Whether the enterprise agent-product surface — the layer where labs and adjacent vendors package existing model capability into something a company actually deploys (seats, IDE integrations, workflow bundles) — keeps shipping at the pace 2026-08-20 revealed. Five items landed in one day (08-20), all on this exact axis, and only one (Mistral's Agentic Search) had an existing thread. Track: whether the pace holds or 08-20 was a one-off pileup; whether any lab treats this as a distinct product line with its own roadmap rather than a bundling exercise; and whether enterprise adoption numbers (seats, IDE install counts) ever get disclosed to test whether the packaging actually converts to usage.
Whether AI-buildout debt is becoming its own underwritten, rated, publicly-watched risk category rather than a footnote inside each borrower's story. Three data points converged on 2026-08-10 at three very different scales: Bloomberg Opinion launched "AIndicators," a credit-market framework built specifically to flag whether lenders are mispricing AI infrastructure; JPMorgan led a $441M debt facility for Global AI, a two-year-old AI-infrastructure firm; and Nvidia's $500B+ MOUs propose special-purpose-entity bonds collateralized by GPU capacity itself. Track: whether rating agencies publish an explicit methodology for compute-collateralized paper, whether any of it prices wide of comparable real-asset debt, the first covenant breach or restructuring, and whether Nvidia's CDS (already record-wide on 07-27) moves with issuance volume. The open question this thread exists to answer is whether GPU compute behaves like an infrastructure asset with a 20-year life or like equipment on a 3-4 year depreciation clock — the two produce very different recovery rates in a default.
Reported confidential IPO filing at ~$965B (single-source — thin, needs a second source) plus an IR-director hire read as fall-2026 prep. Track: filing confirmation or flip-to-public, valuation vs the $965B print, and how the Opus 5 cycle + the AMD/Azure/Google-TPU compute deals get positioned in the roadshow narrative.
Anthropic has signed four large infrastructure arrangements in roughly three months — SpaceX (~$45B), Volta ($10B, six-year, a 133MW Nvidia Vera Rubin site in Norway), Riot Platforms ($9.1B, 20-year, announced 2026-08-10) and, the same day, the Theseus Infrastructure venture with Macquarie Asset Management and Singapore's GIC, which will OWN purpose-built data centres and lease them back with Anthropic as anchor tenant. The pattern worth naming: it is renting and anchoring rather than owning, so the capex sits on someone else's balance sheet while the lease obligation sits on its own. Track: total contracted capacity and its implied annual lease cost against a company that has confidentially filed an S-1; whether any arrangement converts to ownership; the counterparty mix drifting from clouds toward miners and infrastructure funds; and the Theseus consumer-electricity commitment, which is the first contractual concession this map has seen a lab make to ratepayer politics. Read against ai-circular-financing-risk, which asks whether this money is circular — this thread asks a narrower question, who ends up holding the asset.
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 Google TPU deal — diversification, not displacement); OpenAI $50B investment + separate $100B/8yr compute deal CONFIRMED (02-27 announcement). Rainier live (500K Trainium2, $26B IN); 1.4M Trainium deployed; Talen 1.92GW/17yr + X-Energy 5GW-by-2039. Watch 07-30: capex guide, backlog, FCF (reportedly borrowing $25B).
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 here, not a new thread): Samsung/SK Hynix/Micron expansion responses, CXMT's state-funded push (charging ABOVE Samsung; new modules with no price relief; DoD ban overhang), and who blinks first on capex.
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 circular-financing stakes (Nvidia's 10-Q investments in its own buyers).
Nvidia's turn from chip vendor to the buildout's lender of last resort: the $250-500B OpenAI financing guarantee (figure unresolved), equity stakes across its own customer base (Nebius 9.3%, Naver $1B, Intel $5B, Groq ~$20B licensing), and the first market pushback (NVDA -4%, CDS wider, 07-27). Track: the guarantee's actual structure and figure, whether rating agencies treat it as debt-equivalent, and each new customer-equity position. Sits inside the circular-financing story but is its own actor-move. ADDENDUM 2026-08-15: the guarantee's structure/figure question is now largely resolved — under $120B, phase-1 only (down from up to $750B in talks), with signing reported as possible "this weekend." Retarget the watch toward: does the weekend signing happen, does a phase-2 guarantee follow, and rating-agency treatment (still untracked).
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; pre-construction, DRI filed 07-24, Coastal Regional Commission review (≤30d) pending. DISTINCT from SoftBank's ~10GW Ohio (Portsmouth) campus. Track: the local secrecy backlash (10K+ petition), generation-mix disclosure, and whether a GC/financing partner surfaces.
SpaceX (Nasdaq: SPCX) — the public parent holding xAI + X ("SpaceXAI" is only the AI subsidiary). IPO'd 06-12 at $135/~$1.75T (record ~$75B raised), peaked ~$2.6T 06-16, now ~$1.56T (07-23, BELOW the IPO price). The cash engine is Starlink (~$4.4B FY25 op income), funding Starship R&D + xAI's Colossus burn ($7.7B capex in Q1-26 alone) — the IPO is the AI financing bridge. Watch: the Aug-6 lock-up (911.5M shares); Musk's ~82% voting control; whether the Anthropic lease-back is real revenue or circular financing. Q7 material.
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 production date), Apple (NEW — $30B thru 2031), ByteDance (stale 2024 sourcing, unconfirmed) — the explicit COUNTER-case to inhouse-silicon's diversification thesis: "in-house" is largely outsourced to one designer. ②The VMware annuity turned adversarial: Tesco pulled 40,000 servers and is suing >£100M. Capex ~10% of D&A — designs, doesn't build. Track: Jalapeño tape-out, the sixth customer, the litigation spread, next earnings (~early Sept, unconfirmed).
$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 1M-TPU deal is the clearest volume signal). Solar-dominant PPAs + first Gen-IV utility deal (TVA/Kairos, 2030). Payoff case ($514B backlog, Cloud +82%) is press-corroborated, primary filing unreached. Watch: Q3 guidance, the FCF-negative reaction, TPU externalization (Meta eval).
The Grok line closing the frontier gap as a FAST FOLLOWER, not mid-pack: Grok 4.5 (07-08) posted the biggest single-gen leap on the AA Intelligence Index (38→54), 4th at launch, strong on coding (76, ties GPT-5.5) and price ($2/$6, 500k context) — the real weakness is factuality (~54% hallucination). Track Grok 5 (training on Colossus 2; ship date slipping Q1→Q3+ amid Musk hype), the $60B Cursor bet arming coding, and whether owning the mines sustains a lead vs. OpenAI/Anthropic.
Microsoft shipped 7 in-house MAI models and says it was "set free" from the OpenAI contract to pursue superintelligence — while still holding ~27%/$135B of OpenAI and embedding Anthropic's Claude in Copilot. Track how far the in-house models displace OpenAI in Copilot, the MSFT–OpenAI commercial terms, and whether this is genuine independence or a hedge.
~$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). Track sites actually energized, financing/grid bottlenecks, sovereign terms, and slippage.
Commercial datacentres are now deliberately struck in wartime, with published damage assessment — the first time. Iran hit AWS sites in the UAE (03-01) and Bahrain (04-01), claimed further strikes ~07-21, and satellite imagery corroborated damage at Zallaq and Askar on 07-28; Iranian state media named Microsoft, Google, Apple, Meta, Oracle, Intel and Nvidia as targets on 03-31. This is a STRUCTURAL VARIABLE, not an episode: it puts a war-risk term on siting, insurance and sovereign exposure across every capex thread we run, and none of them price it. Watch: further strikes or credible threats; whether any hyperscaler discloses it in a filing or on a call (Amazon reports 07-30 — the first live test); war-risk underwriting and premia for datacentre assets; siting decisions that cite geopolitical risk; and whether the precedent outlives this particular war, which is the thread's main mortality risk.
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, Apple/Nvidia named. The REAL constraint is CoWoS packaging (Nvidia holds >half of 2026-27), not wafers; Taiwan keeps ~85% of sub-5nm through 2030 (MOEA). China friction now runs BOTH directions (Taiwan export curbs weighed; China reportedly weighing a local-use ban, 07-21). "Nvidia overtook Apple as top customer" is contested, not settled. The ~90% advanced-node share is carried, not re-verified. Next: Q3 earnings.
At WWDC 2026 (06-08) Apple shipped Siri AI on Apple Foundation Models "custom-built in collaboration with Google and its Gemini models" (Apple's own wording), partly hosted on Google's cloud on Nvidia Blackwell GPUs under Private Cloud Compute. Deal terms remain unconfirmed by either company (~$1B/yr is reporter-sourced only, never officially stated). The real regulatory angle is the EU: Apple withheld Siri AI from EU iPhones at launch (06-09) citing DMA compliance after the EU rejected its exemption request — not US antitrust. Track: EU launch resolution, whether the $1B/yr figure ever gets confirmed, whether Apple builds back toward its own frontier model.
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 (Broadcom, Marvell). Watch whether displacement reaches frontier training.
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 ran a soft-demand-then-upsized-to-$12.5B arc in one week. MTIA at "millions" of H100-equivalents, "Iris" to production Sept-26, Broadcom extended thru 2029. Superintelligence framing softening ("more slowly than expected"); Moody's credit warning 07-24. The $76B TTM basis is retired. Meta's own 07-29 filing gives Q2 capex $31.08B and FY2026 guidance NARROWED to $130-145B (from $125-145B); revenue $60.8B +28%, EPS $6.18 vs $7.17 est, shares -7-10% after hours. ADDENDUM 2026-08-18: Meta carries a reported ~$420B of AI obligations OFF the balance sheet against $83.7B of reported debt — future leases incl. the $27.3B Beignet/Hyperion structure with Blue Owl, chip/equipment purchase obligations, and the $35B CoreWeave commitment; five hyperscalers cited at ~$1.65T combined. EY flagged Beignet as a critical audit matter Feb-2026. Read the guidance as what Meta will SPEND, not what it has COMMITTED — track the gap.
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 follow, the emissions accounting fallout, whether the gas capacity is bridge or base, and regulatory/PR consequences.
Ping An running AI-driven underwriting and claims at genuine national scale (~60% of accident/health claims automated, some in 51 seconds; underwriting review cut to ~1.5hr; 93% of new auto policies auto-issued) against a fresh sector-wide NFRA framework (2026-06-18) that puts AI-decision liability on the insurer, not the algorithm. Track: any Ping An-specific regulatory scrutiny under that framework, the pending August 2026 interim results, and whether the automation numbers keep climbing.
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 + environmental-justice fights. Watch the DOJ national- security shield over xAI's unpermitted turbines — the improvisation has outrun the regulatory system.
Reported slip to 2027 chasing a ~$1T valuation; Altman rejecting a lower price. Track confirmed timing, valuation target, and macro/AI-stock spillover (SoftBank fell ~12% on the report).
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.
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 china-stack-independence.
The French national champion running two tracks at once: a fast, broadening product cadence (chat models into robotics, formal math, safety tooling) and a capital-plus-institutional-anchoring push (a pending €3B/€20B round, a French defense framework, Airbus/BMW industrial deals). Both serve the same pitch — "open and European" as a market position that doesn't require beating GPT-5.5/Gemini/ Claude or the fastest Chinese open-weight labs on the leaderboard. Track whether the ~€3B round (Samsung + EQT, in talks since mid-June) closes and at what terms, whether the Microsoft EU-infra deal ever discloses a dollar figure, and whether growing Microsoft compute-channel dependence complicates the "sovereign alternative" framing. Cross-refs asml (the ASML stake lives on that thread).
Whether the ~14GW of announced OpenAI/Meta/Anthropic compute deals actually convert to shipped hardware and vested warrants on their stated H2-2026/H1-2027 timelines, or stay a backlog/announcement story — Q2 2026 earnings (reported 2026-08-04) is the first real test. Also watching: China MI325X licensing status, and whether the Oracle 50,000-GPU Q3-2026 rollout starts on schedule.
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.
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 constraints, and whether owning the mines translates into a frontier Grok. Feeds the "who controls the stack" thesis (radar Q1).
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 2030s-weighted. Watch which actually clear licensing.
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 above junk) or a CDS move; whether the guided RPO-to-revenue conversion ($76.6B vs $67B delivered in FY26) holds; any named counterparty breakdown of the OpenAI-linked RPO slice ($638B total, roughly half tied to OpenAI per multiple outlets).
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 trajectory, any unwinding or write-downs, GPU-demand softening, and credit/equity-market reaction. A correction here hits everything in radar Q1/Q2. ADDENDUM 2026-08-15: the loop now has a public, named-investor bear case — Michael Burry sized it at ~$879B and called the $500B Wall Street financing platform an "Enron"-style stunt, reportedly deepening his Nvidia short (08-11/08-14). Its own lender has also started pulling back rather than expanding: Nvidia cut its OpenAI Ohio guarantee from up to $750B in talks down to under $120B, phase-1 only (08-15). Track whether other lenders follow Nvidia's retreat.
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 four *-capex findings). The standing questions: whether AI revenue catches the spend (payoff layer is press-only), the undisclosed own-silicon vs Nvidia splits (4/4 refuse), power as the binding constraint (nuclear '27-'39 vs Meta's gas pivot), and capex migrating off balance sheets (Blue Owl, bonds, leases → the financing threads). Earnings Wed/Thu (Meta+MSFT 07-29, AMZN 07-30) are the live test.
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, Colossus + the 100s of sites), power (gas turbines, nuclear-for-AI, grid), land/cooling/water. Sibling to hyperscaler-capex-big-picture, which cuts the SAME money by spender; this one follows it to where it physically lands. Most destinations have no thread yet — that's the worklist. ADDENDUM 2026-08-15: the capex-committed-vs-capacity-energized gap now has its first two real outcomes on the record, on opposite sides — AWS's Lusby, MD project withdrew rather than clear Texas- style compliance gates (08-04), while Core Scientific, Vantage, and SB Energy formally committed to Abbott's PUCT/ERCOT standards (08-12). Track which outcome the majority of the project queue lands on, not just whether the gap exists.
The neocloud wager, quantified: $99.4B backlog (Q1-26; OpenAI $18.4B + Meta $21B/to-2032 + Jane Street $6B) vs revenue $2.08B and a $740M quarterly loss — everything rides on conversion. Financed at 9.75% GPU-collateralized just as Nvidia CDS widened (07-27) and Moody's named CoreWeave+Oracle "AI's weakest credit link" (07-25). Disclosure RETREATED: customers anonymized to A-D in the 10-Q (S-1 said 77% from top two). Nvidia's "~11%" stake is journalist-derived, 13F-only. Core Scientific rejection left a 4.5-month capacity gap; plan B is thin. Next print: Q2 earnings 08-11.
Guidance is climbing (€36-40B to €43-45B for FY2026) while the China fight runs on three fronts at once: the MATCH Act would extend controls to DUV tools generally, not just EUV, and passed House committee — but its vehicle is STALLED: MATCH rode as an amendment to the FY2027 NDAA, and the Senate failed cloture on that bill 50-46 on 2026-07-14, blocked by Democrats over Iran-war and Israel provisions rather than anything to do with chips. Thune switched to "no" to preserve a motion to reconsider, so it is stalled rather than dead, and the reconsideration vote is now the real next trigger — if it clears, MATCH's 150-day clock for the Netherlands to align its export controls starts running; Commerce's Lutnick opened a still-unresolved inquiry into whether an EUV tool illegally reached China (ASML denies it); and China's own Shanghai Aishengna is now shipping domestic DUV units to SMIC/Hua Hong/CXMT, though independent forecasts put a genuinely competitive 7nm domestic scanner a decade out, not this year. High-NA adoption is splitting the big three — Intel shipped first, TSMC opted out for A14 on cost, Samsung/SK Hynix are ordering. The Mistral stake (11%, active board seat) has no confirmed follow-on yet. Next: Q3 2026 earnings, MATCH Act's Senate/NDAA fate, Lutnick inquiry resolution.
Whether China's first mass-produced immersion DUV tool is real at manufacturing scale, or an announcement. Reuters (07-28) names Shanghai Aishengna (state-owned) as the maker: ~5 units in 2026, ~20 in 2027, with SMIC, Hua Hong and CXMT the expected first recipients — explicitly "far from matching" ASML and unproven at high-volume manufacturing. THE TEST IS DELIVERY, NOT PRODUCTION: a tool installed and running at a named fab, at a stated node and yield. Track the unit count against the 2026 target, which fabs actually receive them, any yield/uptime disclosure, ASML's order book and guidance for China, and whether export-control policy re-prices on it. A slip here is as informative as a hit — this is the load-bearing assumption under the whole chokepoint picture. ADDENDUM 2026-08-18: 21 days on, the delivery test is STILL UNMET — no confirmation from Aishengna or from SMIC/Hua Hong/CXMT of an installed running tool at any node with any yield figure. HARD QUALIFIER now on record: the tools support 28nm-class SINGLE-EXPOSURE patterning only, advanced nodes need multi-patterning, some critical components still internationally sourced. ASML fell up to 14% on the 07-27/28 report, then RAISED its own 2026 guidance to EUR36-40B with DUV immersion shipments ~130 systems and an order book "largely filled through 2027" — the market repriced, the company did not. No export-control change tied to this. NAMING: "Shanghai Yuliansheng"/"Yuliangsheng" is a PREDECESSOR team absorbed into Aishengna (alongside SMEE), not a rival programme — do not double-count.
Demis Hassabis is stepping down as DeepMind CEO to become chairman and Alphabet's Chief Scientist; Koray Kavukcuoglu is promoted to DeepMind SVP, reporting directly to Pichai. Separately, Jeff Dean — Google's longest-tenured engineering leader, 27 years — is leaving to found "Discovery Loop," an AI-for-science startup Google itself is backing. First leadership change at the top of a frontier lab this map has tracked. Watch: whether this reads as succession planning or a real strategy shift (does DeepMind's model pace or priorities change under Kavukcuoglu?); what Hassabis's chairman + Chief Scientist role actually does day to day vs. is honorary; what Discovery Loop builds and whether Google backing a departing exec's startup becomes a pattern; whether other frontier labs see similar moves.
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 ($7B, 2.7GW); Maia 200 confirmed DUAL-PURPOSE (OpenAI inference + MAI training) with the split undisclosed; TMI/Crane restart tracking 2027; $37B AI run-rate +123%. The $250B OpenAI commitment has NO disclosed capex mapping — nobody publishes the OpenAI-vs-own split. Watch Wednesday: FY26 full-year capex, FY27 guide, any split disclosure.
Qualcomm's datacenter re-entry, eight years after Centriq died: Dragonfly C1000 CPU + AI300 accelerator (Investor Day 06-24, stock +15%), Meta and Microsoft named launch customers, $15B datacenter revenue by 2029 target, paired same-day with the announced $3.9B Modular acquisition (close expected H2-26 — audit 07-28) (an explicit anti-CUDA software play). The Tenstorrent talks died (CEO denial within two weeks); Ventana/RISC-V status unclear — does Dragonfly supersede it? First live test: Q3 FY26 earnings 07-29 (tomorrow) — does Dragonfly move guidance?
The ~$800B-gravity ISA chokepoint: SoftBank's 87.1% leaned on as LOAN COLLATERAL, not sold. Post-Qualcomm-trial (Arm's last claim rejected 2025-10-01; appealing) the design-around threat is funded in the open — Qualcomm bought Ventana, chased Tenstorrent $8-10B. Arm answered by ending pure-rentier status: first in-house chip in 35 years (the "AGI CPU," Mar-26, Meta launch customer) — now competing with its own licensees. Royalties real (4 straight >$1B quarters, v9+CSS). Stock -28% into earnings 07-29 — the live test.
⚠️ 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 going negative (-$3M vs +$277M a year ago). Revenue $1.786B, non-IFRS GM 29.9%, a beat the market sold anyway. What is still open: whether the government's two new ~1%-scale equity stakes (quantum + photonics) keep accumulating toward an Intel-style larger position; whether Mubadala's stake keeps trimming below 73% even as its board influence deepens; and whether the ~45% stock decline since the May Mubadala sale gets explained or reversed by the Q2 print.
Allianz ranked #1 on the 2026 Evident AI Index for Insurance and signed a named global partnership with Anthropic (2026-01-09, Claude free to all employees, agentic claims AI for motor/health co-built with Anthropic). Project Nemo (live in Australia since Jul 2025) is already settling small claims ~80% faster. Track: whether the Anthropic-built claims agents or Nemo expand scope/geography, and any EU/Solvency-II regulatory response to agentic claims AI.
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.
