Reference Index — Sector 06: Operators, Deployment, Workforce, Fleet Tooling

Entries compiled 2026-07-14. Every figure carries an as-of date and a working link.

GXO Logistics

What it is: The world's largest pure-play contract-logistics provider and, in the thesis, the only deployer today running multiple humanoid brands side by side in live warehouses — the closest existing thing to a neutral multi-brand robot operator. Status: Public (NYSE: GXO). Spun off from XPO in August 2021. Market cap / valuation: $5.67B as of 2026-07-14 (share price $49.28). Latest financials: Q1 2026 (quarter ended 2026-03-31): revenue $3.3B, +10.8% y/y (+4.1% organic); net income $5M vs. a $95M loss in Q1 2025; adjusted EBITDA $200M vs. $163M; adjusted diluted EPS $0.50. Record sales pipeline of $2.7B. FY2026 guidance raised to adjusted EBITDA $935–975M and adjusted diluted EPS $2.90–3.20. TTM revenue $13.50B. Cited for: GXO runs a three-brand "operational incubator" — Agility Digit, Apptronik Apollo, Reflex Robotics — concurrently in live warehouses; first logistics provider to deploy humanoids in a live operating facility; the Flowery Branch, Georgia site's pilot→RaaS→100,000-tote progression with Agility's Digit; "customer demand for a cross-brand operator is revealed today (GXO)". VI.1; IV.1; IV.2; ARGUMENT IV §"the operator already exists"; VI ARGUMENT §163–165. Links:

Formic (Formic Technologies)

What it is: The pure-play Robots-as-a-Service operator in fixed industrial automation — the working template for the thesis's operator economics (no capex to the customer, the operator owns the asset and sells productive hours against an uptime guarantee), and now extending toward humanoids. Status: Private. Chicago-based; founded 2020. Investors include Blackhorn Ventures, Lux Capital, Initialized Capital, Translink Capital, Mitsubishi HC Capital. Market cap / valuation: No valuation disclosed by the company. Total funding reported at ~$104M by Crunchbase (accessed 2026-07-14); the last announced priced round was a $27.4M Series A extension led by Blackhorn Ventures (June 25, 2024), taking the Series A to more than $52M since January 2022. Latest financials: Private, no P&L disclosed. Company-published operating metrics (formic.co, accessed 2026-07-14): 99.30% fleet uptime over the trailing 12 months; 9.5M+ items produced weekly across the fleet; 645,520 hours of training data; average deployment lead time 12 weeks; 893 worker injuries prevented year-to-date. Pricing model stated as "$0 CapEx required," one fixed monthly subscription price, "20% OpEx reduction," 24/7 support and maintenance included. Cited for: "Formic runs the RaaS operator model in fixed automation: $0 upfront capex, customer pays only for productive running hours … humanoids slated for its production line in 2026–2027" (VI.1); "Formic already brokers robot access as a fully-managed per-hour service" (IV.2); Formic's uptime-guaranteed RaaS contract as the closest working analogue to a robot SLA (IV.5); "Formic brokers conventional automation, not humanoids across brands" (ARGUMENT IV, open positions). CAUTION — thesis numbers that do not match the primary source: Formic's own site states a 20% OpEx reduction, not "~42% below current opex," and publishes no "100+ integrator partners" or "5,000+ deployments" figures. Earlier Formic/Robotics-247 coverage cited 60+ facilities and $20M+ total contract value with a 97% renewal rate. Those thesis figures need re-sourcing or correction. Links:

Jabil

What it is: The contract manufacturer to which Apptronik handed Apollo production — the thesis's evidence that even the best-capitalized humanoid OEM outsources the hardware and keeps the deployment layer, because manufacturing is the commodity and deployment is not. Status: Public (NYSE: JBL). One of the world's largest electronics manufacturing services (EMS) providers; ~100 sites, HQ St. Petersburg, Florida. Market cap / valuation: $33.85B as of 2026-07-14 (share price $323.03). Latest financials: Fiscal Q3 2026 revenue $8.8B; core EPS $3.16 (beat consensus $3.12). TTM revenue $33.59B (+17.8% y/y). FY2026 guidance raised to ~$35B revenue and core EPS ~$12.70, with AI-related revenue guided to ~$13.6B for FY26. Cited for: "Apptronik handed Apollo production to Jabil rather than build the line itself" (ARGUMENT IV §79, §533; OUTLINE IV.1 "Jabil contract manufacturing"). The Apptronik–Jabil strategic collaboration (announced Feb 2025) covers building, testing and deploying Apollo at Jabil's St. Petersburg, FL operation, with Apollo units also working on Jabil production lines — "Apollo to build Apollo." Links:

Palantir Technologies

What it is: The named precedent for the thesis's core claim that forward-deployed services can compound into product margins rather than staying a body-shop — the counterexample to the staffing agencies, and the only one with a public P&L proving it. Status: Public (NASDAQ: PLTR). Market cap / valuation: $321.98B as of 2026-07-14 (share price $134.31; P/E ~146). Latest financials: Q1 2026 (quarter ended 2026-03-31): revenue $1.63B, +85% y/y; ARR above $6.5B. FY2025: revenue $4.48B (+56.2% y/y), net income $1.63B. TTM revenue $5.22B (+67.7%); TTM net income $2.28B. Headcount 4,429. Cited for: "Palantir hit ~$1.5M annualized revenue per employee" (Q1 2026: $1.63B × 4 ÷ 4,429 ≈ $1.47M) and "~$1.01M FY2025" ($4.48B ÷ 4,429) — ARGUMENT IV §234–250, §428, §550; OUTLINE IV.2 §142, §181. Used as the multiple against a staffing major's revenue-per-head, and to support "compounds like Palantir, not like Randstad." Also cited for the forward-deployed engineering (FDE/FDSE) model itself as the deployment motion the robot operator should copy (OUTLINE IV.2 §118–129). CAUTION: OUTLINE IV.2 §181 states "headcount ~3,100." Palantir's current headcount is 4,429 (2026). The ~3,100 figure is stale and, if used with Q1 2026 revenue, would overstate revenue/employee to ~$2.1M. The ~$1.5M annualized figure is right only at ~4,430 heads. Links:

Randstad

What it is: The world's largest staffing firm and the thesis's "floor" comp — the century-old proof that a pure labour-brokerage business, however large, never earns product margins; the thing a robot operator must not become. Status: Public (Euronext Amsterdam: RAND). HQ Diemen, Netherlands. Market cap / valuation: €5.26B as of 2026-07-14 (share price €30.00; down ~28% over the trailing year; 52-week range €21.30–€44.34). Latest financials: FY2025: revenue €23.1B (−2% y/y), EBITA €720M, underlying EBITA margin 3.1%; proposed dividend €1.62/share; >€100M net structural savings delivered. Q4 2025: revenue €5.8B (−2.1% organic), EBITA €191M, underlying EBITA margin 3.3%. Net income FY2025 ~€291M. Corporate headcount ~38,000. Digital marketplaces ~€4B annualized revenue; 1.4M self-scheduled shifts in Q4 2025 alone. Cited for: "Randstad posted a Q4 2025 EBITA margin of 3.3% on revenue around €23–27B" (ARGUMENT IV §209, §547); "Randstad FY2025 revenue €23.07B" (OUTLINE IV.2 §151, §186); the majors averaging ~20% gross / ~3% EBIT margin over the past decade (OUTLINE IV.2 §150); "Randstad's ~€23–27B over a workforce in the tens of thousands" as the revenue-per-head denominator against Palantir (ARGUMENT IV §428–429); "compounds like Palantir, not like Randstad" (ARGUMENT IV §250). Note: The thesis's "€23.07B" is essentially right (€23.1B reported). The "revenue around €23–27B" range is loose — FY2025 revenue was €23.1B; the €27B figure appears to be a stale prior-year number. Randstad's own reported corporate headcount is ~38,000, giving ~€600k revenue per corporate employee — a clean ~2.5x gap to Palantir's ~$1.5M, worth stating precisely rather than as "tens of thousands." Links:

The Adecco Group

What it is: The second staffing major used, with Randstad and ManpowerGroup, to establish the structural margin floor of the labour-brokerage model — ~20% gross, low-single-digit operating. Status: Public (SIX Swiss Exchange: ADEN). HQ Zurich. Brands: Adecco, Akkodis, LHH. Market cap / valuation: CHF 3.19B as of 2026-07-14 (share price CHF 17.79; 173.3M shares outstanding; −29.5% over the trailing year; 52-week range CHF 14.54–27.26). TTM revenue CHF 21.43B; TTM net income CHF 281M; P/E ~11. Latest financials: FY2025: revenues +1.3% y/y (Adecco +2.5%, Akkodis −4%, LHH flat); gross margin 19.2%; EBITA margin excluding one-offs 3.0%; operating income €572M (+8%); net income €295M (+2%); 245 bps of market-share gain. Q4 2025: revenue +4.9% y/y (Americas +21%, APAC +7%); gross margin 19.1% (stable organically); EBITA margin ex-one-offs 3.8% (+60 bps y/y). Cited for: "Adecco Q4-2025 gross margin 19.1%, low-single-digit operating margin" (OUTLINE IV.2 §151; ARGUMENT IV §209–211, §547); the decade-average "~20% gross margin and ~3% EBIT margin" across Adecco/Randstad/Manpower (OUTLINE IV.2 §150). Both the 19.1% Q4 gross margin and the low-single-digit operating margin (3.8% Q4 / 3.0% FY EBITA ex-one-offs) check out against the company's own release. Links:

ManpowerGroup

What it is: The third of the three staffing majors the thesis uses to fix the labour-brokerage margin floor — and, at a $1.83B market cap on $18.4B of revenue, the most vivid illustration of how little the market pays for a pure body-shop. Status: Public (NYSE: MAN). HQ Milwaukee. Brands: Manpower, Experis, Talent Solutions. Market cap / valuation: $1.83B as of 2026-07-14 (share price $39.28; 52-week range $25.15–$47.34). Market cap is ~0.10x revenue. Latest financials: Q1 2026 (quarter ended 2026-03-31): revenue $4,510.4M, +10.3% reported / +2.9% constant currency; gross profit $723.0M; gross margin 16.0%, down from 17.1% y/y; operating profit $28.3M, operating margin 0.6%; GAAP EPS $0.05, adjusted EPS $0.51. TTM revenue $18.38B (+4.8%); TTM net income −$16.4M (loss). Cited for: "Across the past decade the majors (Adecco, Randstad, Manpower) averaged ~20% gross margin and ~3% EBIT margin" (OUTLINE IV.2 §150) — the structural floor of the staffing model that the thesis argues a robot operator must escape. CAUTION: Manpower's Q1 2026 gross margin is 16.0%, not ~20%, and operating margin is 0.6%, not ~3% — the decade-average claim is directionally right but the current numbers are materially worse. If the thesis states "~20% gross / ~3% EBIT" as a present-tense fact rather than a decade average, it is wrong for Manpower today. Links:

Amazon Delivery Service Partner (DSP) program

What it is: Amazon's owner-operator delivery franchise — the structural chassis the thesis borrows wholesale for the robot operator: low-capital entry, leased assets, an investment-grade counterparty guaranteeing demand, and a published scorecard that gates work allocation. Status: A program of Amazon.com, Inc. (NASDAQ: AMZN), not a separate entity. Launched June 2018. Operates in 19 countries. Market cap / valuation: Not separately capitalised. Parent Amazon.com, Inc. is a multi-trillion-dollar public company; the DSP program itself has no standalone valuation. What Amazon does publish: the network has generated $58B in cumulative revenue for DSP owners over five years. Latest financials (program-level, Amazon-published): ~4,400–4,500 DSP small-business owners globally; 390,000 driving jobs created; $58B in revenue generated for those businesses; 20M+ packages delivered daily; Amazon investment in the program of $12.3B+ since 2018 (a later Amazon release puts total support at $16.7B over seven years), including a $2.1B tranche with $660M of annualized rate increases and bonuses; driver national-average pay pushed toward ~$22–23/hour; 10,000+ Rivian electric vans deployed; in-van cameras credited with a 48% accident-rate reduction since 2020. Unit economics (Amazon's own published projections): startup capital from ~$10,000; fleet of 20–40 vans; projected annual revenue $1M–$4.5M and annual profit $75k–$300k per DSP (Amazon disclaims these are projections, "not based on actual results"). Vans are leased, not bought, through Amazon-negotiated fleet programs (Element Fleet Management is the named fleet lessor/manager). Route allocation is gated on Amazon's weekly scorecard with the tiers Fantastic / Great / Fair / Poor. Secondary market: DSP books of business trade at ~40% of annual revenue (range 25–55%) or 2–3.5x annual net income — versus 60–85% of revenue / 4–5x for FedEx Ground routes, a gap the thesis should note as the price of counterparty control. Cited for: "$10–30k operator entry, 20–40 vans, Element Fleet leasing, scorecard-gated route allocation — the franchise chassis, with an investment-grade counterparty guaranteeing demand" (OUTLINE IV.2 §167–168; OUTLINE V.3 §210–229; VI.3 §59); "a DSP runs 20–40 vans / 10–40 drivers, grosses [$1M–$4.5M]" (OUTLINE V.3 §225); the scorecard as "route-allocation currency" (ARGUMENT IV §505); the litigated failures the thesis designs out (Fli-Lo Falcon class action on behalf of ~2,500 US DSPs; Battle Tested Strategies termination Apr 14 2023 and the NLRB joint-employer finding; the $44.6M verdict; scorecard threshold creep) — OUTLINE V.3 §234–252. Note on the "$10–30k" figure: Amazon publishes "startup costs begin at around $10,000." The upper bound of $30k is a third-party estimate (Route Consultant and similar), not an Amazon number. Links:

Element Fleet Management

What it is: The largest publicly traded fleet-management company in the world and the fleet lessor behind Amazon's DSP vans — the thesis's proof that the operating layer of a vehicle fleet is a separate, profitable, third-party business that OEMs ceded. Status: Public (TSX: EFN; also OTC: ELEEF). HQ Toronto. Market cap / valuation: C$11.87B as of 2026-07-14 (share price C$29.62; 52-week range C$25.32–38.26; −15.7% over the trailing year). Latest financials: Q1 2026 (quarter ended 2026-03-31): record net revenue $324M, +17% y/y; net financing revenue $138M, +23%; adjusted EPS $0.35 (+24%); adjusted free cash flow per share $0.45 (+25%); operating margin 56.2% (+150 bps y/y); ROE 20.3% (+360 bps). Vehicles under management 1.565 million, +3% y/y; client revenue retention 98%. FY2026 guidance: net revenue $1.28–1.305B (+8–10%), adjusted EPS $1.40–1.45. Cited for: Amazon DSPs lease their vans "through Element Fleet" (OUTLINE IV.2 §168; OUTLINE V.3 §222; VI.3 §59); Element as one of the neutral third parties that rolled up the fleet-operating layer the automakers gave away — "rolled up PHH Arval in 2014 and GE Capital Fleet Services in 2015" (ARGUMENT IV §106–107, §536). The 56.2% operating margin and 20.3% ROE are the numbers that make the "the operator, not the OEM, earns the durable margin" argument concrete. Links:

Holman (formerly ARI / Automotive Rentals Inc.)

What it is: The century-old, family-owned fleet-management house the thesis names as the founding proof that the operating layer of a vehicle fleet became a separate multi-billion-dollar industry owned by neutral third parties, not by the automakers. Status: Private. Fourth-generation family-owned; HQ Mount Laurel, New Jersey. Founded 1924 as the Rice & Holman Ford dealership; its fleet division was established in 1948 as Automotive Rentals Inc. (ARI). ARI was folded into the unified Holman brand in a 2023 rebrand — "ARI" no longer exists as a separate operating name. Market cap / valuation: No disclosed valuation; never raised institutional venture/PE capital. Third-party trackers place revenue in the multi-billion range (Owler: "revenue range of up to $10 billion"); Holman itself does not publish financials. Could not source a company-confirmed valuation or revenue figure to a primary link as of 2026-07-14 — Holman publishes no financial statements. What is company-confirmed: it operates 7 business units, 59 dealerships across nine states, and describes itself as the largest privately owned fleet leasing and management company in the US. Latest financials: Not disclosed (private, no filings). Fleet scale: Holman states it manages fleets across North America, the UK, Germany and Canada; secondary sources put vehicles under management at ~2.4M globally across 40+ countries (via the Holman Global Fleet Industries partner network). The thesis's "1M+ vehicles managed" for ARI is a historical, pre-rebrand figure. Cited for: "a separate multi-billion-dollar industry owned by neutral third parties — ARI (founded 1924, 1M+ vehicles managed), Element Fleet …, LeasePlan …, Holman, Wheels" (ARGUMENT IV §105–108, §536). CAUTION for the thesis: ARI and Holman are the same company, listed separately in ARGUMENT IV §108 as if they were two of five distinct players. That is an error. Also, "founded 1924" is the Holman dealership; the *fleet* business (ARI) dates to 1948. Links:

Ayvens (ex-LeasePlan / ALD Automotive)

What it is: Europe's largest full-service vehicle leasing and fleet-management group — the third leg of the thesis's "the OEMs ceded the operating layer" argument, and the largest single pool of managed vehicles in the world outside the OEMs. Status: Public (Euronext Paris: AYV). Majority-owned by Société Générale (~52.6%). Created by ALD Automotive's acquisition of LeasePlan, completed May 2023; the combined group rebranded to Ayvens in May 2024. HQ Puteaux, France; ~13,000 employees. Market cap / valuation: €9.13B as of 2026-07-14 (share price €11.66; P/E ~9.9; dividend yield ~5.1%). Latest financials: FY2025: revenue €25.13B (−0.5% y/y); net income €922.1M (+51.1% y/y). TTM revenue €25.04B; TTM net income €969.3M. Total fleet 3.2 million vehicles as at 2025-12-31. FY2024 profit before tax was €994.3M (vs. €1.218B in 2023). Next results due 2026-07-30. Operating under the "PowerUP 2026" strategic plan; system migrations complete in 14 of 21 overlapping countries as of late 2025. Cited for: "LeasePlan (1.5M vehicles across 32 countries before its 2022 ALD merger)" (ARGUMENT IV §107–108, §537) — the fleet-operating layer as a neutral third-party industry. CAUTION for the thesis: ARGUMENT IV dates the ALD/LeasePlan merger to 2022; the deal was announced in January 2022 but completed in May 2023, and the Ayvens rebrand followed in May 2024. The LeasePlan "1.5M vehicles / 32 countries" figure is the pre-merger standalone number; the combined group now manages 3.2M vehicles, which is the stronger figure for the argument. Links:

Foxglove

What it is: The data and observability platform for physical AI, and the author of MCAP — the thesis's best-capitalized claimant on the robot record substrate, and its central "ship the format, win adoption, let committees canonize" precedent. Status: Private. Founded 2021 by alumni of Cruise, Aurora, Amazon Robotics, Stripe and Coinbase. Investors: Bessemer Venture Partners (lead, Series B), Eclipse, Amplify Partners. Market cap / valuation: No post-money valuation disclosed. $40M Series B led by Bessemer Venture Partners, announced 2025-11-12; total raised more than $58M since founding (the thesis's "$58.75M total" is consistent with company statements). Latest financials: Private, no revenue disclosed. Company states "hundreds of customers" across automotive, aerospace, defense, logistics, agriculture, construction and consumer robotics; named customers include NVIDIA, Amazon, Anduril, Wayve and Dexterity. Cited for: "Foxglove raised a $40M Series B led by Bessemer (Nov 12 2025; $58.75M total) for a multimodal data/observability platform for physical AI" (VI.2 §43; ARGUMENT VI §439); MCAP as the "ship the format, let committees canonize" precedent — MCAP launched 2022 as an open self-describing container format and is now shipped by default with ROS 2 and NVIDIA Isaac (OUTLINE IV.5 §435–439; VI.2 §42; ARGUMENT IV §358); the "MCAP/Foxglove gap" in the record/trust layer (Part V, 07-v3 Layer 2). Links:

Formant

What it is: A cloud robot-operations platform (fleet management, teleoperation, mission tasking, incident management, telemetry ingestion) — one of the two named vendors closest to owning the operational-telemetry layer a neutral operator would run. Status: Private. San Francisco. Investors: SignalFire (led the $6M Seed and $18M Series A), BMW i Ventures (led the most recent $21M round). Market cap / valuation: No valuation disclosed. Total funding reported at ~$39.1M (Crunchbase/startup trackers, accessed 2026-07-14), across a $6M Seed, an $18M Series A, and a $21M round led by BMW i Ventures. Reporting on the round size is inconsistent — some sources describe a $45M Series B (Oct 2023) that appears to double-count earlier rounds. Could not source a company-confirmed post-money valuation to a primary link as of 2026-07-14. Latest financials: Private, no revenue disclosed. Company states it powers "tens of thousands of production robots globally" and reported a >5x year-over-year revenue increase (SignalFire, investor blog). Cited for: "Formant (teleop + fleet management + data ingestion) and InOrbit … the fleet-ops tooling vendors closest to owning the operational-telemetry layer" (OUTLINE IV.5 §440–443); "Formant and InOrbit are the direct competitors on fleet telemetry ingestion and cross-brand data portals" (VI.2 §44); "Partial — Foxglove/Formant/InOrbit, none coupling to underwriting fields" (ARGUMENT VI §328). Links:

InOrbit.AI

What it is: The robot-orchestration platform that actually runs heterogeneous, multi-brand fleets — the software an independent neutral operator would use, and the one vendor of the three that has bet on interoperability rather than lock-in by shipping an open-source Open-RMF fleet adapter. Status: Private. Mountain View, CA. Investors: L'ATTITUDE Ventures and Globant Ventures (co-leads, Series A). Market cap / valuation: No post-money valuation disclosed. $10M Series A, announced 2025-09-30, co-led by L'ATTITUDE Ventures and Globant Ventures. Latest financials: Private, no revenue disclosed. Operating scale (company-published, 2026): the InOrbit Connect ecosystem spans 30+ robot brands; supports ROS 2, AMRs, manipulators and custom robots (incl. Universal Robots, FANUC, KUKA). At Automate 2026 (Chicago, June 22–25) InOrbit ran a live federated orchestration demo across eight robotics companies — Ati Robotics, InOrbit.AI, Kärcher, Neura Robotics, Omron, Peer Robotics, Quasi Robotics and Unitree. Cited for: "InOrbit provides genuine multi-brand fleet operations (ROS2, AMRs, manipulators, custom robots; UR/Fanuc/KUKA) — the software an independent operator would run across a heterogeneous fleet" (VI.1 §33); "InOrbit ships an open-source Open-RMF fleet adapter, positioning toward interoperability rather than lock-in" (VI.2 §44); InOrbit as a fleet-ops tooling vendor close to owning the operational-telemetry layer (OUTLINE IV.5 §440–443). Links:

Intrinsic (Alphabet) — and Open-RMF

What it is: Alphabet's industrial-robotics software company, and the corporate owner of the commercial arm behind Open-RMF — the open middleware that manages heterogeneous multi-fleet traffic, task allocation and shared-infrastructure deconfliction (doors, lifts), i.e. the interoperability layer a robot record standard would sit alongside. Status: Subsidiary of Alphabet Inc. (NASDAQ: GOOGL/GOOG); founded out of X, the moonshot factory, and made an independent Alphabet company in 2021. In December 2022 Intrinsic acquired Open Source Robotics Corporation (OSRC) and OSRC-Singapore (OSRC-SG) — the entity that had led Open-RMF. Market cap / valuation: Intrinsic is not separately capitalised or valued; it sits inside Alphabet's "Other Bets" segment. Parent Alphabet is a multi-trillion-dollar public company. Alphabet has never disclosed a standalone Intrinsic valuation, revenue, or the OSRC purchase price. Could not source a standalone Intrinsic valuation to a primary link as of 2026-07-14; the correct statement is that Intrinsic is an unpriced Alphabet Other Bet. Latest financials: Not separately reported. Alphabet's Other Bets segment is reported in aggregate in Alphabet's 10-K/10-Q and has run at an annual operating loss; Intrinsic's own P&L is not broken out. Important scope limit: Intrinsic acquired the *commercial* arm (OSRC/OSRC-SG). ROS, Gazebo and the Open-RMF intellectual property remained with the Open Source Robotics Foundation (OSRF) — Intrinsic did not acquire ROS, Gazebo, TurtleBot, ROSCon, the technical roadmaps or the build infrastructure. Governance of ROS/Open-RMF now sits with the Open Source Robotics Alliance (OSRA). Cited for: "Open-RMF (OSRF/Intrinsic) manages heterogeneous multi-fleet traffic, task allocation and shared-infrastructure deconfliction — the open middleware layer a record standard would sit alongside" (VI.2 §46); Open-RMF as one of the open interoperability efforts alongside MassRobotics AMR Interoperability (OUTLINE IV.5 §432). Links:

Samsara

What it is: The connected-operations telematics vendor that turned fleet telemetry into a data asset insurers price against — the thesis's scale precedent that a neutral telemetry layer can become the underwriting substrate, and the closest working model for what a robot-fleet record layer would be worth. Status: Public (NYSE: IOT). Fiscal year ends 31 January. Market cap / valuation: $21.85B as of 2026-07-14 (share price $37.49). Latest financials: FY2026 (year ended 2026-01-31): revenue $1,618.6M, +29.6% y/y (from $1,249.2M in FY2025); net loss narrowed to $9.1M from $154.9M — effectively at break-even. Ending ARR $1.890B, +30% y/y. Q4 FY2026 revenue $444.3M, +28% y/y. 12,000+ Core Customers at ≥$25k ARR; 3,194 customers at >$100k ARR, contributing $1.2B of ARR (+37% y/y). Data scale: more than 25 trillion data points ingested annually (10-K, FY2026). Cited for: "Samsara (FY2025 revenue $1.249B, 14T data points, customers cut insurance …)" — ARGUMENT V §613, §681; OUTLINE V.4 §345–354: "customers report a 5–18% discount for sharing Samsara telematics data," and Samsara as the telemetry-vendor scale precedent (OUTLINE V.4 §348, §381, §389). CORRECTIONS the thesis needs: (1) The FY2025 $1.249B / 14T data points figures are now a full fiscal year stale. The current numbers are FY2026 revenue $1.618B and 25T+ data points annually, with $1.89B ARR. (2) The "5–18% discount" band could not be sourced to a single Samsara statement. What is verifiable from Samsara's own insurance-partner marketplace: Northland Insurance offers up to a 10% premium discount (or up to $300/vehicle/year subsidy); Nirvana Insurance offers up to 20%; an RLI case study shows a 5% premium reduction; Samsara states customers referred by an insurance partner receive a 15–25% product discount (not a premium discount). The honest range for *insurance-premium* reduction from sharing Samsara data is roughly 5–20%, and it is partner-specific, not a Samsara-published single figure. Links:

Eastworld Labs

What it is: The humanoid-robot accelerator that states the teleoperation wage-arbitrage trade openly — remote human operators in low-wage countries driving humanoid robots doing G7 work, with the resulting sensor data used to bootstrap autonomy. The thesis's proof that the "teleop arbitrage" it describes is not hypothetical: someone put it in a press release. Status: Private. Launched 2026-02-23 by Virtuals Robotics, the embodied-AI division of Virtuals Protocol (a crypto/agent protocol). Based in Kuala Lumpur, Malaysia. Market cap / valuation: No equity valuation disclosed — Eastworld Labs is a division/accelerator of Virtuals Protocol, funded "onchain" rather than through a priced venture round, and has never disclosed a raise or a valuation. Could not source an Eastworld Labs valuation or funding figure to a primary link as of 2026-07-14; what is disclosed is the operating stack below. Latest financials: No revenue or financials disclosed. Company-published operating scale (Feb 2026 launch release): a fleet of 30+ full-sized humanoid robots, including Unitree G1 platforms; a high-throughput data pipeline (SeeSaw + teleoperation) producing 500,000+ recorded tasks to date, described as one of the world's most diversified egocentric humanoid-training datasets; an integrated stack of robot fleets, datasets, autonomous agents and onchain funding for portfolio founders. Cited for: "Eastworld Labs / Physical Labor Layer — the arbitrage stated openly. Feb 2026 …" and "at least one venture (Eastworld) has stated the arbitrage as its explicit [strategy]" (OUTLINE IV.3 §211, §228, §273); "the platform teleop-arbitrage operators (Eastworld) buy in bulk" (OUTLINE IV.1 §77); "Eastworlds KL (30+ humanoids)" as a humanoid-facility precedent (VI §101). Verification of the thesis's "explicit wage-arbitrage into G7" characterization — CONFIRMED by the company's own words: the launch release states Eastworld "will start by using humanoid robots and low-latency teleoperation to capture wage arbitrage in G7 economies while building the foundation of full agent autonomy," with teleoperation from lower-wage countries as a bridge to autonomy over 12–18 months. OUTLINE IV.3 §273 flags this as needing confirmation "before quoting it as their words" — it is quotable. Links:

BMW Group

What it is: The flagship automotive humanoid deployment — Figure AI's robots working inside BMW's Spartanburg plant — and, in the thesis, one of the two OEM customers proving that "where the paying deployments actually are" is heavy manufacturing, not the home. Status: Public (XETRA: BMW; also BMW3 preference shares). Market cap / valuation: €35.16B as of 2026-07-14 (share price €57.32; −33.0% over the trailing year; P/E ~5.2; dividend yield ~7.6%). Latest financials: FY2025: revenue €133.45B (−6.3% y/y); net income ~€7.29B (roughly flat y/y). TTM revenue €130.70B; TTM net income €6.82B. Q1 2026: EBIT −36% y/y, revenue −8.1% y/y, under pressure from China, US tariffs and competitive intensity. Cited for: "Figure at BMW Spartanburg" as a live paying deployment (ARGUMENT IV §27, §166); "Figure 02 supported production of 30,000+ BMW X3 vehicles over ~10 months" and the Figure 03 phase (logistics sequencing — picking components from unsorted containers into just-in-sequence trolleys) with no robot count, cycle rate, or commercial terms disclosed (OUTLINE IV.1 §36–41, §95; BMW press releases T0458778EN and T0455864EN). Verification: BMW's own release confirms Figure 02 "supported the production of more than 30,000 BMW X3 vehicles over ten months," inserting sheet-metal parts in the body shop; the successor Figure 03 now runs complex sequencing applications in logistics at Spartanburg, adding wireless charging, tactile hand sensors, palm cameras and speech-to-speech audio. BMW has separately announced a first European humanoid pilot at Plant Leipzig and a new "Center of Competence for Physical AI in Production." BMW discloses no unit counts or commercial terms — the thesis's caveat holds. Links:

Mercedes-Benz Group

What it is: The second flagship automotive humanoid customer — Apptronik's Apollo running intralogistics in Mercedes plants — and, unusually, both a *customer* and an *investor* in the robot maker, which is exactly the OEM-capture dynamic the thesis argues a neutral operator would break. Status: Public (XETRA: MBG). Market cap / valuation: €40.94B as of 2026-07-14 (share price €45.11; −15.1% over the trailing year; P/E ~8.7; dividend yield ~7.9%). Latest financials: TTM revenue €130.59B (−8.6% y/y); TTM net income €4.88B (−45.2% y/y); EPS €5.08 (−43.9%). Q2 2026: core car-business sales fell 8% y/y, driven by competition in China. Cited for: "Apollo at Mercedes" as a live paying deployment (ARGUMENT IV §28, §166); "Apollo deployed for intralogistics at Mercedes' Digital Factory Campus (Berlin-Marienfelde) and Kecskemét, Hungary" (OUTLINE IV.1 §52–55); Mercedes-Benz as a repeat investor in Apptronik's $935M Series A (OUTLINE IV.1 §54). Verification and the numbers: Mercedes confirms it is testing Apollo at Berlin-Marienfelde (Digital Factory Campus) and Kecskemét, Hungary, focused on repetitive intralogistics tasks — delivering assembly kits to workers, quality-checking components, moving containers between manufacturing stages. Mercedes has stated only a single-digit number of robots are in test mode, and that it invested a "low double-digit million" euro amount in Apptronik (agreement signed at Marienfelde, March 2025). No commercial terms, unit economics or cycle rates disclosed. Links: