶ watch ← E142 all episodes E144 →
Every number here is replayed from score_events — the same ledger the pool ranks on. Decay is what the 116 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 E142
1 hit · 1 partial · 2 miss — windows that closed after 2023-08-18 and up to 2023-08-25, auto-scored against price data and never hand-set. verdict · R · α
| idea | verdict | R | α | closed |
|---|---|---|---|---|
| 🏛️ Blue-state permitting has broken housing construction — building migrates to light-regulation states | HIT | +43.0% | +35.7 | 2023-08-20 |
| ⚡ IRA production credits flip climate-tech unit economics — capital floods in | MISS | -27.5% | -34.8 | 2023-08-20 |
| 🏛️ Medicare price negotiation compresses pharma pricing power and biotech returns | PARTIAL | +9.8% | +2.5 | 2023-08-20 |
| 🏛️ The IRA kills the carbon tax — carbon markets and DAC are dead ends | MISS | +1.1% | -6.3 | 2023-08-20 |
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 (15 ideas)
ordered by how hard each idea moved · quotes are verbatim from the transcript, timestamps deep-link into the episode
Chamath's rule-of-capitalism call off the Q2 blowout: nothing about what Nvidia does is monopolistic, so the visible revenue and profit pool pulls in competitors — Google TPU, Amazon's own silicon, Microsoft FPGA, RISC-V as an open ISA, and potentially Tesla open-sourcing or selling the Dojo/FSD chip platform — and the result is decaying margins. Sacks supplies the same conclusion from the other end: today's near-software gross margins exist only because a GPU shortage lets Nvidia charge whatever it wants. The same mechanism hits Arm's ISA-licensing model as RISC-V spreads and Apple in-sources design, so future profits in merchant silicon are lower than historical profits.
when the market observes that there's a company just printing enormous revenues and profits, they wanna compete with them naturally to get their share of those revenues and profits. And that typically happens in all markets, and the result of that are just decaying margins... I think that that probably decays the Nvidia margin and upside over time.
Chamath's explanation of why a flat, shrinking Arm is being pushed public into a bad tape: SoftBank the telco operator carries ginormous debt against a contracting core business, so it is at risk of breaching covenants and running out of free cash flow. It therefore has to monetize whatever is liquid — Alibaba already sold down, Arm next — regardless of price. That makes SoftBank a price-insensitive seller and puts a supply overhang on whatever it lists.
Chamath's disclosed positioning in the Vivek climate-hoax debate: he has put zero dollars into anything sold to him as a climate investment, and hundreds of millions into technology that reduces energy and materials dependence on other countries, with lower emissions as a by-product. His claim is that the dogmatic climate framing actively repels capital while the energy-independence/fewer-endless-wars framing mobilises it, so the money in the transition goes to domestic supply-chain and hard-asset plays rather than to ESG-branded ones.
if it's happening at the software layer already now, just like it did in Web2 software, and then we see certain elements of Web2 hardware been open sourced, I think it makes pretty logical sense that you can expect the same things to happen in the AI world.
basically these big companies have decided, no, we're just gonna make all these models extremely good, extremely useful and very, very free. And so a lot of the resources are going there to subsidize, economically subsidize, and by implication, economically destroy the value of that category. That's going to be good for startups.
And those businesses, by the way, that we saw on the.com boom and the global financial crisis that were able to do that during the Doldrum Death Valley March that everyone's going through right now, they emerged victorious... On the other side of this trial era, a number of very high quality companies will emerge
there are a lot of companies which have built up huge prep stacks because they raise too much money at the peak at valuations are too high. Going public does allow you to reset your whole prep stack because all the preferred with all the rights and preferences converts to common.
Episode digest
written during extraction and stored in data/extractions/ep143.json — the auditable source of truth, including everything market-adjacent that did not earn a capture
DIARIZATION DEFECT: Chamath has zero labelled turns but is present and addressed by name throughout — every one of his turns is merged into Jason's label, so all Chamath timestamps here are the start of the containing Jason turn (6:48, 19:19, 31:09, 36:52, 47:25, 1:36:48, 1:42:09) rather than the moment he starts speaking. On Nvidia's blowout quarter Chamath called it 'peak in one way' — not on execution but on capitalism: TPUs, Amazon silicon, RISC-V and a possible Tesla decision to open-source or sell the Dojo/FSD platform compete the margin away, a new bearish idea that Sacks reinforced by tying today's software-like margins to a temporary GPU shortage. Friedberg re-upped his own gen-AI bubble call with an efficient-frontier argument (compute ROI stops scaling linearly, so a lot of current spend is inefficient), Jason pushed the dark-fiber overbuild analogy, and Sacks took the other side — capacity gets absorbed as applications grow. Chamath's Arm call is the sharpest number in the episode: SoftBank is marketing a $60-70B print for what he says is honestly a $15-20B company, and he explains it as forced deleveraging by a debt-choked telco (new idea) rather than a real IPO; Friedberg says there is simply no IPO window, Sacks says only fundamentally good companies get out and only at a big down round, and notes going public is now the cheapest way to wash a bloated preference stack. Chamath also declared the M&A market 'effectively dead' with the EC/CMA now probing Adobe-Figma after their Activision win. On the GOP debate the four split on whether Trump's nomination is locked: Chamath bought the 'deep out of the money option' on Vivek, Jason thinks Trump doesn't run, while Sacks (MAGA faction wins) and Friedberg (Trump won the night by not showing up) say it holds. Not coined: Chamath's claim that a Starlink IPO is the only listing that could catalyse the window (SpaceX is unlistable) and his Tesla-sells-Dojo speculation, which is folded into the silicon-margin idea's thesis.