The AI buildout · 30 threads · 525 stories · 14 open predictions
The Treasury mechanism behind last week's "financial repression" finding got a name and a number — the TGA sits at $935-950bn, $350-400bn above the prior administration's target, reportedly funding the doubled buyback ceiling; framing hardened over the week from "repression" to "fiscal dominance." SoftBank's ¥1tn retail bond got firm terms (banks reportedly declined the institutional risk, forcing the retail route) and a second $10-20bn wholesale leg opened two days later with existing notes selling off same-day — two financing-stress signals on one entity inside 48 hours. Anthropic's IPO pitch escalated three times in five days (>$100bn raise → >$2tn valuation → a claimed >$30tn TAM) before drawing its first analyst pushback.
Working note of 2026-08-27 (62 days old; no newer one has been written). The threads, stories and predictions below are the live evidence behind it.
$1,741.6B counted across 330 flow claims in the corpus's graph, export of 2026-10-28 (corpus commit 52b068e).
Not counted: $99.5B across 27 deals still in talks, and 5 guarantees ($3.5B stated), recorded as who is liable and not as money paid.
Recorded transactions, not unique dollars: money raised and then spent can appear in both. An all-time total, and some claims are undated. It moves as the corpus files new claims.
| Capex | $556.2B |
| Capacity/service payment | $402.0B |
| Equity | $371.4B |
| Debt | $274.2B |
| Asset purchase | $126.1B |
| Grant / subsidy | $10.3B |
| Prepayment | $1.5B |
| Guarantee | — |
| Non-cash consideration | — |
Point-figure, USD-denominated flows only — click a row for the members and sourcing.
| N/a — financing | $802.4B |
| Land, shell & materials | $406.2B |
| Purchased compute capacity (services) | $398.2B |
| Memory & storage | $93.7B |
| Power infrastructure | $16.3B |
| Compute silicon & systems | $15.6B |
| Intellectual property / licensing | $8.7B |
| Networking & optics | $0.6B |
| Unnamed vendors (via TSMC's capex) | $294.4B |
| Unnamed vendors (via Anthropic's capex) | $125.2B |
| OpenAI | $122.0B |
| Unnamed vendors (via SpaceXAI's capex) | $119.6B |
| Amazon | $116.2B |
| Unnamed vendors (via Adani Group's capex) | $100.0B |
| Anthropic | $93.3B |
| Nscale | $53.8B |
| Softbank | $52.4B |
| Stargate | $52.0B |
| Unnamed vendors (via Micron's capex) | $49.4B |
| Microsoft | $44.9B |
| Unnamed vendors (via SK Hynix's capex) | $41.9B |
| SpaceXAI | $40.0B |
| Meta | $37.0B |
The data-center buildout is diverting wafer capacity to HBM — a reported THREE-YEAR HBM deficit — spiking DRAM/NAND prices and forcing consumer-hardware price hikes (Apple, Microsoft). Now also the capacity-race ledger (backlog W5: widened …
Earnings 07-30. As of 07-27: the AGI-cuts vs $200B-plan "contradiction" is coherent — retreat from frontier research, double down as the labs' landlord: $364B backlog EXCLUDING Anthropic's new $100B+/decade commitment (grew 10 days AFTER …
The demand side of the chip spend — a ~$1T Blackwell+Rubin book (a management aggregate, not audited backlog); the $100B OpenAI deal is dead, and the constraint has shifted to supply (HBM4). Watch the per- customer concentration and the …
OpenAI-direct build (Octans GA LLC/Butler Snow — no named GC or Stargate-JV backer): Savannah Gateway Industrial Hub, Rincon, Effingham County GA; 3.2GW/25yr Georgia Power PPA phased 2028-2032; $20-30B cost unreconciled across outlets; …
Broadcom's two tolls. ①The custom-ASIC co-design book — FIVE majors now: Google (a $21B TPU order + the Anthropic multi-GW deal), Meta (extended thru 2029; Hock Tan left Meta's board the same week), OpenAI (Jalapeño unveiled 06-24, no …
$195-205B guided for 2026 (07-22), destination traced 07-27: a 15+ state land-grab (MO $15B, MI 1GW "Cannoli", TX 1GW+, NextEra 3×GW-scale) + India $15B hub — not a few megasites. TPU v7 "Ironwood" vs Nvidia split UNDISCLOSED (Anthropic's …
~$500B / ~10 GW joint venture (OpenAI/SoftBank/Oracle/MGX); ~7 GW + >$400B committed over 3 years; SoftBank carrying a $40B OpenAI loan; sovereign co-investment via "OpenAI for Countries" (UAE live 2026; UK/Norway/Argentina/South Korea). …
The chokepoint is cashing in: capex raised 15% to $60-64B + another $100B onto Arizona ($265B total US, 12 facilities — no per-fab breakdown public) at the 07-16 Q2 print; 2027 price talks run 5-10% baseline + up to 25% AI surcharge, …
The custom-silicon revolt (TPU/Trainium/Maia/MTIA) — real but INFERENCE- only, anchored by Anthropic running mostly off Nvidia on Trainium + TPU. Nvidia's absolute demand still grows; the clearest winners are the co- design houses …
The week's reversal: Meta QUIT RE100 amid a 7.5GW gas buildout (07-23/24) and SOLD 80% of Hyperion (5GW, $50B+) to Blue Owl — capex migrating off its balance sheet while El Paso ($10B/1GW) faces live local pushback and the BlackRock bond …
Meta quit RE100 amid a 7.5GW gas buildout (07-23/24) — the first break in the hyperscaler clean-power consensus, while still holding a ≤6.6GW nuclear RFP and 2.3GW NextEra solar. Track: whether Google/Microsoft/Amazon hold their pledges or …
Where the capex lands in POWER — the real bottleneck. Nuclear-for-AI (SMRs, Oklo/X-Energy), gas turbines (the Colossus permit fights), grid deals, and the energy-justice/permitting battles. Category meta under where-the-capex-lands.
The grid is the binding constraint (not chips) — interconnection queues exploding, hyperscalers self-supplying via gigawatt-scale gas, an equipment ceiling (turbines sold out to 2030, transformers ~3yr), and the ratepayer + …
Up-to-$9.8B STAR Market listing (allotment 07-22, lists 07-27) funding Chinese DRAM/HBM capacity into the global squeeze. Track first-day reception, capacity-buildout plans vs. the shortage, and read-through to ai-memory-shortage and …
Where the capex lands in SILICON — the $ flowing to Nvidia/AMD/TSMC/ Broadcom, plus the custom-silicon revolt (Jalapeño, Trainium, Maia, TPU) as the giants try to stop renting from Nvidia. Category meta under where-the-capex-lands.
The systemic risk under the whole AI capital thesis: ~$800B+ of loops where Nvidia invests in/supplies OpenAI, OpenAI commits to Oracle/cloud, those providers buy Nvidia GPUs. OpenAI on track to lose ~$14B in 2026. Track the loss/revenue …
Where the capex lands in CONCRETE — the specific gigawatt campuses. Stargate, Colossus, Camellia + the 100s of sites. Category meta under where-the-capex-lands; the specific-sites worklist lives here.
The combined 2026 plan is now ~$610-650B across the four (Google $195-205B · Microsoft ~$100B+ FY26 · Amazon ~$200B · Meta $135-145B guidance) — destination TRACED per actor 07-27 (see the comparable table in this thread's timeline + the …
The nuclear-for-AI wave — but ~0 net-new MW power any AI datacenter before 2027, and the first firm power is a RESTART (TMI-1), not a new build. The ~13GW announced is <20% of projected demand; every genuinely new SMR/Gen-IV project is …
Whether the FY27 capex step-up (~$70-95B, depending on source) and a reported (thinly sourced) OpenAI target-miss show up in the 2026-09-14 Q1 FY27 print; any further credit-rating action beyond S&P's 2026-07-09 downgrade to BBB- (one notch …
SpaceXAI's self-owned training compute (Colossus, Memphis) — the "owns its mines" fact that makes it structurally distinct from OpenAI/Anthropic, who rent. Track the live GPU count + power draw, the reported scale-to-1M-GPU plan, grid/power …
The ~$750B question posed as a DESTINATION, not a spender: where does the aggregate AI capex actually land, on what projects? The taxonomy — chips (Nvidia/AMD/TSMC/Broadcom + custom silicon like Jalapeño), datacenters (Stargate, Camellia, …
Earnings 07-29 is the near-term test — the "$97B TTM" figure is derived arithmetic (unconfirmed; $80.1B 9-mo FY26 is the IR number). As of 07-27: WI+GA "AI superfactory" (~2GW, fiber-linked) + a distinct Pecos TX ~2GW on Chevron's Kilby PPA …
⚠️ RESOLVED at the Q2 2026 print (2026-08-05, already on this thread's timeline): the Q1 capex/depreciation near-parity did NOT hold — it widened the other way, capex $411M against $307M D&A (+158% YoY capex), with adjusted free cash flow …
OpenAI's first custom inference ASIC, co-developed with Broadcom (9-month cycle, ~50% claimed cost/GPU). Track end-2026 deployment, real perf/cost vs. claims, and the next-gen parts in the multi-generation platform.
517 older stories are on the threads above.
