Allometry · Lending · T1 · Insurance-Linked Capital
T1 24 mo vintage to unlock

Insurance prices distributions. Banks price points.

Banks underwrite a single borrower against equipment residual and personal guarantees. Insurers must believe a population of borrowers behaves like the population already in their book. That is why insurance is the hardest tier to unlock — and the most valuable. With 100+ operators in the cohort and 24 months of attested data, Allometry becomes the actuarial table no carrier has access to today. Munich Re. Swiss Re. Lloyd's. Eventually: Allometry IS the carrier.

$1M – $250M+Typical coverage limit
6–18 monthsNew program · Jan 1 cycle
100+ assetsStatistical credibility floor
$21.1B → $23.85BParametric market 2025→26 (13% CAGR)
§ 01 · Who's in this tier

Thirteen named carriers. From global reinsurers to parametric MGAs.

Three layers: global reinsurers (Munich Re, Swiss Re, Hannover Re) who provide capacity to everyone else; Lloyd's market syndicates (Beazley, Hiscox, Atrium, Nephila) writing specialty subscription business; and emerging parametric MGAs (Parametrix, Descartes) who measure events independently and pay fast. Lloyd's and Hannover Re backed Parametrix specifically because Parametrix could independently measure cloud outages. The same template applies to field-service operators — independent operator-side measurement is what makes the cohort insurable.

Munich Re
#2 global reinsurer (IFRS 17) · leads parametric / nat-cat R&D · renewables, smart-city infra, climate parametric
$25M–$500M+ · treaty
Swiss Re
#1 global reinsurer (2024 AM Best ranking) · treaty + facultative on solar / wind / data center · structured solutions
$25M–$500M+ · treaty + fac
Hannover Re
#3 global · backed Parametrix · pioneer in cloud-outage parametric bonds · first-mover on parametric structures vs ILS
$5M–$200M · facultative
Lloyd's syndicates
Beazley, Hiscox, Atrium, Nephila S2358/S2359 · specialty subscription market · coverholder paper for MGAs · cyber, energy, MAP
$1M–$140M · per syndicate share
Nephila Capital (Markel)
ILS fund mgr · $7.7B 3rd-party AUM · Q1'26 rev +59% YoY · Syndicate 2358 GWP +129% to $508.6M in 2025
$5M–$100M+ · collateralized
Markel (Programs)
Largest non-affiliated program carrier · warranty, casualty, package · targets MGA programs at $10–15M annual premium each
Program-level partnerships
AXA XL
Energy, power, renewables · Structured Risk Solutions writes Performance Insurance for new tech · operational PD + BI
$10M–$250M+ · energy
Allianz Commercial
Renewables, mining, conventional power BI · expanded cyber JV with Coalition (2025) · operational PD + BI on power gen
$10M–$500M · operational
Tokio Marine HCC (Surety)
Contract performance / payment / maintenance bonds · energy operator + ROW bonds · up to $50M single-bond capacity
$1M–$50M · surety
Coalition
Active cyber MGA · IoT-originated incidents · Allianz-backed · OT/cyber-physical for connected equipment
$1M–$25M · cyber tower
Descartes Underwriting
Parametric MGA · new data-center product suite Apr 2026 · climate, cyber, infra · up to $140M per policy flexible capacity
$5M–$140M · parametric
Parametrix
Lloyd's coverholder · data-center SLA + cloud-outage parametric · raised $27M Series B (Hannover Digital led) · $50M corp cloud-outage · 15-day pay
$1M–$50M · parametric SLA
NormanMax / FloodFlash
IoT-sensor parametric flood MGA · Lloyd's coverholder · NormanMax acquired FloodFlash 2025 · location-bound parametric
$50K–$10M · parametric flood
§ 02 · The products

Five product types. One operator data substrate underneath.

Insurance products relevant to asset-heavy operators all share one structural requirement: independent, verifiable measurement of the underlying risk. The carrier (or its reinsurer) needs to know whether the SLA actually breached, whether the asset actually failed, whether the contracted delivery actually happened. Today most of that measurement is operator-reported, which is uninsurable at scale because of adverse selection. With ZKP-attested operator data, the same products become writable at cohort scale — and at materially tighter combined ratios.

▸ Parametric · fast-pay

Parametric SLA wraps

Payout triggered by measured performance drop (uptime, throughput, response time). Parametrix is the canonical example: 15-business-day payment, no adjuster. The trigger is the data — which is exactly what the vault produces.

▸ Warranty · OEM-extension

Equipment warranty wraps

OEM warranty extensions backed by insurance paper. Blink × EVSTAR's 5-year EV-charger protection plan is the template. The carrier needs failure-rate distributions; the vault produces them per SKU and vintage.

▸ Operational · all-risk + BI

Business interruption

AXA XL, Allianz Commercial cover physical damage + revenue loss from asset failure. Rated against replacement cost, typically 0.3–1.5% of insured value. Loss-cost ratio improves when historical loss data is hash-chained.

▸ Surety · contracted delivery

Performance bonds

Tokio Marine HCC up to $50M per bond on solar O&M contracts. 1–3% of bonded amount per annum. The vault evidences delivery history — the substrate for premium reduction.

▸ Hybrid · cyber × OT

Cyber-OT towers

Coalition (IoT origin), Beazley Cyber Risks, Allianz/Coalition JV. Covers operational disruption from connected-equipment compromise. $1M–$5M retention, $5M–$25M limits. The vault's IoT-sourced data is the underwriting evidence.

▸ Performance · new-tech

"Performance Insurance"

AXA XL Structured Risk Solutions writes guarantee paper on new / renewable tech performance. Emerging hybrid bridging warranty and insurance — the vault is the data layer that lets it scale.

§ 03 · Why insurance is the hardest tier

A four-subject sample cannot fit a distribution.

Banks underwrite a point estimate of probability-of-default times loss-given-default. They can fall back on equipment residual, personal guarantees, and AR collateral when the model is uncertain. Insurance has no such fallback. The carrier must fit the entire shape of the loss curve — mean, variance, tail — and price the tail at 99th or 99.5th percentile. That requires a population. Until the population exists, the math doesn't.

The credibility math everyone underweights.

~1,082 claims ~27,000 exposure units 100+ comparable risks

Actuarial Standard of Practice 25 ties full credibility to claim counts. At standard parameters, full credibility requires roughly 1,082 claims. If expected frequency is 4 claims per 100 units, that's ~27,000 exposure units. Partial credibility kicks in much earlier — but a cohort below 100 comparable risks is rarely credible enough for a new program. Banks lose principal; insurers lose multiples of premium when tails fire. The asymmetry justifies the data appetite.

The Parametrix precedent is the model. Hannover Re, Lloyd's syndicates, and Lockton backed Parametrix specifically because Parametrix could independently measure cloud outages with sensor-grade fidelity. Operator self-report would have been uninsurable. The same template applies to field services — independent measurement is what unlocks the cohort.

§ 04 · The endgame

Phase 1: MGA. Phase 4: own the syndicate.

Routing premium to Munich Re is the obvious play. Becoming the carrier is the better one — because the proprietary cohort data is the underwriting moat, and the underwriting moat is the rating-agency capital adequacy that lets a carrier exist at all. The path is well-trodden: Lemonade, Hippo, Coalition all walked it. The binding constraint is not regulatory capital. It's cohort data + reinsurer willingness to cede capacity. The vault produces both.

MGA on rated paper

▸ Now → 24 mo

Program-business MGA on Markel / AXA / Lloyd's paper. Markel targets $10–15M annual programs — achievable on a few hundred operators at $100–300K each. Allometry handles distribution, underwriting, and claims; the rated carrier holds the policy.

Rented cell / PCC

▸ 24 → 36 mo

Vermont, Bermuda SAC, or Guernsey protected-cell company. Setup in weeks. Minimum participation premium ~$500K. Capital often <$250K per cell vs $5M+ standalone captive. Allometry takes 10–30% quota share alongside fronting carrier.

Own captive · fronted

▸ 36+ mo

Own captive or sponsored captive (Green Mountain Vermont model). Fronted by AIG or Marsh-managed facility. Reinsured by Munich Re / Hannover Re / Nephila against tail risk. Allometry holds proportionally more of the underwriting profit.

Lloyd's syndicate · scale

▸ Long horizon

Lloyd's syndicate or full carrier license. Bermuda Class 3A typically $5M minimum capital. US admitted carrier $5–10M+ plus RBC. The proprietary cohort data is the asset that unlocks rating-agency capital adequacy.

§ 05 · What the vault unlocks

Adverse selection collapses when the data can't be cherry-picked.

The fundamental problem with insuring field-service operators today is adverse selection: operators with hidden churn, hidden downtime, hidden warranty exclusions self-select into coverage. Carriers respond by loading premium with an "ignorance premium" — pricing the worst-case operator into every quote. Cryptographically attested operator data eliminates adverse selection structurally. Operators can't hide what's in the hash chain.

Vault evidence (continuously attested)What it unlocks · insurance economics
Real-time loss-event streams
vs annual loss-run snapshotsIBNR uncertainty eliminated
Replaces the annual "as-of" loss run that's already stale at submission. Reserve calculations tighten; carrier confidence in reserves rises; combined ratio target tightens.
Hash-chained operational metrics
uptime, MTBF, MTTR per SKU and vintageAdverse selection priced out
Operators cannot cherry-pick which events to report. Carriers can fit distributions from full population data, not curated submissions. Loss-cost ratio improves systematically.
ZKP-attested SLA compliance
cohort-level without exposing per-operator dataCohort priced as a cohort, not worst-of
Letting carriers verify proofs without seeing the underlying operator data lets the cohort be priced collectively — the price reflects cohort risk, not the worst observed operator.
Equipment manufacturer warranty data
OEM warranty terms · recall history · residualsTail risk priced from data, not from theory
Carriers can fit failure-rate distributions per OEM-vintage cohort. PML calculations tighten. Reinsurance attachment points move favorably.
Contract terms parsed
SLA definitions · liquidated-damages clauses · concentration exposureWording disputes eliminated
Parametric triggers can be set against parsed-contract definitions, not interpreted ones. Basis-risk on parametric products tightens.
Workforce data
technician certifications · training hours · supervisor ratiosOperational risk underwriting
Workforce quality becomes a measured underwriting input. Carriers can price the human-execution component of operational risk, not assume worst-case.
Geographic / climatic location data
COPE per site · nat-cat exposure · supplier concentrationPer-site precision · not per-portfolio averaging
Nat-cat reinsurance attachments price more accurately. Supplier-concentration coverage becomes underwritable for the first time.
§ 06 · When T1 unlocks · 24-month vintage

Twenty-four months is where the distribution stabilizes.

The 24-month threshold is anchored to insurance-specific math. Actuarial credibility builds with both claim count and exposure-year accumulation. Six months produces a single seasonal slice; twelve captures one annual cycle; twenty-four captures two — which is the threshold where failure-rate distributions, SLA compliance variance, and tail behavior begin to stabilize against expected parameters. Below 24 months, partial credibility ratios are so low that carriers price the cohort at "ignorance premium" — usually 3–5× rational pricing — or decline.

For operators today: T1 unlocks parametric SLA wraps, equipment warranty wraps, and performance bonds priced from real distributions rather than theoretical worst-case. For Allometry as a future carrier: T1 vintage is the point where MGA programs (Phase 1) become economically viable at scale. The 24-month threshold is where the actuarial table goes from theoretical to real.