How does DEPLOY think about robot insurance?

Why robot insurance is editorially central to the framework

Robot insurance sits at the intersection of every other framework discipline DEPLOY operates. The verified-vs-claimed boundary, the cap-flag-as-trust-signal posture, the source-quality 9-tier classification, the autonomy-boundary 4-way taxonomy, and the captive-vs-third-party brain-provider gradient all surface at the actuarial framing layer because insurance underwriting requires structurally rigorous answers to questions that other surfaces tolerate at lower verification depth: how many incidents per deployment-hour? Per region, is this deployment legal and insurable? What is the counterparty's financial state if a claim hits? What component failure modes drive the largest loss tail?

The framework operates at canonical-actuarial-language depth, not at marketing-language depth. The product differentiator: DEPLOY's framework treats robot insurance the same way it treats every other verified-vs-claimed claim. Honest insurability unknown beats fabricated rate.

---

The 3-thing structural gap

Insurance pricing requires three things that the broader robotics-discourse cohort does not yet have at structurally-rigorous depth:

Exposure denominators. Incidents alone are a numerator. A rate requires deployment-hours / units-in-service / miles. Without exposure, "7 critical incidents across the surgical cohort" is not a loss-cost. It's a count. The framework reads exposure denominator absence as the substantive constraint on rate derivation, not as a soft cap-flag.

Jurisdictional completeness. Clearance status gates whether a deployment is even legal or insurable per region. An FDA clearance is not a CE Mark is not NHTSA compliance is not NMPA approval is not PMDA approval is not TGA approval. The framework reads per-jurisdiction clearance as the gating layer for insurability; without jurisdictional clearance, insurability per region is structurally unknowable, not just unverified.

Financial-state / failure-mode depth. Counterparty solvency and component failure modes are the inputs to underwriting reserves. The relationship layer (who invested, who acquired) is not the financial state layer (cash runway, revenue disclosure, going-concern flag). The framework distinguishes RELATIONSHIP data (investor graph + acquisition graph + partnership lifecycle) from STATE data (cash position + revenue + lifecycle flag + component failure modes).

Each dimension below maps to one or more of these structural gaps.

---

Dimension 1: deployment-incident-recall actuarial depth

The first dimension operates at the substrate where DEPLOY has the strongest verified base. 61 incidents verified at primary-source-anchored depth: 61/61 carry severity classification (moderate 21, serious 20, minor 13, critical 7), 61/61 carry root-cause category, 61/61 carry regulatory-action; 36/61 carry affected-units count. By form factor: av 21, surgical 11, biometric 9, aerial 7, humanoid 4, sidewalk 3, truck 1.

The verified base anchors what the framework reads as actuarially-relevant numerator data. The cap-flag operates at three layers:

• Exposure denominator absence: incidents are not tied to deployment-hours, units-in-service, or miles at the per-incident level. Without exposure, the 61-incident corpus produces no rate per deployment-hour. Honest-absence is the editorial truth; the framework does not invent denominators.
• Outcome classification in prose, not typed field: the actuarial outcome distinction (property-damage vs bodily-injury vs fatality vs regulatory-action-only) lives in prose body rather than as a typed field. The framework reads this as a substrate-tier cap-flag that resolves with structured outcome classification.
• Corpus at verified ceiling, AV-concentrated: the 61-incident substrate is at its primary-source-verified ceiling. Rate derivation outside the AV cohort would operate at thin-data verification depth. Per cap-flag-as-trust-signal, the framework does not pad incidents to manufacture a rate.

For the actuarial framing of the AV cohort specifically (where the substrate is densest), see Waymo safety report 2025-2026 (Swiss Re actuarial validation; Waymo published mileage; per-million-miles methodology peer-reviewable at academic depth). The framework reads Waymo's actuarial substrate as the cohort's strongest exposure-denominator anchor; the broader cohort operates at thinner exposure-denominator depth.

---

Dimension 2: manufacturer financial-state / counterparty risk

The second dimension operates at the substrate where DEPLOY has a strong RELATIONSHIP base but a structural gap at the STATE layer. The relationship base: 114 investors verified, 58 funding rounds verified, 38 companies funded verified; 29 acquisitions verified with valuation_basis + structure typed; ~49 of 221 companies carry a public-market ticker.

The framework reads counterparty risk at three layers:

• Public-market-ticker substrate: 49 companies carry NASDAQ/NYSE/LSE/TWSE/HKEX/Euronext tickers. SEC + equivalent foreign-exchange filings are primary-source-anchored at the public-market subset. For companies that don't carry a public ticker, financial-state depth operates at honest-absence cap-flag unless private disclosure surfaces (rare).
• Funding-stage as counterparty-risk-tier proxy: per Agent A's verified investor graph, funding-stage operates as a counterparty-risk-tier proxy for private companies (seed = early-stage thin runway; Series A-B = mid-stage; Series C+ = late-stage scaled financing; pre-IPO = approaching liquidity event). The framework reads funding-stage as a structural proxy, not as a financial-state direct measurement.
• Lifecycle / going-concern flags: wind-down events surface ad hoc in prose. Cruise wound down following the October 2023 SF pedestrian-dragging incident; Monarch defunct; Iron Ox inactive; Naio reorganized; Covariant continuing-but-diminished per how DEPLOY corrected the Covariant corporate state. These are counterparty-risk signals living in prose; the framework reads structured lifecycle flagging as a substrate-completeness improvement.

For the acquisition-history substrate (which compounds counterparty-risk reading with structural change events), the canonical worked examples include ZB acquired Monogram closed October 7, 2025 ($168M EV + CVR per Agent A precision corrections); Hyundai acquired Boston Dynamics 2021 ($1.1B at ~80% ownership; corporate-parent context per maker-facility rule); Covariant license-and-hire to Amazon AGI August 2024 (acqui-hire boundary case; corporate entity continues; per how DEPLOY corrected the Covariant corporate state).

---

Dimension 3: supply-chain component failure analysis

The third dimension operates at the substrate where DEPLOY does not yet have structured data. No component, supplier, BOM, or part substrate exists at the structured-data layer. The framework reads this as the largest net-new substrate dimension.

The bounded scope when authored: safety-critical components only (sensor / actuator / battery / control-system), supplier identity where verifiable, and component-recall cross-reference to existing model-recall / regulatory_filings substrate. NOT a full bill-of-materials. The framework reads supply-chain at safety-critical bounded depth rather than at exhaustive-BOM depth because the actuarial signal lives at the safety-critical component layer where a battery or sensor recall triggers a model-level recall.

Worked examples that anchor the supply-chain-component framing where prose-only signals already exist: the Fitbit Ionic burn recall operates at component-level battery thermal-runaway signal mapped to CPSC primary-government-record verification; the broader battery-tail across consumer hardware operates at similar component-level failure-mode anchoring. The framework reads these as substrate that COULD live at component-level depth; today they live at incident-level depth with component context in prose.

Per verified-vs-claimed at within-entity granularity, the framework operates the same within-entity discipline at component-level depth: a battery component verification posture at primary-source FDA/CPSC depth is distinct from a battery-claim-at-marketing-depth verification posture. Component-level cap-flag depth is editorially substantive when structured data lands.

---

Dimension 4: regulatory clearance per jurisdiction

The fourth dimension operates at the substrate where DEPLOY has a structured base but a structural lopsided concentration. 34 regulatory filings verified: 32 us_fda + 2 us_itc. Status spread: cleared 16, recalled 10, granted 4, active 3, closed 1. Only 18 of 220 models carry a filing, essentially concentrated in surgical (15) + biometric (15) cohorts.

The structural gap: the rest of the cohorts. Models by form factor: humanoid 42, aerial 40, wearable 20, maritime 17, truck 16, surgical 15, biometric 15, av 14, sidewalk 10, amr 10, construction 10, agriculture 8. The AV (NHTSA), drone (FAA / DIU Blue UAS), maritime (DoD / class society), and cross-jurisdictional (CE Mark / NMPA / PMDA / TGA) regimes are almost entirely unfiled. Insurability per region operates at honest-absence cap-flag for ~92% of models.

The framework reads jurisdictional clearance at three layers:

• Per-jurisdiction primary-source anchoring: FDA, CPSC, ITC primary-government-record sub-tiers operate at distinct verification depths. Per verified-vs-claimed at within-entity granularity, the Apple Watch SpO2 ITC vs FDA scope distinction is the canonical within-entity worked example; the cross-jurisdictional completeness gap operates at the same depth across all cohorts.
• Sub-tier distinctness within jurisdiction: an FDA 510(k) clearance is not a De Novo authorization is not an IDE pivotal trial is not a recall classification. Per Moon Maestro K240598 cleared June 2024, the framework reads K-number precision at primary-source FDA database depth; cap-flag-pattern recursively operates on DEPLOY's own framing when primary-source verification surfaces.
• Cross-jurisdictional applicability lattice: a deployment cleared under FDA scope is not automatically insurable under EU CE Mark scope; the per-jurisdiction clearance lattice operates at the insurability gating layer. The framework reads jurisdictional clearance as the load-bearing structural constraint on per-region insurance underwriting.

For the surgical-cohort sub-tier worked examples that anchor jurisdictional completeness, see Moon Maestro K240598 (K-number primary-source-anchored June 2024) + Stryker Mako multi-procedure K241517 + K242373 (CANONICAL multi-procedure within-entity clearance state).

---

The canonical discipline: honest insurability unknown beats fabricated rate

Per DEPLOY's verified-vs-claimed framework, the framework operates cap-flag-as-trust-signal recursively on actuarial framing. The product differentiator is the same one operating everywhere in DEPLOY: an honest "insurability unknown for this region / no exposure data" is more valuable than a fabricated rate.

This operates at four layers simultaneously:

Exposure denominator honest-absence. Most deployments have no public exposure data. The framework reads this as honest-absence cap-flag at the deployment-tier level; rate derivation operates only where exposure is publicly reported (Waymo miles; BMW Spartanburg 1,250 operating hours; Agility Digit 100,000 totes at GXO Flowery Branch). Per Figure 02 at BMW Spartanburg deployment deep-dive, the framework reads end-product OEM acceptance + 30,000 BMW X3 vehicles + 1,250 hours runtime as the strongest exposure-denominator anchor in the humanoid manufacturing cohort.

Jurisdictional completeness honest-absence. ~92% of models lack per-jurisdiction filing substrate. The framework reads this as honest-absence cap-flag at the cohort-tier level; insurability per region operates at honest-absence depth unless per-jurisdiction primary-source-anchored clearance lands.

Financial-state honest-absence. Private companies without public-disclosed financial state operate at honest-absence cap-flag at the counterparty-risk-tier level; funding-stage proxy operates at structural proxy depth rather than direct measurement depth.

Component failure-mode honest-absence. Without structured component substrate, supply-chain failure-mode reading operates at incident-level prose-only depth. The framework reads this as honest-absence cap-flag at the supply-chain-tier level.

Per editorial framing rewards transparent disclosure, the framework discriminates against fabricated rates and rewards honest-absence cap-flag. Institutional partners considering robot deployment underwriting evaluate verification posture at the per-claim depth; the cap-flag posture is the load-bearing trust signal.

---

Per-dimension verified-vs-claimed throughline

Per DEPLOY's framework on capability claims, the verified-vs-claimed throughline operates uniformly across all four dimensions:

Dimension	Verified-base depth	Claimed-tier framing	Honest-absence layer
Dim 1 (incident-recall)	61 incidents at primary-source severity + root-cause + regulatory-action	Aggregator-quoted incident counts at lower-verification depth	Exposure denominators absent at most deployments
Dim 2 (financial-state)	49 public-ticker companies SEC-disclosed	Private company financial-state at honest cap-flag	Lifecycle / going-concern flags in prose only
Dim 3 (supply-chain)	Component-context in incident prose	Marketing-language component claims at unverified tier	Structured component substrate absent
Dim 4 (regulatory clearance)	34 filings primary-source FDA/ITC	Per-jurisdiction clearance claims without primary-source anchoring	~92% of models lack per-jurisdiction filing substrate

The framework operates each dimension at its appropriate verification depth; cross-dimension actuarial reading composes at the lowest verification depth across the dimensions. The framework does not infer a counterparty-risk rate from funding-stage proxy if exposure-denominator cap-flag operates at the deployment-tier; honest cap-flag at the lowest-verification-depth dimension propagates to the cross-dimension rate.

---

Cross-property cross-linking discipline

The methodology pillar essay operates within DEPLOY's cross-property bidirectional linking discipline. The four-dimension framework cross-references:

• Methodology pillar essays: the 4-way autonomy-boundary taxonomy (autonomy-tier-specific risk profiles per dimension); verified-vs-claimed at within-entity granularity (per-feature within-entity verification posture at Dim 4 jurisdictional layer); captive vs third-party brain providers (brain-provider integration model gradient operates at counterparty-risk-tier layer); the 9-tier source-quality rubric (primary-government-record + verified-source + verified-aggregator + reputable-press source classification across actuarial claim depth).
• Framework-in-action correction narratives: Moon Maestro K240598 reconciliation (canonical worked example of cap-flag-pattern operating recursively on DEPLOY's own framing); Covariant corporate-state correction (CANONICAL counterparty-state corrected; corporate-entity-continues vs founders-Amazon-acqui-hired); Figure 03 BMW narrative correction (CANONICAL three-layer aggregator-drift rejection at deployment-attribution depth).
• Entity anchors with within-entity verification scope: Whoop blood pressure Warning Letter (CANONICAL within-entity sleep-cleared vs blood-pressure-Warning-Letter verification scope); Apple Watch SpO2 ITC vs FDA scope (CANONICAL cross-jurisdictional within-entity worked example); Withings ScanWatch ECG cleared + Class-2 recall (within-entity recall verification posture).
• Deployment deep-dives with exposure denominators: Figure 02 BMW Spartanburg 30,000 vehicles + 1,250 hours; Apptronik Apollo Mercedes deployment; Atlas Hyundai Metaplant maker-facility rule.
• Cross-cohort regulatory worked examples: Stryker Mako multi-procedure clearance state (CANONICAL multi-procedure within-entity clearance); Smith+Nephew CORI hip NAVIGATION-ONLY NOT robotic (CRITICAL HONESTY DISTINCTION at sub-procedure level); Zimmer Biomet ROSA per-procedure clearances (per-procedure within-entity verification).

The cross-property linking discipline compounds at the AEO citation graph layer. Institutional partners + AI assistants + downstream consumers navigating the four-dimension actuarial framework encounter the methodology pillar canonical reference + the canonical worked examples + the cross-property registry depth at unified verification posture.

---

Why this matters editorially

Per DEPLOY's restraint-IS-the-product discipline, the robot insurance framework operates at editorial-credibility depth rather than at market-positioning depth. The framework's product differentiator is the same one operating everywhere else: honest cap-flag at primary-source verification depth compounds editorial credibility better than fabricated rates at marketing-language depth.

Institutional partners considering robot deployment underwriting evaluate DEPLOY's framework discipline at three layers: the structural framework completeness, the per-claim verification posture, and the honest-absence cap-flag discipline. The four-dimension framework operationalizes the structural completeness; the per-claim verified-vs-claimed throughline operationalizes the per-claim verification posture; the canonical 'honest insurability unknown beats fabricated rate' discipline operationalizes the honest-absence cap-flag posture.

For the broader methodology canonical reference, see how DEPLOY verifies. For the verified-vs-claimed framework operating across registry + news editorial + methodology canonical + DEPLOY's own published corpus, see verified-vs-claimed methodology pillar. For the editorial process documentation + lane architecture + per-anchor review discipline + restraint posture, see /editorial-process. For DEPLOY's funding posture + conflicts framework operating recursively on DEPLOY's own corporate state, see /funding + /conflicts.