07 — Machine finance, ratings, securitization, insurance

Reference entries for the organizations the thesis names in the financing / ratings / insurance sector. All figures researched and linked; market caps as of 2026-07-14 unless otherwise stated.

CoreWeave, Inc.

What it is: The AI "neocloud" that proved GPUs are pledgeable collateral — the thesis's central proof that a new machine class becomes bankable the moment lenders can see utilization, contracted revenue and residual value. Status: Public (NASDAQ: CRWV), IPO March 2025. Market cap / valuation: ~$43.3B at $79.34/share (2026-07-14, StockAnalysis). Note the stock is down sharply from its 2025 highs (~-47% over the trailing year). Latest financials: Q1 2026 revenue $2.08B (vs $981.8M in Q1 2025, +112% y/y); revenue backlog $99.4B; ~3.5 GW total contracted power; FY2026 revenue guidance $12–13B. TTM revenue $6.23B, TTM net loss ~-$1.59B. Cited for: The GPU debt ladder — $2.3B facility (Aug 2023, Magnetar + Blackstone); $7.5B (May 2024); and the maturation marker, the $8.5B DDTL 4.0 facility closed 2026-03-31, the first investment-grade-rated GPU/HPC-backed financing, rated A3 (Moody's) / A(low) (DBRS), floating tranche SOFR + 2.25%, fixed tranche ~5.9%, maturing March 2032, anchored by Blackstone Credit & Insurance, secured by CoreWeave Compute Acquisition Co. VIII assets and an associated customer contract (the ~$19B Meta contract). "GPU debt went from exotic to investment grade in ~30 months." Also the ~$28B of equity+debt commitments raised in the prior 12 months. (I.1; I ARGUMENT ledger table; III.6; V.1; V.2; VI.3) Links:

Lambda (Lambda Labs, Inc.)

What it is: The GPU cloud that ran the first GPU-backed asset-backed securitization — the thesis's template deal and its residual-value cautionary tale. Status: Private (Nvidia-backed; reported to be preparing a 2026 IPO). Market cap / valuation: Last disclosed primary round: Series E, over $1.5B, led by TWG Global with US Innovative Technology Fund, announced November 2025. Lambda declined to disclose a post-money figure; press reporting around the round put it in the $4–6B range, and secondary-market marks in mid-2026 (Forge, June 2026) imply ~$9B. Treat any specific valuation as a secondary-market estimate, not a company disclosure. Total disclosed equity: $320M Series C (Feb 2024) + $480M (2025) + $1.5B+ Series E (Nov 2025). Latest financials: Lambda does not publish financials. It has publicly disclosed only the financing structure and its GPU fleet build-out (H100/H200/Blackwell purchases funded by the facility). Could not source revenue or margin to a primary link as of 2026-07-14. Cited for: "Lambda closed the first GPU-backed ABS, ~$500M, led by Macquarie Group (announced April 2024), with H100s placed in a special-purpose GPU financing SPV backed by both the physical chips and the cloud-rental cash flows they generate." Also the cautionary half: the deal is described in the thesis (and in market commentary) as underwritten to an assumption of ~50% residual value at three years, which the Blackwell step-down broke — an H100 that cost ~$30k in 2023 has traded near ~$8k in 2026. Important caveat: the 50%-at-3-years figure is trade/analyst commentary on the deal, not a Lambda or Macquarie disclosure; the primary releases state the size, the SPV structure and the lead, not the residual assumption. (V.1 template deal #3; V.2(a) residual-value cautionary case; V ARGUMENT §2a, §5) Links:

Macquarie Group Limited

What it is: The Australian infrastructure-and-asset financier that led Lambda's $500M GPU facility — in the thesis, the incumbent balance sheet that decides when a new machine class is financeable. Status: Public (ASX: MQG). Market cap / valuation: ~A$93.3B at A$253.04/share (2026-07-14, StockAnalysis). Latest financials: FY26 (year ended 31 March 2026): net profit after tax attributable to ordinary shareholders A$4,847M, up 30% on FY25; net operating income A$19,477M (+13%); operating expenses A$12,748M (+5%); total FY26 ordinary dividend A$7.00/share. Cited for: Leading the ~$500M Lambda GPU-backed facility (April 2024), the first GPU ABS — the anchor for the thesis's claim that a mainstream infrastructure lender, not a crypto-adjacent one, wrote the first chip-collateralized securitization. (V.1 template deal #3; V ARGUMENT §2a) Links:

SolarCity (now Tesla Energy / Tesla, Inc.)

What it is: The residential-solar installer that ran the first securitization of distributed solar assets — the thesis's foundational proof that a machine class becomes bankable only once standardized production data exists. Status: Formerly public (NASDAQ: SCTY); acquired by Tesla, Inc. (NASDAQ: TSLA) in November 2016 for ~$2.6B in stock; SolarCity's business is now Tesla's energy generation & storage segment. Market cap / valuation: SolarCity has no independent market cap. Parent Tesla, Inc.: ~$1.48T at $395.35/share (2026-07-14, StockAnalysis). SolarCity's own last standalone value was the ~$2.6B all-stock Tesla acquisition (closed 21 Nov 2016). Latest financials: No standalone reporting since 2016; the business is folded into Tesla's energy generation & storage segment. Tesla group TTM revenue $97.88B, net income $3.86B (as of 2026-07-14). At the time of the 2013-1 deal SolarCity was a loss-making growth installer; the securitization was the financing innovation, not a profitability event. Cited for: SolarCity 2013-1 — the industry's first securitization of distributed solar. Priced 13 Nov 2013, completed 21 Nov 2013: $54,425,000 aggregate principal of Solar Asset Backed Notes, Series 2013-1, 4.80% coupon, scheduled maturity Dec 2026, rated BBB+ by Standard & Poor's, secured by a pool of 5,033 PV systems across 14 states with their leases and PPAs (71.1% residential / 28.9% commercial-government). The thesis uses this as template deal #1: single SPE, pooled PV + lease + PPA, single tranche. (III.6; V.1 template deal #1; V ARGUMENT §4 timeline "Nov 21 2013") Links:

Sunrun Inc.

What it is: The largest US residential-solar company and the dominant repeat solar-ABS issuer — the thesis's evidence that a securitization category, once opened, industrializes into routine, cheaper, larger deals. Status: Public (NASDAQ: RUN). Market cap / valuation: ~$3.03B at $12.70/share (2026-07-14, StockAnalysis). Latest financials: TTM revenue $3.17B (+52.4%); TTM net income $565.3M; P/E ~6.0 (as of 2026-07-14). In 2025 Sunrun raised ~$2.8B of non-recourse debt, including ~$2.4B of senior debt across five securitizations plus subordinated financings (Q3 2025 results, 6 Nov 2025); Q3 2025 alone: three securitizations, ~$1.4B senior non-recourse; the Sep 2025 $510M deal priced at 6.21% yield / 240bp spread. Cited for: The industrialization of solar ABS — the largest single residential-solar securitization on record, Sunrun's $886.3M senior deal; repeat issuers (Sunrun / Sunnova / Solar Mosaic) making up ~77% of issuance; Sunrun's ~$2.8B of non-recourse debt raised in 2025 across five securitizations; H1-2024 sector issuance >$3.3B across 11 deals at an average $298M. Advance rates on mature-asset ABS run 75–80%. (V.1; V ARGUMENT §4 timeline) Links:

Sunnova Energy International Inc.

What it is: The #2 US third-party-owned residential solar company and a top-three repeat solar-ABS issuer — and, since 2025, the sector's failure case: a securitization-funded fleet whose *sponsor* went bankrupt while the asset class survived. Status: Formerly public (NYSE: NOVA); filed Chapter 11 on 9 June 2025, delisted from the NYSE, equity wiped out. Substantially all assets were sold via a §363 process to Solaris Assets, LLC (an ad hoc bondholder group backed by affiliates of GoodFinch Management) for ~$118M, with operations transitioning to SunStrong Management. Market cap / valuation: No market cap — delisted June 2025. The relevant valuation marker is the ~$118M §363 asset-sale price against ~$8.9B of long-term debt at filing. Latest financials: At the Chapter 11 filing (June 2025): ~$8.9B long-term debt, ~$13.5M cash, >500,000 customers; >700 employees (>55% of workforce) laid off pre-filing; $90M DIP facility approved 12 June 2025 ($15M interim). The Trump administration's termination of Sunnova's $3B DOE loan guarantee — the largest federal solar commitment ever made — was a proximate cause alongside rates and the California policy change. Cited for: Named as one of the repeat solar-ABS issuers (Sunrun / Sunnova / Solar Mosaic ≈ 77% of issuance) whose repeat issuance industrialized the category; and, for the thesis's risk symmetry, the demonstration that a bankable asset class can outlive its sponsor — Sunnova's ABS collateral was sold and serviced onward while the equity went to zero. (V.1; V.2 risk framing) Links:

First Solar, Inc.

What it is: The rare western PV module maker that survived the Chinese commoditization flood — the thesis's existence proof that a differentiated technology position (CdTe thin-film) can outlast an order-of-magnitude price collapse. Status: Public (NASDAQ: FSLR). Market cap / valuation: ~$23.9B at $222.70/share (2026-07-14, StockAnalysis). Latest financials: FY2025 (reported 24 Feb 2026): net sales $5.2B (vs $4.2B FY2024, +24%), net income ~$1.5B, diluted EPS $14.21, record 17.5 GW of modules sold; net cash position $2.4B at year-end. TTM (to 2026-07-14): revenue $5.42B, net income $1.67B. This is the basis of the thesis's "2025 was its most profitable year ever." Cited for: Survivor of the module-price collapse — residential module ~$8.70/W (2010) → <$0.10/W by Q3 2024; "nearly every western maker died; First Solar the rare survivor… 2025 was its most profitable year ever." (III.6; III ARGUMENT §"Existence proof") Links:

Switch, Inc. (Switch, Ltd.)

What it is: The largest single issuer of data-center ABS — the thesis's evidence that the machine-securitization pattern (SolarCity → CoreWeave) has already generalized to a third asset class. Status: Private. Taken private on 6 Dec 2022 by DigitalBridge Group (DigitalBridge Partners II) and IFM Investors for $34.25/share in cash, ~$11B including debt repayment (announced 11 May 2022). Formerly NYSE: SWCH. Market cap / valuation: Last disclosed: ~$11B enterprise value at the December 2022 take-private. No public mark since; Switch has not disclosed a subsequent valuation. Latest financials: Switch does not publish financials as a private company. Disclosed operating scale from its ABS programme: after the Series 2026-1 issuance (~$768M, closed 14 April 2026), its securitized pool comprises 11 data centers across five US markets serving 550+ customers; all Switch ABS issuances qualify as secured green bonds. Prior deals include a $3.5B securitized debt financing (announced 2025) and a $659M fourth ABS offering. Cited for: "Switch is the largest issuer of data-center ABS, ~$4.2B across five deals since 2024" — the anchor for the thesis's claim that data-center ABS is a fast-forming category (~$40B+ outstanding projected by 2027, per DoubleLine/Morgan Stanley). (V.1) Links:

Triton International Limited

What it is: The world's largest intermodal-container lessor — the thesis's canonical "neutral operator of a commoditized machine," the endpoint every fleet-financing argument in Parts IV and V points at. Status: Private. Acquired by Brookfield Infrastructure (NYSE: BIP / BIPC) in a take-private that closed in 2023; formerly NYSE: TRTN. Market cap / valuation: $13.3B take-private, announced 12 April 2023, at $85.00 per Triton common share ($68.50 cash + $16.50 in BIPC class A exchangeable shares), a 35% premium to the 11 Apr 2023 close and 34% to the 30-day VWAP. Parent Brookfield Infrastructure Partners market cap ~$17.65B at $38.11/unit (2026-07-14, StockAnalysis); BIP TTM revenue $24.01B. Latest financials: No standalone public reporting post-take-private; Triton continues to file as a debt issuer. The thesis's operating figures: fleet >7.0M TEU (~13% global share); utilization 97.4% — note this is a Sep-2020 cycle peak, not a current figure, and the thesis's own claims-to-verify list flags it as such. Container leasing is a consolidated ~$10B+ market (Triton, Textainer ~3.5M TEU, Florens ~2.8M TEU). Cited for: The container-lessor endpoint — the neutral operator that owns the box, not the ship or the cargo; the $13.3B Brookfield take-private as the terminal valuation of that role; the ~7.0M TEU / ~13%-share figure; utilization 97.4% (label as cycle peak). (IV.5; IV ARGUMENT §"container value-capture"; V.1) Links:

Textainer Group Holdings Limited

What it is: The #2 intermodal-container lessor — with Triton, the pair whose take-privates by infrastructure capital mark the terminal valuation of a neutral machine-fleet operator. Status: Private. Acquired by Stonepeak (infrastructure PE, ~$57.1B AUM at announcement); deal completed 14 March 2024. Formerly NYSE: TGH. Market cap / valuation: ~$7.4B aggregate enterprise value, announced 22 Oct 2023; $50.00/share in cash to common holders (common equity ~$2.1B), a ~46% premium to the last close before signing. Preference shares and depositary shares were redeemed post-close. Latest financials: No standalone public reporting since March 2024. Thesis operating figure: fleet ~3.5M TEU (vs Triton's >7.0M), in a container-leasing market of ~$10B+ and consolidating. Cited for: The consolidation endpoint — "Textainer acquired by Stonepeak, ~$7.4B EV"; fleet ~3.5M TEU as the second pole of the container-lessor oligopoly. (V.1; IV ARGUMENT §"container value-capture") Links:

AerCap Holdings N.V.

What it is: The world's largest aircraft lessor — the mature-state proof that a mobile, high-value capital good ends up owned by independent lessors rather than by its operators, and the direct analogy for a robot-fleet lessor. Status: Public (NYSE: AER; also Euronext Dublin). Market cap / valuation: ~$23.3B at $147.67/share (2026-07-14, StockAnalysis). Latest financials: TTM revenue $8.68B (+7.8%); TTM net income $3.93B (+83.7%) as of 2026-07-14. FY2025 activity: 705 transactions — 371 lease agreements, 145 asset purchases, 189 asset sales for ~$3.9B of proceeds. Portfolio: ~1,515 aircraft owned + 166 managed (≈1,681), plus 1,200+ engines and 300+ helicopters, and an order book of ~410 aircraft; ~$72B of assets. Cited for: **"AerCap (1,676 aircraft, world's largest), built by *rolling up* ILFC in 2014 and GECAS in 2021" — the roll-up that created the modern lessor; and the leasing-penetration series, ~2.4% of the world fleet leased in 1980 → ~51% in 2023–2025. Note the correction below: the current 20-F count is ~1,681 owned+managed aircraft (1,515 owned), not 1,676. (V.3; V ARGUMENT §3, §5 "Aircraft: ~51% fleet leased") Links:**

Guinness Peat Aviation (GPA) → GE Capital Aviation Services (GECAS)

What it is: The historical cautionary case at the birth of aircraft leasing — the pioneer that created the asset class, over-levered against a fleet ordered on optimistic residual assumptions, and was absorbed by the balance sheet that rescued it. The thesis's structural warning for whoever tries to build the first robot-fleet lessor. Status: Both defunct as independent entities. GPA (founded 1975 at Shannon, by Tony Ryan with Aer Lingus and Guinness Peat Group) collapsed after its 1992 flotation failed; its assets were managed and absorbed by GE Capital Aviation Services (GECAS), formed by GE Capital in 1993. GECAS was itself sold to AerCap in a deal completed 1 November 2021 (~$30B; GE took ~$24B cash plus ~46% of AerCap). GPA's residual entity was renamed AerFi and later absorbed into what became debis AirFinance → AerCap. Market cap / valuation: No live valuation for either. Historical markers: the aborted IPO of ~$850M of shares on 18 June 1992 (applications for only ~50M of 85M shares offered); GE Capital's 1993 rescue — creating GECAS to manage GPA's assets and buying 44 of GPA's 464 aircraft outright for US$1.4B — with an option on ~90% of the equity at a low price. The final chapter: AerCap acquired GECAS from GE for ~$30B (closed Nov 2021). Latest financials: Neither entity reports. Historical: GPA had roughly $5B+ of debt (some accounts put total obligations near $10B including order commitments) and posted a loss of roughly $700M in the fiscal year ending March 1993. Note: the thesis's "~$5B debt" and press accounts of "~$10bn in debts" differ because one counts balance-sheet debt and the other includes the aircraft order book — flag which you mean. Cited for: GPA as world's-largest lessor in the 1980s with 300+ aircraft (its fleet under management reached 464 by the collapse); 18 June 1992 aborted ~$850M flotation; post-Gulf-War downturn; over-levered against an optimistic residual; GECAS rescue; industry consolidation into AerCap (which absorbed ILFC in 2014 and GECAS in 2021). (V.3; V ARGUMENT §"SECONDARY — Aircraft-leasing formation (GPA → ILFC/AerCap)") Links:

Moody's Corporation

What it is: The ratings house — the thesis's valuation comp for "the synthesized-record business," the company that proves underwriting data compounds into tens of billions of enterprise value, and the origin story (John Moody, 1909) of rating machines from public records. Status: Public (NYSE: MCO). Ratings arm: Moody's Investors Service (MIS); data arm: Moody's Analytics (MA). Market cap / valuation: ~$86.1B at $492.90/share (2026-07-14, StockAnalysis). This confirms the thesis's "~$85B, July 2026" and supersedes the older "~$80B." Latest financials: Q1 2026 (record): revenue $2.079B (+8%); company adjusted operating margin 53.2% (from 51.7%); adjusted diluted EPS $4.33 (+13%); free cash flow $844M (+26%). MIS (ratings) segment: record revenue ~$1.2B on >$2 trillion of rated issuance, adjusted operating margin ~66.7% (≈67%). Moody's Analytics ARR $3.6B, adjusted operating margin 33%. FY2025: revenue $7.72B (+8.9%), net income $2.46B. TTM (to 2026-07-14): revenue $7.87B, net income $2.50B. Cited for: (a) Moody's ≈ $85B of market value as the ratings comp for the thesis's platform ("priced like a small Moody's") — II.4/V.5/V ARGUMENT §5 ("~$85B market cap, ~$4.0B EBITDA, July 2026"); (b) MIS operating margin ~60–63% adjusted (60.1% in 2024; 66% in Q1 2025) as the "toll-booth economics" number — II.4, VI.1, VI ARGUMENT §"×~2.3 Moody's MIS"; (c) **April 1909: John Moody's Analyses Publishing Company issues *Moody's Analyses of Railroad Investments*, applying letter ratings to railroad securities — the origin of ratings from organized public records (V ARGUMENT §"DOMINANT"); (d) the investor-pays → issuer-pays flip around 1970 after sixty years on the wrong side of the two-sided market (VI ARGUMENT §1); (e) A3 rating on CoreWeave's DDTL 4.0 (I.1, V.1). Corrections it forces:** the MIS margin has *risen*, not settled at ~60% — Q1 2026 adjusted operating margin was ~67%. Use FY figures, not the quarter, if you want a conservative number. Links:

S&P Global Inc.

What it is: The other half of the ratings duopoly — the thesis's second data point that a scope-owning referee earns ~60%+ operating margins forever, and the agency that rated SolarCity 2013-1 BBB+. Status: Public (NYSE: SPGI). Divisions: S&P Global Ratings, Market Intelligence, S&P Dow Jones Indices, Commodity Insights/Platts, CRISIL. (Mobility was spun off in July 2026.) Market cap / valuation: ~$129.4B at $437.04/share (2026-07-14, StockAnalysis). Latest financials: FY2025 (reported 10 Feb 2026): revenue $15.336B (+8%); adjusted operating profit $7.730B (+11%); company adjusted operating margin 50.4% (+140bp). Ratings division FY2025 revenue $4.724B (+8%). Ratings operating profit margin: 65% for the nine months to 30 Sep 2025 (67% adjusted); 61% in Q4 2025 — which blends to the thesis's full-year ~64% (65% adjusted). TTM (to 2026-07-14): revenue $15.73B, net income $4.78B. Cited for: (a) "S&P Global's Ratings segment posted a 64% operating margin (65% adjusted) for FY2025" — the toll-booth-economics number (VI.1; VI ARGUMENT §1; II.4 "Moody's/S&P run ~60% operating margins"); (b) S&P rated SolarCity's 2013-1 pool BBB+ (V.1); (c) the absence check — no rated robot-fleet securitization appears in S&P's asset-class taxonomy (V ARGUMENT §6). Links:

KBRA (Kroll Bond Rating Agency, LLC)

What it is: The post-2008 challenger NRSRO that dominates structured-finance ratings — in the thesis, the agency whose asset-class taxonomy is the test of whether "robot fleet ABS" exists as a category (it does not). Status: Private. Founded 2010 by Jules Kroll; an SEC-registered NRSRO. Majority ownership passed to Wharf Street LLC, which acquired ~90% of the agency in late 2021. Never publicly listed. Market cap / valuation: Last disclosed marker: the Wharf Street transaction (announced late 2021) valued KBRA at more than $300M for its ~90% stake, per trade reporting. KBRA itself has never published a valuation; treat the >$300M figure as press-reported, not a filed number. Latest financials: Not disclosed; KBRA is private and files no financial statements. Its published output is the ratings and research itself. Cited for: "2026 US ABS new-issue is forecast at ~$385B" — precisely, KBRA's 2026 U.S. ABS Sector Outlook (published Nov 2025) forecasts record total new-issue volume of ~$385.2B, ~+5% y/y; also its surveillance stability data (through Oct 2025: 77% affirmations, 19% upgrades, 4% downgrades, downgrades concentrated in non-prime auto and solar ABS). And, critically, the absence check: robot/humanoid fleets appear in no rating agency's asset-class taxonomy, KBRA's included. (V.1; V ARGUMENT §5, §6) Links:

Morningstar DBRS (DBRS, Inc.)

What it is: The world's fourth-largest credit rating agency and, alongside Moody's, one of the two agencies that put an investment-grade rating on a GPU-backed loan for the first time — the thesis's proof that the ratings apparatus, not just the lenders, has accepted compute as collateral. Status: Subsidiary of Morningstar, Inc. (NASDAQ: MORN). Morningstar agreed to acquire DBRS on 29 May 2019 for $669M; the deal closed 2 July 2019 and DBRS was merged with Morningstar Credit Ratings to form DBRS Morningstar, since rebranded Morningstar DBRS. Market cap / valuation: DBRS has no standalone market cap. Acquisition price $669M (2019). Parent Morningstar, Inc. is publicly traded (NASDAQ: MORN); its credit-ratings business is reported inside the Morningstar Credit segment. Latest financials: Morningstar DBRS does not report separately. Scale at acquisition: >2,400 issuer families, ~50,000 securities rated, 500+ staff across seven locations. Parent Morningstar's Credit segment (which houses DBRS) is the relevant reporting line in Morningstar's 10-K. Cited for: The A(low) rating on CoreWeave's $8.5B DDTL 4.0 facility (alongside Moody's A3) — "the first investment-grade-rated GPU-backed financing." (I.1; I ARGUMENT ledger; V.1) Links:

Equifax Inc.

What it is: The oldest credit bureau — the thesis's DOMINANT historical case in Part II: a private firm that made itself the referee of a record nobody else kept, and compounded that seat for 125 years. Status: Public (NYSE: EFX). Founded 1899 in Atlanta as Retail Credit Company by the Woolford brothers; rebranded Equifax in 1975. Market cap / valuation: ~$20.0B at $167.94/share (2026-07-14, StockAnalysis). Latest financials: FY2025: revenue $6.075B, +7%; diluted EPS $5.32, adjusted EPS $7.65; Q4 2025 revenue $1.551B (+9%). TTM (to 2026-07-14): revenue $6.28B, net income $698.7M. Data coverage: more than 800 million individual consumers and more than 88 million businesses worldwide; ~14,000 employees; operations in 24 countries. Recent: agreement to acquire Círculo de Crédito (Mexico) for $750M. Cited for: Founding Nov 1899 (Woolford brothers, Retail Credit Company); the rebrand to Equifax in 1975; >800 million consumers on file; $6.07B FY2025 revenue; and the three-bureau structure (Equifax / Experian / TransUnion) as a near-textbook oligopoly sitting on a record nobody can replicate — plus the 2017 breach as "the failure mode of an unaccountable referee." (II.3; II ARGUMENT §"DOMINANT — the credit bureaus"; VI) Corrections it forces: the thesis's ">800 million consumers and 100 million businesses" — the businesses figure is more than 88 million, not 100 million. Use 88M or say "tens of millions of businesses." Links:

Experian plc

What it is: The largest of the three consumer credit bureaus by revenue — the second seat in the thesis's "near-textbook oligopoly" that owns the consumer credit record. Status: Public (LSE: EXPN; FTSE 100; Irish-domiciled, US-listed ADRs as EXPGY). Market cap / valuation: ~£23.8B at 2,654 GBX (2026-07-14, StockAnalysis) — roughly $32B. Note: the stock is down ~31% over the trailing year. Latest financials: FY26 (year ended 31 March 2026, reported May 2026): total revenue US$8,425M, +13% at actual rates / +8% organic; Benchmark EPS +15%; post-tax return on capital employed 17.2%; new US$1B share buyback. TTM (to 2026-07-14): revenue £6.40B, net income £1.14B. Cited for: One of the three bureaus — "Equifax, Experian, TransUnion — a near-textbook oligopoly sitting on [a record no one can rebuild]." The thesis uses the trio, not Experian's individual numbers, as the structural claim. (II.3; II ARGUMENT §"DOMINANT — the credit bureaus") Links:

TransUnion

What it is: The third consumer credit bureau — the smallest of the trio and the one that completes the thesis's oligopoly claim. Status: Public (NYSE: TRU). Market cap / valuation: ~$14.6B at $75.90/share (2026-07-14, StockAnalysis). Latest financials: FY2025: revenue $4.58B (+9.4%), net income $455.4M (+60.1%). TTM (to 2026-07-14): revenue $4.73B (+11.0%), net income $704.4M (+91.7%), EPS $3.60. Cited for: The third seat in the bureau oligopoly (Equifax / Experian / TransUnion) — the thesis's proof that the record-keeping seat, once occupied, consolidates into a small number of unassailable incumbents. (II.3; II ARGUMENT §"DOMINANT — the credit bureaus") Links:

PICC (The People's Insurance Company (Group) of China) / PICC Property and Casualty

What it is: China's largest P&C insurer and one of the three carriers writing humanoid-robot policies — the thesis's proof that physical robot insurance is *already writable*, in China, today. Status: Public. PICC Group: HKEX 1339 / SSE 601319. The robot policies are written by its P&C subsidiary PICC Property and Casualty Co Ltd (HKEX: 2328). State-controlled (founded 1949). Market cap / valuation: PICC Group ~HK$336.7B at HK$5.03/share (2026-07-14, StockAnalysis). Latest financials: PICC Group FY2025: revenue RMB 623.40B (+7.2%), net income RMB 46.21B (+9.6%). TTM (to 2026-07-14): revenue HK$701.7B, net income HK$47.9B. Cited for: "Since Sep 2025, PICC P&C, China Pacific P&C and Ping An P&C offer humanoid-robot policies." The documented deal the thesis rests on: the Huazhong University of Science and Technology business incubator (Wuhan) insured two 60kg humanoid robots at a premium of ~5,000 yuan (~$707) per robot per year, with a maximum compensation of 500,000 yuan if damage occurs within the year — the first embodied-robot policy in Hubei province. Cover comprises physical damage (natural disaster, fire/explosion, collision, overturning/falling, electrical failure, cybersecurity incident, abnormal operation) and third-party liability (personal injury or property damage the robot causes). (V.4; V ARGUMENT §"Unlock: physical robot insurance is writable today") Links:

Ping An Insurance (Group) Company of China, Ltd.

What it is: China's largest private insurer/financial conglomerate, and the third of the three carriers writing embodied-robot cover — in the thesis, evidence that robot insurance in China moved from a product to a *financing package*. Status: Public (HKEX: 2318; SSE: 601318). The robot cover is written by Ping An Property & Casualty Insurance Company of China. Market cap / valuation: ~HK$1.01 trillion at HK$53.70/share (2026-07-14, StockAnalysis). Latest financials: TTM (to 2026-07-14): revenue HK$1.08T, net income HK$150.9B; PE ~6.7. FY2025: revenue -0.8%, earnings +6.5%. Q1 2026: operating profit +7.6%; life & health new business value +20.8% y/y. ~258,800 employees. Cited for: One of the three Chinese carriers writing humanoid-robot policies since Sep 2025 — and specifically, in November 2025 Ping An P&C rolled out a comprehensive financial solution integrating robot insurance with credit and IPO services, which is the thesis's strongest data point that insurance is becoming the underwriting spine of robot *finance*, not a standalone product. (V.4; V ARGUMENT §"Unlock: physical robot insurance is writable today") Links:

China Pacific Insurance (Group) Co., Ltd. (CPIC)

What it is: China's #3 insurer and the first carrier to launch a dedicated humanoid-robot commercial insurance product — the earliest dated evidence in the thesis that robot risk is being priced. Status: Public (HKEX: 2601; SSE: 601601). The robot product is written by China Pacific Property Insurance Co Ltd. Market cap / valuation: ~HK$310.2B at HK$27.86/share (2026-07-14, StockAnalysis). Latest financials: TTM (to 2026-07-14): revenue HK$375.8B (+4.9%), net income HK$61.3B (+25.9%), EPS 6.18 (+22.1%). Q1 2026: steady net-profit and operating-profit growth; life value and agency channel improving. Cited for: China Pacific Property Insurance released China's first dedicated insurance for the commercial application of humanoid robots, in September 2025, covering "the whole chain of production, sales, leasing and usage" — the date-stamp for "robot insurance exists in China now." (V.4; V ARGUMENT §"Unlock") Links:

Relm Insurance Ltd.

What it is: A Bermuda-based specialty insurer for emerging-sector risk (crypto, Web3, AI) — in the thesis, the earliest Western carrier to write affirmative AI liability, and the proof that Western AI cover started as a *wrap around an exclusion*, not as a standalone product. Status: Private. Bermuda-domiciled specialty carrier, founded 2019, writing for emerging industries — digital assets/Web3, AI, the space economy, fintech. Acquired/launched Relm Insurance MENA in July 2024. Market cap / valuation: Could not source a valuation or a disclosed funding total to a primary link as of 2026-07-14. Relm has never published a valuation; the only investor traceable in public databases is Plug and Play Insurtech (accelerator round, amount undisclosed). Do not state a Relm valuation in the text. Latest financials: Not disclosed — Relm is private and publishes no financial statements. Cited for: "Relm Insurance 'PONTAAI,' an excess wrap above existing liability programs." Precisely: on 14 January 2025 Relm launched three AI products — NOVAAI, PONTAAI and RESCAAI. PONTAAI is an excess Difference-in-Conditions / Wrap policy designed to fill the exclusions and gaps in a client's existing liability programme created by their use or development of AI; it covers damages, claim expenses and civil fines (where insurable) from negligent acts/errors/omissions in AI services, including IP infringement, discrimination, personal injury, privacy breaches and violations of AI-specific regulation (explicitly framed against the EU AI Act). (V.4; V ARGUMENT §"Western robot/AI insurers, early") Links:

Munich Re (Münchener Rückversicherungs-Gesellschaft AG) / HSB (Hartford Steam Boiler)

What it is: The world's largest reinsurer and the carrier that turned AI performance into an insurable, parametric-style risk — in the thesis, the balance sheet that makes AI/robot underwriting scalable, and the same firm named in Part II as a catalyst for machine-safety standards. Status: Public (XETRA: MUV2; DAX). HSB (Hartford Steam Boiler Inspection and Insurance Company) is Munich Re's US specialty subsidiary — a multi-line specialty insurer and IoT/inspection risk-management provider. Market cap / valuation: ~€64.75B at €512.00/share (2026-07-14, StockAnalysis). Latest financials: FY2025 net result ~€6.1B (+8%). TTM (to 2026-07-14): revenue €61.35B, net income €6.73B, EPS €52.27. Cited for: (a) Munich Re / HSB "AI Liability Insurance" for small business — launched 18 March 2026; fills the AI-related gaps in general-liability policies, covering bodily injury, property damage, and personal & advertising injury from AI-generated content; standard limits $25,000 or $50,000 with a $500 deductible; distributed only through partner carriers (not sold direct), pending regulatory approval. (b) aiSure™, Munich Re's affirmative AI performance cover, is the product Mosaic writes on (below). (c) Munich Re also appears in Part II as a named catalyst in the machine-safety/standards map. (II.4 catalysts map; V.4; V ARGUMENT §"Western robot/AI insurers, early") Links:

Mosaic Insurance

What it is: A specialty (Lloyd's-backed) insurer that, with Munich Re, brought the first parametric-style AI *performance* cover to market — the thesis's clearest example of underwriting written against a measurable telemetry threshold rather than against a narrative. Status: Private. Launched 4 February 2021 by Mitch Blaser, combining Lloyd's Syndicate 1609 with a syndicated capital-management agency structure; anchor investor Golden Gate Capital. Market cap / valuation: No valuation has ever been disclosed. The disclosed capital marker is Mosaic's initial capitalisation of $185M at launch (Feb 2021), anchored by Golden Gate Capital. Could not source a current valuation to a primary link as of 2026-07-14. Latest financials: Not publicly reported (private; syndicate results are disclosed through Lloyd's rather than as company accounts). Cited for: "Mosaic + Munich Re aiSure" — announced February 2026: Mosaic underwrites and distributes Mosaic x aiSure™, providing up to EUR/USD/CAD 15 million of initial capacity to protect AI developers and vendors worldwide against financial loss from defined AI performance failures. It is designed for the probabilistic nature of AI: parametric-style payouts triggered automatically when measurable performance thresholds are breached, settled on performance data rather than lengthy investigation, and underwritten on the *model* — what it does and how its outputs are used — rather than on the insured's industry. Complementary to cyber and tech E&O. This is the single best existing example of the thesis's "underwriting from the telemetry." (V.4; V ARGUMENT §"Western robot/AI insurers, early") Links:

Koop Technologies

What it is: The robotics/AV insurtech that underwrites off a live API feed from robot fleets — the thesis's single sharpest proof that the demand for a standardized, machine-readable operating record already exists and is being met ad hoc, vendor by vendor. Status: Private. Founded 2020, Pittsburgh. A Lloyd's of London coverholder and a Lloyd's Lab alumnus. Market cap / valuation: No valuation disclosed. Total disclosed funding: ~$5.7M — a $2.5M seed (Aug/Sep 2021) led by Ubiquity Ventures, with Bee Partners, Sure Ventures, WestWave Capital and angels; plus a ~$3.2M round (11 July 2023) led by Alley Robotics Ventures with Fusion Fund and existing investors. Latest financials: Not disclosed. Company-stated operating metrics from the 2023 round: book of business grown 3× across 14 robotics use cases in ~12 months, a top-percentile loss ratio across the portfolio, and Lloyd's coverholder status. Cited for: "Koop already pulls hundreds of telemetry fields per second directly from fleets for automated underwriting." Confirmed at source: Koop's own "API Underwriting" page describes evaluating insurable risk by collecting exposure and performance data via API continuously or periodically, and explicitly references accessing "hundreds of data fields sent every second from a fleet of robots", organized around three metrics — performance (troubleshooting/error rates), utilization (time, mileage, jobs completed), and operating domain (warehouse vs public road). (VI.1; VI ARGUMENT §"Koop already pulls hundreds of…") Correction it forces: as of 2026-07-14, koop.ai's homepage presents Koop primarily as an AI-powered GRC / compliance-certification platform (SOC 2, ISO 27001, HIPAA, CMMC 2.0) with insurance as one offering — the robotics-underwriting product is now sub-page material, not the front door. Cite the API-underwriting blog page and the funding releases, not the homepage, and do not describe Koop as a pure robotics insurer without that caveat. Links:

John Deere Financial (John Deere Capital Corporation) / Deere & Company

What it is: The full-stack manufacturer captive that finances the sale, lease and residual of its own machines — the thesis's warning case for how robot financing will *default* to happening if no independent standard exists: the OEM owns the credit, the data and the residual. Status: Public parent (NYSE: DE). John Deere Capital Corporation (JDCC) is a wholly-owned finance subsidiary that files its own 10-K with the SEC (CIK 27673) — a rare, useful primary window into captive-finance mechanics. Market cap / valuation: Deere & Company ~$157.0B at $581.55/share (2026-07-14, StockAnalysis). JDCC has no separate market value; it is wholly owned. Latest financials: Deere FY2025 (ended 2 Nov 2025): net sales and revenues $45.684B, -12% y/y; net income attributable to Deere $5.027B ($18.50/share), vs $7.100B ($25.62) in FY2024. TTM (to 2026-07-14): revenue $47.34B, net income $4.78B. JDCC disclosure: at least 90% of the receivables and leases JDCC administers finance the purchase or lease of John Deere products — i.e. the captive exists to move the parent's machines. Cited for: "John Deere Financial — Deere runs a large captive that finances retail sales and leases of its own equipment", used in V ARGUMENT §3 as the closest live analogue for financing a capital-good fleet (and named again at V.5/OUTLINE V as "the John Deere Financial / auto-captive path" — the path robot finance takes *before* an independent standard exists). (V.3; V ARGUMENT §3; V OUTLINE ~line 511) Links:

GM Financial (General Motors Financial Company, Inc.) / General Motors

What it is: The archetypal captive finance arm — from GMAC (1919), "the model for the modern captive," through GM's post-bankruptcy rebuild of a captive by acquisition. The thesis's historical spine for "the manufacturer finances its own machines until someone independent does it better." Status: Public parent (NYSE: GM). GM Financial is a wholly-owned subsidiary that files separately with the SEC (CIK 804269) — formerly AmeriCredit Corp. Market cap / valuation: General Motors ~$68.7B at $76.20/share (2026-07-14, StockAnalysis). GM Financial has no separate market value. The relevant historical price: GM agreed to acquire AmeriCredit for ~$3.5B in an all-cash deal, announced 22 July 2010, closing in October 2010, and renamed it GM Financial. Latest financials: GM Financial FY2025: net sales and revenue $17.048B; GM Financial interest, operating and other expenses $14.295B (year ended 31 Dec 2025). Parent GM FY2025: ~$10.5B of EBIT-adjusted earnings; TTM (to 2026-07-14) revenue $184.6B, net income $2.43B. Cited for: GMAC (1919) "became the model for the modern captive finance company"; GM bought AmeriCredit for ~$3.5B in Oct 2010 and renamed it GM Financial; by Q2 2020 GM Financial was financing 53% of GM's US retail sales. (V.3; V ARGUMENT §3) Correction it forces: the 53% figure is a Q2-2020 pandemic-quarter peak, not a steady state — GM's own 10-K reports GM Financial penetration of US retail sales at 44% (2021) and 43% (2022). Label 53% as a peak or use the 40s. Links:

The Progressive Corporation

What it is: The telematics-underwriting precedent — the carrier that proved a machine's own data feed can reprice its risk, in both directions, at national scale. In the thesis, the single strongest existing analogue for underwriting a robot off its telemetry. Status: Public (NYSE: PGR). Founded 1937. Market cap / valuation: ~$132.7B at $227.12/share (2026-07-14, StockAnalysis). Latest financials: FY2025: revenue $87.64B (+16.3%), earnings +33.6%. TTM (to 2026-07-14): revenue $89.42B (+13.9%), net income $11.56B (+32.6%). ~70,000 employees; segments: personal auto, commercial auto, residential property. Cited for: (a) Snapshot, Progressive's usage-based auto programme: Progressive advertises average savings of $231/year, with an average sign-up discount of $169 and an average $322/year at program completion; and, crucially for the thesis's "the data actually reprices risk, both ways," **~2 in 10 (one in five) drivers see a rate *increase* from Snapshot. (b) The fleet extension — Smart Haul, Progressive Commercial's telematics programme, which pays fleets to share ELD/telematics data. (V.4; V ARGUMENT §"Unlock: … underwritten against"; V OUTLINE lines 341–347) Corrections it forces (important): the thesis says "Progressive gives a 5–18% discount for sharing Samsara telematics data." That conflates two different things. The 5%-initial-to-18% range is Progressive's Smart Haul program, and the largest savings are reserved for preferred-vendor ELDs (Geotab, Motive, Omnitracs) — with an established safety record some fleets save 15%+. Samsara's own Progressive integration is documented as a 3% premium discount for sharing Samsara data. Rewrite as: "Progressive's Smart Haul pays fleets 5%→up to 18% for sharing ELD telematics; Samsara's direct Progressive data-sharing integration carries a 3% discount." The structural point — a telemetry vendor's feed becomes the underwriting input — survives intact; the number does not. Links:**

Framework Ventures

What it is: The crypto-native VC that has explicitly repositioned tokenization and stablecoins as the *financing layer* for AI compute and robotics infrastructure — the thesis's clearest signal that capital-stack builders are forming around robot fleets before the rated, syndicated market exists. Status: Private venture firm (founded 2019 by Michael Anderson and Vance Spencer). Market cap / valuation: Not a valuation but the relevant figure: fourth fund "FVIV" closed oversubscribed at $400M, ~26 June 2026. LPs undisclosed, described as "a predominantly institutional base anchored by an Ivy League endowment, nonprofits, sovereign wealth funds, and funds of funds." Roughly half of the $400M was already deployed at close. Check sizes run $1M to $50M, pre-seed through Series A. Latest financials: Not disclosed (private fund; no reported financials). Firm-stated portfolio positions include Hyperliquid, Sky and Plasma on the digital-asset side, and distributed-energy network Daylight on the infrastructure side. Cited for: "Framework Ventures closed a $400M fourth fund (FVIV, ~26 Jun 2026), explicitly positioning tokenization/stablecoins as the financing layer for AI and robotics infrastructure; ~half already deployed." In V ARGUMENT §6 it is the "crypto/tokenization plumbing" half of the absence-verification: the capital is forming, but *not* as a rated, syndicated robot ABS. (VI.2; V ARGUMENT §6) Links:

Robonnement

What it is: Europe's open Robotics-as-a-Service platform, and — as of the thesis's searches — the closest thing that exists anywhere to a robot-collateralized financing: a small, private, unrated asset-based facility against industrial robot arms. Status: Private. Swiss (Rheintal); operates an open RaaS platform for robot integrators across Switzerland, Germany and Austria. Market cap / valuation: No equity valuation disclosed. The disclosed figure is the financing itself: €15M asset-based financing, closed December 2024, to expand its robot fleet and open its RaaS platform to European integrators. Latest financials: Not disclosed. Company-stated operating scale: 80+ robots deployed across Switzerland, Germany and Austria, with customers including Porsche, BMW and Schneider Electric; subscription RaaS model bundling maintenance, updates and managed operation. Cited for: The absence proof in V ARGUMENT §6 — "there is no rated robot-fleet securitization anywhere; the closest thing that exists is Robonnement's €15M/$15M private asset-based RaaS facility (Dec 2024, industrial arms in Europe, not humanoids, unrated)." Every qualifier matters: private, small, arms not humanoids, and unrated. (V ARGUMENT §6; V OUTLINE claims-to-verify) Links: