External risk benchmark for robotaxi insurers to quote fleets using live recalls, incidents, and market-expansion signals.

1. A benchmark that refreshes every 12 hours turns AV diligence into a live workflow instead of an occasional market study.
2. Public recalls and crash investigations now create immediate repricing signals for carriers covering robotaxi fleets.
3. Texas has made fleet-scale transparency directly useful to insurers by enforcing commercial AV rules and exposing a public tracker.
4. Fleet sizes and launch plans are changing faster than annual underwriting cycles, so static diligence goes stale quickly.
5. Visible performance gaps between operators mean carriers can now benchmark who deserves capacity and who should pay for volatility.

Catalyst. Autnmy's 12-hour benchmark, Texas's new public AV fleet transparency, and fresh Waymo and Avride safety incidents make external AV risk comparison both newly possible and newly urgent.

The product gives carriers, brokers, and reinsurers a live external- intelligence layer for each robotaxi operator and market pair. It pulls benchmark data, public fleet trackers, recall notices, incident investigations, partnership announcements, and commercialization signals into a normalized exposure model covering scale, safety, regulatory friction, and operational maturity. Underwriters can compare operators side by side, see when a recall or launch changes a risk profile, and export a timestamped diligence memo for quote committees, renewal reviews, or reinsurance partners. Over time, the platform can accept operator-submitted evidence and claims outcomes to calibrate pricing, but its first value is replacing bespoke AV research with a comparable market reference.

What's different. Most AV tooling sells into the fleet operator and helps the vehicle company build or prove autonomy internally. This company starts from the outsider who bears balance-sheet risk and needs an explainable, third-party view across operators and markets. If it becomes the system carriers use to compare, document, and monitor AV programs, it can accumulate proprietary pricing benchmarks, operator histories, and portfolio signals that are hard for point consultants or operator-built dashboards to replicate.

Jobs to be done

flowchart LR
  Buyer[Autonomous mobility underwriter] --> Pain[Cannot price robotaxi fleets with sparse loss data and fast-changing safety signals]
  Pain --> Product[Robotaxi risk-rating workbench]
  Product --> Outcome[Faster quotes, better renewals, and defensible AV capacity decisions]

Executive takeaways

• The wedge is real because public AV signal density is finally high enough to support a recurring underwriting workflow instead of a once-a-year market note.
• The beachhead is narrow and high-touch, so the product should sell around live Texas launches, recalls, and renewals before broadening into reinsurance and adjacent autonomy categories.
• The hardest competition is indirect—broker advisory, reinsurer analysis, and internal spreadsheets—rather than a crowded field of insurer-first software.
• The product should start as explainable decision support, because fragmented regulation and thin claims history still make black-box AV pricing hard to trust.

Market definition

The relevant market is external underwriting intelligence for commercial robotaxi programs: software and analyst workflows that turn recalls, incident reports, fleet authorizations, city launches, safety cases, and partner disclosures into operator-market risk comparisons and committee-ready renewal memos.

Customer and buyer

Daily users are specialty autonomous-mobility underwriters, broker placement teams, and reinsurance or actuarial innovation leads who must compare robotaxi operators without mature loss curves. The economic buyer is usually the chief underwriting officer, head of specialty programs, or business leader who owns autonomous mobility capacity.

Buying triggers

• A new city launch or service-area expansion forces a carrier or broker to reassess exposure before historical claims data has time to accumulate. [7][9][17][25][26][27][35]
• A recall, crash investigation, or regulatory action creates an immediate repricing or renewal-committee event. [13][19][20][37]
• A reinsurer, capital provider, or quote committee asks for a neutral comparison of operators rather than another self-authored safety deck. [2][34][41][44][45][46]

Willingness to pay

Willingness to pay is credible because the product plugs into an expensive existing problem. Commercial auto has suffered persistent underwriting losses, AV liability is drifting across auto, product, cyber, and professional lines, and brokers already describe bespoke multi-party program design as hard manual work. A six-figure annual tool is plausible if it shortens quote cycles, improves committee defensibility, or prevents even one badly priced launch or renewal. [41][44][45][46][47][48]

Category dynamics

Growth signal High-double-digit fleet growth through 2035; BCG sees ~1.32M robotaxis across the U.S., China, and Europe in its base case by 2035 with upside to ~3M globally, while Goldman frames a ~$400B-plus robotaxi market by 2035.

Validation signals

• AUTNMY AI launched a live benchmark product, proving that external AV-comparison workflows are now commercially plausible on public and licensed data.
• Waymo has crossed from pilot novelty into scaled recurring service, with large weekly ride volumes, multi-city expansion, and manufacturing investment.
• Uber is standardizing autonomous distribution across multiple operators, creating a repeatable stream of city launches and partner-risk comparisons for insurers to track.
• Commercial auto insurers remain under pressure from loss severity and underwriting strain, so better diligence has a clear budget and profitability story.

Regulatory & technical constraints

• Texas and California have materially different data, testing, and deployment rules, so a single cross-market risk score must always be jurisdiction-aware.
• NIST still describes standardized AV performance metrics and repeatable testing procedures as work in progress, which limits how deterministic any external rating can be.
• Liability can swing between auto, product, cyber, and professional lines depending on supervision mode and system behavior, which complicates simple one-number risk outputs.
• Some operators disclose richly while others do not, so missing data must be estimated conservatively or withheld rather than presented with false precision.

Competition is real but mostly indirect. Direct benchmark publishers are emerging, yet most real substitutes are still broker or reinsurer advisory teams, operator safety decks, and in-house watchlists stitched together around live launches and incidents. The gap is an insurer-first control layer that stays neutral across operators, time-stamps every source, and turns signal changes into quote and portfolio actions quickly.

Why incumbents do not win by default

• AV benchmark publishers. Broad leaderboards validate the category, but they do not win by default because underwriters need explainable operator-market memos, override logs, and portfolio monitoring rather than a public ranking alone.
• Global brokers. Brokers like Marsh and Aon already help structure AV programs and liability wraps, but their service model is high-touch and episodic rather than a live, neutral system of record for ongoing operator monitoring.
• Reinsurers and capital providers. Reinsurers can fund research and shape underwriting frameworks, but their tools are optimized for portfolio strategy and capacity decisions, not for daily quote-committee workflows inside carriers and MGAs.
• Operator self-disclosures. Waymo, Pony.ai, Baidu, and others publish useful safety and growth data, but each disclosure is self-framed and hard to compare across jurisdictions, ODDs, and supervision models without a neutral normalization layer.
• Mobility platforms. Uber is becoming the distribution layer for multiple AV operators, but platforms do not naturally serve as neutral insurer-grade arbiters of which operator or city deserves capacity.

Robotaxi Risk Rating should start as a Texas-first underwriting intelligence workbench for specialty MGAs, carriers, and brokers quoting U.S. robotaxi excess-liability programs, not as a general AV analytics platform or automated pricing engine. The first customer is a specialty autonomous-mobility program lead facing a live launch, renewal, or post-incident repricing cycle where historical claims are thin and operator disclosures are self-framed. The wedge is credible because public-signal density has crossed a threshold: AUTNMY's 12-hour benchmark, Texas fleet transparency, recalls, crash investigations, and launch announcements now create a recurring diligence workflow instead of a one-off research note. Version one should stay read-heavy and explainable, combining source-linked operator-market scorecards, material-event alerts, and committee-ready quote or renewal memos delivered in hours rather than weeks. The first proof point is not model accuracy in the abstract; it is whether one MGA, broker, or reinsurer changes terms, limits, or monitoring intensity based on a live memo generated from the platform. The beachhead is narrow—research sizes it at roughly $33.8M TAM, $6.8M Texas-led SAM, and $2.7M reachable year-3 SOM—so the venture case depends on multi-team expansion within each program and later reuse across reinsurance, delivery, trucking, or financing workflows. The biggest disconfirming risk is that underwriters keep manual or broker-assisted workflows because they trust human judgment more than a new category-specific system or will not pay six figures for a mostly external-data product. The input research does not quantify current AV diligence budgets or claims-sharing willingness, so the company should raise a pre-seed round only to prove paid pilots, production conversion, and at least one credible adjacency before scaling GTM headcount.

Problem

• Specialty underwriters, MGAs, and brokers must quote or renew robotaxi programs even though long-run claims history is still thin and most operator disclosures are self-reported.
• Recalls, incidents, city launches, and regulatory changes can alter exposure in days, but current workflows still rely on analyst memos, consultant reports, operator decks, and spreadsheet watchlists.
• The default substitutes are episodic and hard to compare across operators, jurisdictions, and supervision models, which makes quote committees slow, inconsistent, and vulnerable to mispricing.

Solution

• Build a live external-intelligence workbench that normalizes benchmark, recall, incident, fleet-authorization, and launch data into source-linked operator-market scorecards across safety, scale, regulatory friction, and operating maturity.
• Start with event alerts, side-by-side comparisons, and timestamped quote or renewal memo export, then add override logging and operator rebuttal workflows so underwriters can defend decisions without relying on opaque automated pricing.

Why we win

• The product sells to the outsider who bears balance-sheet risk, not to the robotaxi operator, so it can stay neutral across competing fleets and markets.
• Public leaderboards and broker advisory do not win by default because underwriters need state-aware workflow, evidence lineage, override logs, and committee-ready outputs rather than a ranking alone.
• A time-stamped history of operator-market events tied to underwriting actions, rebuttals, and portfolio decisions can compound into a proprietary dataset that is harder to copy than raw news monitoring.

Milestones

flowchart LR
  Wedge[Texas robotaxi underwriting wedge] --> MVP[Source-linked risk workbench]
  MVP --> Proof[Paid pilots and faster committee memos]
  Proof --> Expansion[Multi-team portfolio monitoring and adjacent autonomy risk workflows]

Founding team

Experiment roadmap

Risk assessment

What must be true

• At least 3 of the first 5 target MGAs, carriers, or brokers must accept a software-priced pilot instead of asking only for custom advisory.
• At least 50% of paid pilots must convert to $100k+ annual subscriptions after one live quote or renewal cycle.
• In at least 3 recent underwriting events, a recall, incident, or launch signal must have changed terms, limits, or monitoring within 24 hours.
• Texas-first public and licensed data must be sufficient to draft a committee-ready memo for most early programs without confidential operator integrations.
• By month 18, at least 2 production customers must add a second paying team or adjacent autonomy workflow, proving the wedge can expand.

Open diligence questions

• What do MGAs, carriers, and brokers spend today per AV quote or renewal on consultants, research, and internal analyst time?
• Which public or licensed signals actually cause term, limit, or capacity changes in live quote committees?
• How often will a chief underwriting officer pay for a neutral third-party memo instead of relying on broker advisory or operator decks?
• Can a source-linked score survive operator disputes without damaging carrier or broker relationships?
• Which adjacent workflow—reinsurance review, autonomous delivery, trucking, or financing diligence—reuses the same data model first?

Model sanity

• Revenue engine. The base case grows from 3.7 paying teams at Q4Y1 to 19.3 at Q4Y3, with second-team expansion inside shared programs doing almost as much work as first-team wins.
• Must go right. Paid pilots have to convert into $150K production watchlists fast enough that one GTM hire can support roughly one new paying team every one to two months through Y2 and Y3.
• Model breaks if. If production ARR falls toward $140K or sales cycles slip by a quarter or two, the downside case drops Y3 revenue to about $1.6M and drives cash roughly $561K below zero.
• Next-round proof. The next financing is justified once AUTNMY has 2+ six-figure production conversions, California overlays, and at least one shared program with multiple paying teams or an adjacent workflow win by month 18-24.

Scenarios

Sensitivity

flowchart LR
  Leads --> PaidPilots
  PaidPilots --> ProductionTeams
  ProductionTeams --> SharedPrograms
  SharedPrograms --> Revenue
  Revenue --> GrossProfit
  GrossProfit --> Cash

Flags: Base ARPU uses the top of the BP's stated $100K-$150K production range, so pricing proof is a real gating risk. · Revenue is still concentrated in about 8 active programs by Y3, so losing one large program would materially dent growth. · Full-year Y3 EBITDA stays negative even though Q4Y3 is near breakeven, so the next round still depends on expansion proof rather than pure profitability. · Gross margin only reaches the 70% target if licensed data and memo QA stay productized instead of turning into analyst-heavy services.

• Beachhead is narrow. Only a limited number of carriers, brokers, and reinsurers actively underwrite early robotaxi programs today. Mitigation: Start with the highest-urgency mobility programs, then expand the same workflow into autonomous delivery, trucking, and financing diligence.
• Carriers may stay conservative. Some underwriters may prefer manual judgment and wait for more claims history before trusting a new category-specific workbench. Mitigation: Position the product as explainable decision support and committee documentation rather than black-box automated pricing.
• Operators can dispute external scores. Robotaxi companies may challenge adverse ratings if public data is incomplete or contextual details are missing. Mitigation: Keep every score source-linked, allow operator evidence uploads, and make the rating logic transparent enough for underwriters to override with documented rationale.