E274
SpaceX's $2T Case, Nvidia's Shock Selloff, America Turns on AI, Trump Pulls AI Order, Bond Crisis?
2026-05-22 spoken.md · speaker-labeled ▶ watch ← E273 all episodes E275 →
0
ideas born
6
ideas moved
7
captures · 3 voices
3
dissenting
-12.6
conviction added
-39.4
decay · 61 silent
Every number here is replayed from score_events — the same ledger the pool ranks on. Decay is what the 61 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 E273
1 hit · 0 partial · 2 miss — windows that closed after 2026-05-15 and up to 2026-05-22, auto-scored against price data and never hand-set. verdict · R · α
| idea | verdict | R | α | closed |
|---|---|---|---|---|
| 🏛️ Gulf capital realigns to a US bloc, displacing China's Belt and Road | HIT | +66.2% | +40.6 | 2026-05-17 |
| 🏛️ MFN drug-pricing EO compresses pharma profits and pushes R&D offshore | MISS | -40.2% | -65.9 | 2026-05-17 |
| 🏦 The "Big, Beautiful Bill" delivers no real deficit cut and the long end reprices | MISS | -1.6% | -27.3 | 2026-05-17 |
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 (6 ideas)
ordered by how hard each idea moved · quotes are verbatim from the transcript, timestamps deep-link into the episode
there's a narrative that Nvidia is losing share to the TPU, ... within the Western AI world, within data centers that are being built, ... Nvidia's AI business is growing faster than Broadcom's, and faster than a lot of other companies that are seen as part of this ASIC sharegate story. ... these other ASICs are not being submitted for benchmarks. ... I think the reason they're not being submitted is they will lose.
at the end of this Grok transaction last year, my ... we're going to move to these domain-specific architectures. I thought that was like a fait accompli, ... But the reality is that that DSA market evolution is actually happening inside of Nvidia. That's what's so insane to me. ... These guys actually have domain-specific architectures because they're doing these design programs with every,
I don't think a dollar crisis is around the ... the Bundesbank was still in charge and you had the Deutschmark, ... what is globally a bad neighborhood of high debt levels. And we have AI in our corner, and we have energy self-sufficiency.
It ultimately breaks and as it starts to break, you have massive inflation because the value of your underlying currency collapses and then you have money printing and ... is one of those things that could be a catalyst for a credit crisis, because there's a lot of people that are in this carry trade. ... this is water leaking out of the bucket.
▲ GOOGL
🤖 Anthropic IPO (~Oct 2026) is attractively priced at $1.5-2T
watch 54
CONTESTED
▲ +9.1
72.4 → 81.5
What is happening with AI right now, with Anthropic growing faster than any company in history at massive scale? ... they're bigger than 100 different countries for sure. Exactly. And compounding really fast and now profitable.
if OpenAI and Anthropic call it $100 billion of ARR now, with 80%-ish gross margins on inference, the returns are there. ... if we add in Gemini, we add in Cursor, we add in XAI, we add in Open Source, it's not hard to see $200, $300, $400 billion of ARR at the end of this year at high margins. ... there's going to be a really strong ROI this year, even excluding
Episode digest
written during extraction and stored in data/extractions/ep274.json — the auditable source of truth, including everything market-adjacent that did not earn a capture
Sacks out; Gavin Baker (Atreides) guests — his second pod appearance per Jason, so the regular-guest promotion rule (≥2 distinct episodes → weight 1.0) may trigger once the DB sees a second episode from him; professional fund manager (100+ positions), disclosed Atreides is an investor in Excite Labs (in 'essentially every Starlink') — conflict noted. SpaceX S-1 FILED: $75B raise at $1.75T, ticker SPCX confirmed, listing ~June 12; Starlink $11.4B rev/+50%/$4.4B op income, space $4B rev/-$650M, AI $3.2B rev/-$6.4B; EWS (Anthropic renting Colossus 1 + parts of 2) at $1.25B/MONTH — $45B over 3 years, cancellable on 90 days notice; Polymarket 71% closes day one above $2T. ALL pre-IPO → era rule, digest-only for spacex-multiple-ways-to-win: Chamath's underwriting case ($18-19B '25 rev → 25-30 '26 → 40-45 '27 then doubling; terrestrial data centers alone $100-200B revenue by 2030-32; capital moat → technology moat → execution moat; 'we're at the beginning of the beginning'; the Jensen DC-to-DC design-partner angle; 'the only guy [with] civilizational out-of-left-field s**t'), Gavin's build-speed stat (122→91→66 days per data center; 'GPUs will be allocated to who can plug them in'), Friedberg's civilizational-backup frame (space-based comms/data centers beyond government control; discloses he owns NO SpaceX). Chamath's Tesla-merge re-up captured on tesla-second-act. Cursor's Composer 2.5 goes Pareto-dominant after 3-4 weeks of RL on Colossus 2 (base = Kimi K25); Grok Build gives xAI its harness; Gavin: 'extremely significant for XAI and Cursor' — private, digest. Orbital compute: Gavin's point prediction H2'28-H1'30, working H100 already in space (Karpathy trained on it); 'the only people skeptical are those not involved in data centers or space' — digest. Karpathy joins Anthropic to lead recursive self-improvement pre-training: Chamath's order-of-magnitude-yearly/'parabolic' model-quality take and Gavin's 'might seem conservative' — no clean instrument, digest; Anthropic per Gavin now PROFITABLE (captured). NVIDIA blowout: $81.6B rev +85% y/y (+20% q/q), $58B net income, $48B FCF, 75% GM, $80B more buybacks, dividend 1c→25c, 50% of FCF to shareholders — yet stock +16% YTD at ~$5.3T; Gavin's cross-sectional-inefficiency frame (memory 3-5x PE, NVDA low-teens vs rich power/cooling/optical multiples — 'both cannot be true'), the ASIC-sharegate rebuttal (captured as fragmentation oppose), $20B overnight CPU business, and the amortization-bear-case kill: mix-and-match DSAs in front of old GPUs extend useful life to 10-15 years, CoreWeave financing at 6% — 'this quarter Nvidia single-handedly saved the Neo clouds' (Chamath); Burry politely thanked ('we need bears'). Leopold Aschenbrenner's 13F puts read as Iran hedges, not a semis short (Gavin). America-turns-on-AI: booed commencement speeches, possible CCP-funded anti-AI/data-center campaign (Gavin), Friedberg's three-cause diagnosis (asymmetric-diffusion power imbalance, foreign-state interference lineage from KGB playbooks, Copernican anti-humanist ego shock) and Manhattan-Project proliferation logic — can't slow down, need China detente; Chamath's KYC-with-China proposal; Sham Sankar's 'ask the end user' clip; Cloudflare's Matthew Prince fires 20% ('measurers') — Chamath: 'PR school of retards... you all suck at this'; Zuck's dystopian layoff (8K cut while recording all employee screens to train models). Trump PULLS the AI EO hours before signing — would have required federal supervision/review of frontier models; the CEOs (incl. Nikesh) had been invited; policy, digest. Trump-Xi follow-up: soybeans/planes/H100s-H200s to Baidu, no grand deal; Putin with Xi next day (Friedberg: 'no rainbow-colored chapter 3'); Chamath hints the real deal is a behind-closed-doors game-board split; Gavin's oil-chessboard deterrent (Venezuela+Iran off the board → China can't fight a war on Russian oil alone) and 'every day the Strait of Hormuz is closed is relatively good for America' (natgas down in US while world LNG +100-200% — relative-strength point, digest). Bond crisis macro: May CPI seen 4.2%+ (Polymarket 99%), pro forecasters at 6%, 10Y 4.6% (Bessent target <4% broken), Japan 30Y 5.1% record, UK/Germany yields at multi-decade highs, Korean retail levering into AI chips — narrative flipped from Fed cuts to possible HIKES (headwind noted for fed-easing-2026, no capture — Jason reporting, not predicting); Friedberg's debt-spiral doom captured; Chamath's response is discipline not direction (5-or-less concentrated holdings, 'stick to eating'); Gavin's three-things-can-be-true: rates up is concerning AND AI fundamentals accelerating AND Hormuz favors America — dollar-crisis oppose captured; tech bubble comparison (Cisco 100x fwd vs NVDA low-teens); flags AI fundamentals may be SEASONAL (summer lull — watch for it in Q3 prints). Labels: full names used (Jason Calacanis/Chamath Palihapitiya/David Friedberg/Gavin Baker — new variant, first names still unambiguous); SPEAKER_5 = Zuckerberg video clip at 38:42 plus one stray interjection at 1:18:32 — no captures rest on SPEAKER_5.