A circular capital loop — firms booking each other's spending as revenue, solvent only while external capital keeps flowing — gated by physical chokepoints it cannot buy past on the timeline (compute capital, China rare earths, Russian enrichment, the power grid), and embedded in measurement and control layers (official statistics, paper commodity prices, and the identity / surveillance / political-money rails) that determine whether it can be seen and questioned. The financial core is machine-verified; every other layer is evidence-graded and excluded from the proofs. Not one cabal — recurring operators rebuilding the dot-com (vendor financing), Enron (off-balance-sheet SPVs), and LTCM (interconnected leverage) structures in the least-regulated venue.
11firm circular core (SCC; 12 incl. cancelable)
16round-trip cycles
229/234graph nodes/edges
117/117financial/structural edges
$1.03Texternal capital OpenAI needs
57%NVIDIA self-funding (headline)
-81%US home priced in gold since 1998
2adversary chokepoints (CN+RU)
91%credit common factor (PC1 share; diversification illusory)
78cited research blocks
36runnable models
Formal verdicts (every engine)
| Engine | What it shows | Verdict |
|---|---|---|
| Graph SCC | 11-firm robust core; SpaceX cancelable-only | PROVED |
| Z3 T1 round-trip | a money cycle is realizable | SAT |
| Z3 T2 NVIDIA self-finance | ≥$25.8B revenue self-financed | UNSAT(refuted) |
| Z3 T3 OpenAI commitments | needs ≥$1.03T external capital | UNSAT(refuted) |
| Z3 T4 core at zero inflow | insolvent without continued capital | UNSAT(refuted) |
| Z3 T5 SpaceX vs OpenAI | SpaceX SAT / OpenAI UNSAT standalone | PROVED |
| Z3 reflexive marks M1-M4 | AI 'profit' is self-referential paper marks; gains reverse if IPO < private mark | SAT/UNSAT |
| Z3 self-marked value U1-U4 | bank HTM / AI marks / private credit / insurance = one defect; gaps correlate under a common factor (no netting); forced convergence on the event | SAT/UNSAT |
| Z3 depreciation trap D1-D4 | useful-life = the 5th self-marked number; honest 2-3yr life strictly lowers profit (~$176B 2026-28, Burry); asset life < 5-19yr financing tenor -> equity can't stay whole; depreciation is only timing | SAT/UNSAT |
| TLA+ cascade | capital-stop → OpenAI→CoreWeave→Oracle | VIOLATED(trace) |
| TLA+ SpaceX safe | never defaults | HOLDS |
| TLA+ MarkUnwind | down-round/IPO<mark -> Amazon+Google reverse gains (writedown trace) | VIOLATED(trace) |
| Alloy structure | core circular; SpaceX separable w/o cancelable | ALL HOLD |
| Z3 coordination C1-C4 | small-bank stablecoin trap, suppressible | PROVED |
| Z3 Fed trap F1-F3 | no single rate satisfies N targets | UNSAT |
| Z3 defense chokepoint | REE independence infeasible until ~2028 | UNSAT |
| Z3 power adequacy P1 | AI power demand > supply thru 2028 | UNSAT |
| Z3 HALEU chokepoint P2 | Russia enrichment dependence to ~2029 | UNSAT |
| Allied REE min-cut | feasible ~2028; bottleneck = midstream separation, not mining | max-flow |
| Z3 age-verif futility | effective gating UNSAT under breach; surveillance unconditional | UNSAT |
The circular core
Robust core SCC (11): AMD, Amazon, Anthropic, CoreWeave, Crusoe, Lambda, Microsoft, NVIDIA, OpenAI, Oracle, SINK_lenders.
Circular only via cancelable edges: SpaceX (SpaceX: operationally separable, financially cross-held by Google's ~$100B stake).
Edges split into a financial layer (117 capital/credit/compute flows) and a structural layer (117 governance/legal/security/ownership relationships). The SCC computed over the financial layer alone equals the SCC over all edges (structural_edges_add_no_cycle = True): the circular core rests on capital flows; the graded structural edges add no cycle.
Cross-layer connectors
Nodes ranked by distinct neighbor-sectors bridged (degree, source-files, and whether they span both layers). This quantifies the bridge nodes that tie the financial core to the surrounding layers.
| Node | Degree | Sectors bridged | Source files | Layers | Neighbor sectors |
|---|---|---|---|---|---|
| Hedera | 16 | 10 | 1 | structural | bank, crypto_infra, defense, dlt, financial_infra, financier, hyperscaler |
| SpaceX | 11 | 7 | 3 | financial+structural | ai_lab, chip_vendor, exogenous_source, hyperscaler, satellite, state_intel, telecom |
| NVIDIA | 17 | 6 | 4 | financial+structural | ai_lab, chip_vendor, hyperscaler, neocloud, space_real_economy, telecom |
| Anthropic | 11 | 5 | 4 | financial+structural | chip_vendor, financier, hyperscaler, sink, space_real_economy |
| 7 | 5 | 4 | financial+structural | ai_lab, chip_vendor, dlt, ip_rightsholder, space_real_economy | |
| OpenAI | 17 | 6 | 5 | financial | chip_vendor, financier, hyperscaler, ip_rightsholder, neocloud, spv |
| Oracle | 7 | 6 | 4 | financial | ai_lab, bigtech_asia, chip_vendor, private_credit, sink, spv |
| JPMorgan | 5 | 4 | 4 | financial+structural | commodity_market, crypto_infra, other, spv |
| MGX | 8 | 4 | 4 | financial+structural | ai_lab, bigtech_asia, exchange, spv |
| Binance | 6 | 4 | 3 | financial+structural | financier, other, regulator, stablecoin |
| Ripple | 5 | 4 | 3 | financial+structural | bank, financier, political, regulator |
| CoreWeave | 10 | 5 | 3 | financial | ai_lab, chip_vendor, hyperscaler, private_credit, sink |
The three physical chokepoints
| Chokepoint | Constraint | Adversary | Independence | Proof |
|---|---|---|---|---|
| Compute capital | OpenAI needs ≥$1.03T external; solvent only while capital flows | — | capital trap | Z3 T3/T4 |
| Rare earths (defense) | ~5,550 kg/yr samarium, ~100% China; auto-deny Dec-2025 | China | ~2028 | defense_chokepoint UNSAT |
| Power + fuel (energy) | firm-power gap widens to 2028; HALEU ~100% Russian | Russia | ~2028-29 | power_adequacy UNSAT |
Scenario engine (formal verdicts)
| Scenario | OpenAI cap gap | solvent@tap | REE indep | power ok | verdict |
|---|---|---|---|---|---|
| BASE | $1,090B | YES | 2028 | >2028 | FRAGILE |
| BULL | $40B | YES | 2028 | 2026 | RESILIENT |
| BEAR | $1,331B | NO | 2029 | >2028 | BREAKS |
BASE = FRAGILE (solvent only while the capital tap stays open) · BULL = RESILIENT (near self-financing) · BEAR = BREAKS (carry-unwind shuts the tap → the cascade).
Everything in hard money (gold lens)
| Year | US median home $ | in gold | $ idx (1998=100) | GOLD idx |
|---|---|---|---|---|
| 1998 | $152,000 | 517.0 oz | 100 | 100 |
| 2007 | $247,000 | 355.4 oz | 162 | 69 |
| 2012 | $245,000 | 146.8 oz | 161 | 28 |
| 2020 | $336,000 | 189.8 oz | 221 | 37 |
| 2022 | $457,000 | 253.7 oz | 301 | 49 |
| 2026 | $418,000 | 97.2 oz | 275 | 19 |
The gold lens SPLITS the market: S&P 500 -69% in gold since 2000 (debasement) vs NVIDIA +1985% in gold since 2016 (real value capture). The bubble = broad debasement + extreme real concentration into the circular oligopoly.
US home +175% in dollars since 1998 but -81% in gold. CRE peaked in gold ~2001 (now ~16 vs 100). DoD budget +3.7x nominal but -75% in gold. OpenAI's $1.4T = 0.53x all of TARP in gold.
The weavers (temporal meta-graph, betweenness)
| Node | Layer | Betweenness | Layers | Eras |
|---|---|---|---|---|
| OpenAI | ai_core | 2889 | 6 | 3 |
| a16z | crypto | 1797 | 4 | 3 |
| Ethereum | crypto | 975 | 3 | 2 |
| Stargate | ai_core | 966 | 4 | 3 |
| Larry Summers | finance_dereg | 752 | 3 | 2 |
| Circle/USDC | crypto | 726 | 4 | 3 |
| Peter Thiel | payments | 687 | 5 | 3 |
| PayPal | payments | 619 | 2 | 1 |
| Fairshake PAC | crypto | 592 | 3 | 3 |
| GLBA 1999 | finance_dereg | 577 | 2 | 2 |
| NVIDIA | ai_core | 549 | 3 | 3 |
| BlackRock | private_credit | 521 | 4 | 2 |
Bank vulnerability (biggest hidden HTM holes)
| Bank | HTM loss/equity | AFS/eq |
|---|---|---|
| USAA FEDERAL SAVINGS BANK | -50.2% | -1.09B |
| CHARLES SCHWAB BANK SSB | -43.3% | -3.59B |
| BANK OF HAWAII | -34.7% | -0.16B |
| BANK OF AMERICA NA | -33.6% | -2.18B |
| CHARLES SCHWAB PREMIER BK SS | -26.9% | -0.13B |
| WOODFOREST NATIONAL BANK | -24.1% | -0.08B |
| MORGAN STANLEY PRIVATE BK NA | -22.7% | -0.44B |
| WASHINGTON TRUST BANK | -22.6% | -0.02B |
8 mid-tier banks flagged on ≥2 axes (CRE + securities loss + uninsured). Foreign-branch artifacts excluded.
Cross-sectional analysis
At each moment, compare the whole cross-section of segments — credit-rating buckets, sovereigns, muni states — and measure dispersion, relative value, and the shared common factor. PCA first-principal-component share on monthly spread changes (Collin-Dufresne–Goldstein–Martin basis). Engine: models/graph/cross_section.py; full charts + correlation heatmap + RV snapshot on the charts page.
The common factor (PC1 share)
| Cross-section | Segments | Avg pairwise corr | PC1 share (common factor) | Dispersion now |
|---|---|---|---|---|
| US corporate credit (OAS ladder) | 5 | 0.883 | 91% | elevated (z=0.78) |
| Developed sovereign 10Y | 26 | 0.604 | 65% | mid-range (z=-0.4) |
| Municipal (per-state/quality) | 4 | 0.466 | 70% | compressed (z=-0.65) |
| Corporate breadth by tier (FINRA TRACE) | 3 | 0.805 | 87% | compressed (z=-0.67) |
A high PC1 share = the segments move as one. US credit ≈ 91% confirms the CDGM common factor and means cross-credit diversification is largely illusory at the system level — the empirical agreement with the self-marked-value claim (self_marked_value U1–U4: the gaps correlate under a common factor; no netting). Most stress priced right now: Japan; tightest: HY muni (HYD).
Bank vulnerability cross-section (peer-relative z)
| Bank | St | Assets $B | HTM loss/eq | Uninsured | CRE/T1 | Composite z | Pctile |
|---|---|---|---|---|---|---|---|
| BANK OF HAWAII | HI | 23.9 | -34.7% | 47.0% | 165.0% | 1.52 | 100.0% |
| WASHINGTON TRUST BANK | WA | 10.7 | -22.6% | 57.0% | 298.0% | 1.51 | 99.0% |
| CITIZENS BUSINESS BANK NA | CA | 15.5 | -15.5% | 54.0% | 323.0% | 1.2 | 99.0% |
| FARMERS&MERCHANTS BK LONG BE | CA | 11.9 | -15.4% | 55.0% | 290.0% | 1.13 | 98.0% |
| BANK OF AMERICA NA | NC | 2672.2 | -33.6% | 44.0% | 38.0% | 1.08 | 98.0% |
| STATE STREET BANK&TRUST CO | MA | 386.5 | -14.5% | 93.0% | 7.0% | 1.08 | 98.0% |
| WOODFOREST NATIONAL BANK | TX | 9.2 | -24.1% | 29.0% | 305.0% | 1.06 | 97.0% |
| USAA FEDERAL SAVINGS BANK | AZ | 109.7 | -50.2% | 5.0% | 0.0% | 0.99 | 96.0% |
Cross-layer connectors (bridging score)
| Node | Degree | Sectors bridged | Both layers | Bridge score |
|---|---|---|---|---|
| Hedera | 16 | 10 | no | 4.64 |
| NVIDIA | 17 | 6 | yes | 3.14 |
| OpenAI | 17 | 6 | no | 2.64 |
| SpaceX | 11 | 7 | yes | 2.34 |
| Anthropic | 11 | 5 | yes | 1.23 |
| CoreWeave | 10 | 5 | no | 0.5 |
| Oracle | 7 | 6 | no | 0.38 |
| 7 | 5 | yes | 0.32 |
Evidence-graded overlays (excluded from the proofs)
These layers extend the map beyond the financial core. Each item is graded fact | contested | weak | unsupported and is excluded from the Z3/TLA+/Alloy proofs.
| Layer | What it documents | File |
|---|---|---|
| Regulatory capture (SEC/SDNY/FDIC) | Choke Point 2.0 FOIA, ConsenSys/MetaMask, Ripple-timing+JPM, Kraken/Binance/LBRY sweep, Fairshake $169M | research/spec-sec-sdny-regulatory.json |
| Privacy-tool prosecutions | Tornado Cash + Samourai: real Lazarus $7.6B use vs theories courts rejected (Van Loon) vs developers jailed (code-is-speech) | research/spec-tornado-samourai.json |
| Exchanges + Asia/Gulf | Mt.Gox unwind, Binance CZ-pardon/USD1/MGX, SBI/Ripple ~9%, Ant HKDA+PBoC pause, ByteDance/TikTok->Oracle+MGX | research/spec-exchanges-asia.json |
| Telecom + satellite + ISR | Salt Typhoon broke the CALEA backdoor; Amazon/Apple Globalstar; Starshield/NRO; carrier direct-to-cell | research/spec-telecom-satellite.json |
| Disclosures + surveillance | Snowden (Bullrun/Dual_EC/Room 641A), In-Q-Tel/Palantir, DARPA-TrailOfBits, FBI ANOM, Epstein Files Act, Pandora | research/spec-disclosures-surveillance.json |
| PQC / quantum-sat / DLT | SEALSQ/WISeKey global build-out + Hedera 34-member council overlaps (Google/IBM) + SEALCOIN bridge | research/fin-hedera-connections.json |
| Official-data integrity (epistemics) | Zero-trust read of BLS/BEA/Fed: -911k jobs benchmark, OER/CPI methodology, Boskin fiscal motive, 2025 collection cuts + BLS-chief firing, the understate-inflation incentive map | research/macro-official-data-integrity.json |
| Rent vs CPI (ALNRI analysis) | Apartment List / Zillow / BLS New-Tenant rents lead official CPI shelter by ~1 year; shelter ~1/3 of CPI -> headline misreports turning points | research/macro-rent-cpi-divergence.json |
| Gig / contingent labor | Full-time independents 13.6M->27.7M; misclassification (Lyft NJ $19.4M); CES/CPS + multiple-jobholders distort the jobs headline | research/macro-gig-labor.json |
| Futures vs physical (commodities) | COMEX/LBMA/LME/SHFE: 2025-26 silver backwardation, >$2.50 gold premium, ~4.2:1 paper:registered, 28% copper tariff spread, JPM $920M spoofing | research/macro-futures-vs-physical.json |
| SEC filings (primary source) | The circular-funding thesis in the filers' own 10-Ks: CoreWeave 67% Microsoft + $21B debt; NVIDIA concentration 36%->61%; Amazon/Google fair-value GAINS vs Microsoft equity-method LOSSES | research/spec-sec-filings-primary.json |
| Bank HTM marks (zero-trust) | HTM 'held to par' is a self-chosen value, not a price; FDIC unrealized-loss series ~$690B(Q3-22)->$306B(Q4-25)->$325B(Q1-26); withdrawn problem-bank totals reconstructable from RC-B + FDIC API | research/macro-bank-htm-marks.json |
| Private credit marks (zero-trust) | 3rd self-marked asset class (manager NAVs); First Brands/Tricolor marked ~par then ~12-33c; bad-PIK ~3x 2021; risk migrating to insurance annuities + 401(k) retail | research/macro-private-credit-marks.json |
| Insurance / Bermuda endpoint | Where the risk lands: PE-owned insurers (>$700B, ~25% of US life) fund annuities with manager-marked private credit, cede liabilities to Bermuda CAPTIVES they control (~60% offshore = internal 'transfer'); IMF/FSB/FIO/NAIC warning | research/spec-insurance-bermuda.json |
| Jobs / inflation / Fed-vs-bond-market (CHARTS) | Disambiguated jobs + inflation over 10yr; the funds rate tracks the 2Y Treasury (bond market), not the 2%/full-employment mandate; see the Charts page | research/macro-jobs-inflation-fed.json |
| Stablecoin -> Treasury rail | GENIUS Act mandates 1:1 T-bill reserves -> stablecoin growth = legislated T-bill demand; Tether >$100B (>UAE/Germany); Bessent's stated motive; China exits to gold; ~$900B demand-supply gap | research/macro-stablecoin-treasury-rail.json |
| Timing of the unwind | STRUCTURE certain (insolvent at zero inflow; marks must reverse), DATE unforecastable in principle (reflexivity/Minsky/Keynes); watch the trigger panel, not the calendar | research/spec-unwind-timing.json |
| Digital-ID convergence (Worldcoin→eIDAS→Digital Euro→Chat Control→OSA) | Five tracks rebuild ONE identity+payment+scanning stack from five angles; each sold on child-protection/anti-deepfake; EUDI wallets + Digital Euro + Chat Control trilogues all land 2026-27. The systems leak on arrival (Discord/Persona ~70k IDs + watchlist/biometric pipeline; EU age app broken in <2min; the Commission's own Europa breach) — the empirical futile-under-breach case | research/digitalid-worldcoin-eid-convergence.json |
| Digital-ID orchestration — the REAL incentive | Who coordinates it (World Bank ID4D, UN DPI, India-Stack/G20, BIS unified ledger, WEF, TBI) and why beyond the defensible frame: BIS's Carstens on record that CBDC gives 'absolute control … technology to enforce'; fiscal repression + access-chokepoint + AI-transition + de-anonymization | research/digitalid-orchestration-real-incentive.json |
| Identity-theft supply-chain wave | Shai-Hulud self-propagating npm worm (+2.0, 25k repos), Team PCP/Mini Shai-Hulud (UNC6780), the Scattered LAPSUS$ Hunters OAuth-token extortion (700+ orgs via Salesloft/Drift), and the Nightmare-Eclipse anti-Microsoft zero-day grudge — one mechanism: steal the trust attached to an identity | research/spec-supplychain-shaihulud-extortion.json |
| Threat-actor attribution map (UNC / DPRK) | How Mandiant/Google UNC→TEMP→APT/FIN graduation works (epistemic-honesty device), and where DPRK's financially-motivated state hacking sits — TraderTraitor/UNC4899, Bybit $1.5B, Contagious-Interview npm overlap | research/spec-unc-nk-attribution-map.json |
| AI-capex depreciation & duration-mismatch trap | Useful-life as the 5th self-marked number (Burry ~$176B understated depreciation 2026-28; Meta 5.5yr, MSFT 6 vs 3yr); Oracle FY2026 -$23.7B FCF + $248B 15-19yr leases vs ~3yr GPUs; formalized in models/z3/depreciation_trap.py (D1-D4). From the justdario article | research/fin-ai-depreciation-debttrap.json |
| Critical-minerals supply-response equities | The allied answer to the China rare-earth + ANTIMONY chokepoint: MP (DoD largest shareholder + $110/kg NdPr price floor + Apple $500M), Perpetua (only mined US antimony, EXIM $2.9B, 2029), US Antimony (only NA mil-spec smelter), Ucore (DoD REE separation) + the stockpile-pumped tail. Quasi-nationalization = mirror of the AI-capex circular financing | research/macro-critical-minerals-equities.json |
| Contested resource states (force/territory/money/denial) | Venezuela (US captured Maduro Jan-2026; oil/gold/REE), Greenland (Trump 'framework'; REE + Pituffik/GIUK Arctic), Argentina ($40B US backstop for lithium vs China's Ganfeng), Iran (2026 war; Hormuz blocked; cutting China off the oil/REE axis) - the chokepoint contest gone kinetic | research/geopolitics-contested-resource-states.json |
| Allied intel & strategic-geography nodes | Physical chokepoints of intel/compute/data: Pine Gap signals (US spy-sats over 1/3 of globe), Korea/Japan HBM + Taiwan packaging (the allied memory chokepoint), Singapore cables vs Hong Kong's bifurcation, and the Greenland/GIUK Arctic frontier | research/geopolitics-allied-intel-nodes.json |
| Taiwan / TSMC silicon shield (the keystone) | ~95% of advanced chips + nearly all advanced packaging (CoWoS) on one island under China gray-zone/quarantine pressure - the master compute single-point-of-failure. Arizona ($165B) is a hedge, not independence: wafers still fly to Taiwan for packaging, runs behind the leading edge, inputs via Japan (materials) + ASML (EUV). The AI build is a bet the Strait stays calm | research/geopolitics-taiwan-silicon-shield.json |
| Chip-chokepoint war: equipment vs materials | West controls the equipment (ASML EUV monopoly + Zeiss optics + Japanese tools; reported kill switch on Taiwan's machines); China controls the materials (gallium/germanium/antimony/heavy-REE + an extraterritorial 0.1% rule) - mutual deterrence. Sticky NOT permanent: China's multi-track indigenous EUV (LDP, Tsinghua SSMB), Canon nanoimprint, and the tacit-knowledge wall espionage can't copy | research/geopolitics-chip-chokepoint-war.json |
| Russia - energy, nuclear fuel & the Arctic | Gates HALEU (~44% of enrichment, the SMR fuel; Rosatom un-sanctioned), controls the Northern Sea Route + the only nuclear icebreaker fleet, and routes oil to China/India via a ~100-vessel shadow fleet. Sanctions re-routed the levers, didn't close them | research/geopolitics-russia-energy-arctic.json |
| AI power & grid bottleneck | Power-constrained, not just capital-constrained: ~$650B+ 2026 capex meets ~127-week transformer lead times (shortage to 2029), sold-out gas turbines, multi-year interconnection queues, Russia-gated nuclear fuel - ~half of US datacenter projects risk delay. BYOP privatizes the energy chokepoint; idle GPUs compound the depreciation trap | research/macro-ai-power-grid-bottleneck.json |
| Physical information layer (cables + space) | ~99% of intercontinental data on undersea cables being cut (Baltic/Taiwan/Red Sea) by deniable shadow-fleet/anchor-drag vs a too-small aging repair fleet; orbital ISR/comms concentrating into one private actor (SpaceX Starshield/NRO). The data substrate is itself a contested chokepoint | research/geopolitics-cables-space-layer.json |
| Defense-industrial base (arsenal + materials + defense-tech) | US arsenal emptied by Ukraine + the 2026 Iran war (~45% of PrSM, ~half THAAD/PAC-3); 155mm stuck ~40k/mo vs a 100k goal; rebuild gated by China's REE/antimony (no F-35s/missiles) - 'needs Beijing's permission to reload'. Plus the AI-core-funded defense-tech wave (Anduril/Palantir/SpaceX ~88%; a16z American Dynamism) and its narrative-beta split | research/geopolitics-defense-industrial-base.json |
| Efficiency counter-thesis (our own thesis, doubted) | DeepSeek (frontier at a fraction of cost; Nvidia -$600B/day) could mean over-built - or via Jevons, justified. The tell: 2026 capex grew to ~$602B anyway, so the narrative justifies building under OPPOSITE facts (unfalsifiable, bubble-shaped). China routes around the chokepoint via efficiency + a domestic stack (Huawei Ascend). The proofs don't depend on demand being real; honesty says it's genuinely uncertain | research/fin-ai-efficiency-counter-thesis.json |
| US fiscal trap (the macro root) | Debt ~$38-39T; FY2026 interest (~$1.0T) now EXCEEDS defense (~$947B), ~$88B/mo = defense+education. An over-indebted sovereign has only repressive exits (inflate away, captive buyers, capped real yields) - which is WHY the stablecoin->Treasury rail + programmable-money/digital-ID enforcement layer are rational for the state. Interest>defense makes the control rail rational, no cabal required | research/macro-us-fiscal-trap.json |
| Datacenter water, siting & the orbital escape | Past power: water (~200-300B gal/yr; 2/3 of 809 planned on drought land), community consent ($18B blocked/$46B delayed; Maine's 1st statewide moratorium), and household bills (Dallas Fed +~50% wholesale; $100->$281). The orbital-datacenter escape (SpaceX et al.) is graded a moonshot (heat rejection in vacuum, launch cost, servicing depreciating GPUs) | research/macro-ai-datacenter-water-siting.json |
| Gulf sovereign capital (who funds the loop) | The concrete identity of the 'external capital' the proofs require: Abu Dhabi's MGX (>$100B AUM target) now backs all three frontier labs (OpenAI/Anthropic $380B/xAI) + Stargate UAE (1GW); Saudi PIF/HUMAIN ($100B, 1.9->6.6GW) races it. The price of US chips is cutting China (G42 divested; IGAA/RTE bars Chinese access). If the petro-sovereigns pause, the zero-inflow insolvency case (T4) goes live | research/fin-gulf-sovereign-ai-capital.json |
| OpenAI's conversion & the $1.4T-vs-$25B gap | The Oct-2025 for-profit conversion (OpenAI Group PBC; nonprofit controls 26%; Microsoft 27% ~$135B; Delaware AG cleared) legally unlocked the mega-capital. The number it raised against: ~$1.4T of compute commitments vs ~$25B revenue (~56x) - every counterparty booked a slice. Exactly where the proofs bite (T3: needs >=$1.03T external; T4: insolvent at zero inflow) | research/fin-openai-conversion-governance.json |
| India - the swing power | Courted by both blocs, captive to neither: wrote the DPI/Aadhaar blueprint the West copies, building chips (Tata/PSMC Dholera, Micron Sanand - behind but moving), keeps discounted Russian oil (~33% of crude), and signed a US critical-minerals pact vs China - taking chips+minerals while keeping oil+BRICS optionality | research/geopolitics-india-swing-power.json |
| The operator-network (not one cabal) | A small recurring network - Thiel/PayPal & a16z + Altman + Musk + the Gulf funds - now holding founder, funder AND official positions across the AI core, defense-tech (Palantir Pentagon backbone, Anduril), crypto (Sacks czar, Bitcoin Reserve), digital-ID (World/Persona) and the executive branch (Vance VP, Musk/DOGE). POSITION not shared intent - the members feud (Musk is suing OpenAI/Altman; some frame their moves as defensive), which is why it's NOT a cabal; no malign intent assigned to anyone | research/influence-operator-network.json |
Primary sources
Sample of the government and court records cited:
- US Supreme Court — Murthy v. Missouri (standing)
- US Treasury / OFAC — Tornado Cash delisting (SB0057)
- 5th Cir. Court of Appeals — Van Loon v. Treasury opinion
- US DOJ / SDNY — Samourai founders plead guilty
- US SEC — Kraken staking settlement (2023-25)
- US Congress — Epstein Files Transparency Act (H.R.4405)
- ICIJ — Pandora Papers
- Hedera Council — Governing-council roster
- NIST CSRC — Post-quantum FIPS 203/204/205
- US FDIC — FOIA 'pause letters' coverage
- US BLS — Preliminary payroll benchmark (-911k)
- SSA — Boskin Commission report (1996)
- US BLS — New Tenant Rent Index (leads CPI ~1yr)
- SEC EDGAR — CoreWeave 10-K FY2025 (67% Microsoft, $21B debt)
- SEC EDGAR — NVIDIA 10-K FY2025 (customer concentration)
- LME — CME-LME copper arbitrage (education)
models/audit.py (cross-document number/coverage checks) and models/cross_review.py (edge-amount conflicts, connectors) run via scripts/new-research.sh. Current audit: 0 flags.
Source index (78 cited research blocks)
research/age-verification-abolition.mdresearch/commodities-metals.mdresearch/digitalid-corporate.mdresearch/digitalid-orchestration-real-incentive.mdresearch/digitalid-regulatory.mdresearch/digitalid-worldcoin-eid-convergence.mdresearch/fin-ai-depreciation-debttrap.mdresearch/fin-ai-efficiency-counter-thesis.mdresearch/fin-coreweave-oracle.mdresearch/fin-disney-openai-sora.mdresearch/fin-google-amazon-anthropic-meta.mdresearch/fin-gulf-sovereign-ai-capital.mdresearch/fin-hedera-connections.mdresearch/fin-meta-family.mdresearch/fin-microsoft-openai.mdresearch/fin-nvidia-openai.mdresearch/fin-openai-conversion-governance.mdresearch/fin-sealsq-wisekey-global.mdresearch/geopolitics-allied-intel-nodes.mdresearch/geopolitics-cables-space-layer.mdresearch/geopolitics-chip-chokepoint-war.mdresearch/geopolitics-contested-resource-states.mdresearch/geopolitics-defense-industrial-base.mdresearch/geopolitics-india-swing-power.mdresearch/geopolitics-russia-energy-arctic.mdresearch/geopolitics-taiwan-silicon-shield.mdresearch/influence-ad-censorship.mdresearch/influence-china-tech.mdresearch/influence-congress-funding-compromise.mdresearch/influence-meta-childsafety.mdresearch/influence-operator-network.mdresearch/influence-tbi-policy.mdresearch/macro-ai-datacenter-water-siting.mdresearch/macro-ai-power-grid-bottleneck.mdresearch/macro-bank-hierarchy.mdresearch/macro-bank-htm-marks.mdresearch/macro-carry-trades.mdresearch/macro-cre-privatecredit.mdresearch/macro-critical-minerals-equities.mdresearch/macro-cross-sectional-analysis.mdresearch/macro-crqc-quantum-landscape.mdresearch/macro-fdic.mdresearch/macro-fed-trap-regional.mdresearch/macro-firstbrands-ubs.mdresearch/macro-futures-vs-physical.mdresearch/macro-gig-labor.mdresearch/macro-gold-silver-reprice.mdresearch/macro-history-dereg-manipulation.mdresearch/macro-jobs-inflation-fed.mdresearch/macro-official-data-integrity.mdresearch/macro-private-credit-marks.mdresearch/macro-quantum-computing.mdresearch/macro-rent-cpi-divergence.mdresearch/macro-stablecoin-failures-manipulation.mdresearch/macro-stablecoin-treasury-rail.mdresearch/macro-us-fiscal-trap.mdresearch/sec-filings.mdresearch/spec-cambodia-scam-fara-paid-influence.mdresearch/spec-crypto-banking-debanking.mdresearch/spec-disclosures-surveillance.mdresearch/spec-exchanges-asia.mdresearch/spec-insurance-bermuda.mdresearch/spec-onchain-threat-actor-addresses.mdresearch/spec-sec-filings-primary.mdresearch/spec-sec-sdny-regulatory.mdresearch/spec-supplychain-shaihulud-extortion.mdresearch/spec-telecom-satellite.mdresearch/spec-tornado-samourai.mdresearch/spec-unc-nk-attribution-map.mdresearch/spec-unwind-timing.md
Reports: report/UNMASKING.md · report/TEMPORAL-WEB.md · report/EXECUTIVE-SUMMARY.md. Every figure's URLs are in the matching research/*.json.