Elon's Anthropic Deal, The Next AI Monopoly?, "FDA for AI" Panic, Trading the AI Boom
2026-05-08 spoken.md · speaker-labeled ▶ watch ← E271 all episodes E273 →
Every number here is replayed from score_events — the same ledger the pool ranks on. Decay is what the 67 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 E271
0 hit · 0 partial · 3 miss — windows that closed after 2026-05-01 and up to 2026-05-08, auto-scored against price data and never hand-set. verdict · R · α
| idea | verdict | R | α | closed |
|---|---|---|---|---|
| 📈 Asset-light domestic services platforms are the tariff-immune trade | MISS | -10.9% | -39.5 | 2026-05-02 |
| 🏛️ Closing the unregistered-foreign-importer loophole hits Amazon's 3P marketplace | MISS | -41.2% | -69.8 | 2026-05-02 |
| 📈 Trans-Pacific container volume collapse seizes up the physical logistics economy | MISS | -55.3% | -83.9 | 2026-05-02 |
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 (3 ideas)
ordered by how hard each idea moved · quotes are verbatim from the transcript, timestamps deep-link into the episode
80% of our exposures or more have been in compute, AI, memory, ... It's hard to imagine a more Goldilocks situation for the United States ... At some point, you just have to acknowledge USA is winning.
I think we have kind of call it 500 days where you just got to be net long. But I think it's literally in the hundreds of days from now, 500, you're going to have to have an important reckoning moment. The people that are paying for all these tokens need to see it in actual benefit.
Anthropic and OpenAI's revenue performance has nothing to do with demand, zero. It is entirely to do with the supply constraints that exist in data centers and specifically in power. If they had infinite power, ... their revenues would probably be even more parabolic. ... the five-year view for those two companies is quite robust.
let me be the first to congratulate Dario on winning the AI race. ... from January 1st to March 31st, they grew from roughly 10 billion of ARR to 30 billion. ... unless something about their current trajectory changes, Anthropic will be the most powerful monopoly ever created in human history.
When Tesla and SpaceX merge, and we have all things Elon and Elon Corp, which will happen probably by the end of the year, maybe the middle of next year... you'll have this one asset that will trade at a valuation premium... which is logically explainable — everybody else has stopped innovating.
Episode digest
written during extraction and stored in data/extractions/ep272.json — the auditable source of truth, including everything market-adjacent that did not earn a capture
Chamath's E271 call lands in five days: Anthropic leases ALL of Colossus 1 (220K+ NVIDIA GPUs, 300MW+) — 'Elon Web Services' is born, Claude rate limits lift immediately, and Brad estimates $4-5B incremental EWS revenue de-risking the now-$2T SpaceX IPO story (pre-IPO, era rule, digest). The panel's Anthropic consensus peaks: Sacks fully reverses into 'congratulations Dario on winning the AI race' — 44B April ARR, ~100B exiting 2026, possibly 'the most powerful monopoly ever created in human history' (his strongest capture); Chamath re-ups multi-trillion-both with the power-not-demand frame; the only debated constraint is compute/energy and the competitive response (OpenAI's Spud/Codex resurgence, Google, Elon+Cursor). Chamath re-predicts the Tesla-SpaceX merger by end of 2026 ('Elon Corp'), and NVIDIA+Pulte+Span put mini GPU clusters beside new homes (digest). Trading the AI boom: Brad discloses Altimeter fully risk-on since Dec/Jan (80%+ in compute/AI/memory; 25% of the fund in SK Hynix/Samsung/Micron at 5-7x earnings) — S&P 'not the stuff bubbles are made of' at Meta 17x/NVDA 19x; Chamath flips back from his E269 risk-off to 'net long for ~500 days' but sets the reckoning marker: zero evidence yet of AI lifting S&P operating margins ('I bet any amount of money [the 200bps margin expansion] is not AI'), and the fork — OPEX-shrink vs revenue-grow — decides how society and markets respond around 2027-28. Sacks credits Trump's rescission of the Biden compute-approval regime for the whole boom (policy). The 'FDA for AI' panic segment and Spencer Pratt's LA mayoral run — politics, digest. Labels: clean; Brad guesting for Friedberg.