Insurance · Deep dive

Corgi

An AI-native insurer for startups that quotes in 30 seconds and books premium like software revenue — with the risk sitting in an unrated, member-owned risk retention group it also controls.

emerging

The question that decides it: Corgi's risk sits in Technology Risk Retention Group, an unrated Arizona RRG reinsured inside its own corporate family, and it wrote long-tail D&O, E&O and cyber on barely a year of loss history. Do the 2025-2026 accident years develop inside its pricing assumptions — or does the first adverse year land on a carrier with no AM Best rating and no state guaranty fund behind it?

Snapshot

Corgi sells business insurance to startups — general liability, D&O, tech and AI E&O, cyber, EPLI. A founder logs in, uploads a pitch deck, and gets a bindable quote in about 30 seconds. Founded in 2024 by two Y Combinator alumni with no insurance background, it crossed $40M in annualized premium by December 2025 and has raised roughly $374M. Its valuation went from $630M in January 2026 to $1.3B on 6 May to $2.6B on 28 May — doubling in three weeks, largely from the same investors. Underneath the AI-native carrier story is the structure most coverage skips: the risk does not sit on a rated insurer’s balance sheet. It sits in an unrated Arizona risk retention group owned by the policyholders themselves.

Founding story

Nico Laqua’s father spent his career as a lawyer for USAA. The story Laqua tells is watching him hunt-and-peck at a keyboard, processing claims by hand, and concluding that insurance was an industry made of words, waiting for a machine that could read them. He studied neuroscience and computational biology at Columbia, then built games — Basket Entertainment, a portfolio he says reached over 200 million monthly actives. Emily Yuan dropped out of Stanford in her junior year; she and Laqua had already built Picnic, a Gen Z social app, after he cold-sent her a project for feedback. She made Forbes 30 Under 30 in 2024 and is reported to have been a product manager at OpenAI.

GPT-3.5 was the trigger. The two went through Y Combinator’s 2024 batch and did what almost every insurtech does first: they became a broker. It worked — by their own account it was growing and profitable. Then they shut it down and spent the better part of two years in stealth building a carrier instead, because a broker earns a commission on someone else’s product and can never fix the product. That decision is the whole company. It is also the source of every risk on this page.

How it works

The most important structural fact about Corgi is the one it markets around. Corgi Insurance Services, Inc. is the program administrator. The paper is issued by Technology Risk Retention Group, Inc., an Arizona-domiciled risk retention group. An RRG is a member-owned liability insurer created under the federal Liability Risk Retention Act. It can write across state lines without licensing in each one — which is how a two-year-old company reached 49 states so fast. It is also confined by law to third-party liability, exempt from state guaranty funds, and carries no AM Best rating. Buy a Corgi policy and you become a member-owner of the carrier insuring you. A captive reinsurer inside Corgi’s own corporate family sits behind it.

So: who carries the risk? Corgi’s own policyholders and Corgi’s own captive — not a rated third-party carrier, not a fronting partner. Before RRG approval in July 2025 it placed business as a broker/MGA. After it, Corgi kept the underwriting margin and the float. It also kept the tail.

The front end is the part investors bought. A founder signs in with a company email or drops a YC or Crunchbase URL and Corgi auto-populates the file. Its underwriting engine — internally called Hammurabi — ingests pitch decks, SOC-2 reports and GitHub repositories alongside a short questionnaire on headcount, stage, revenue and cloud stack, and returns priced, toggleable coverage modules in roughly 30 seconds. Inside auto-bind limits the founder pays by card or ACH and the certificate issues immediately, with policy data available as JSON over an API. Anything outside the box routes to a human underwriter with a model-generated recommendation attached.

Claims run through the same stack. On 29 June 2026 Corgi launched Corgi Claims, an in-house TPA pairing a network of more than 5,000 licensed adjusters with an AI layer that scores severity, flags coverage issues and pulls missing documents at first notice of loss, before an adjuster opens the file.

Product and business overview

The startup program is sold as modules: commercial general liability, D&O, tech and AI errors and omissions, cyber, EPLI, media liability, fiduciary, hired and non-owned auto — pre-bundled by stage (pre-seed/seed, Series A, growth). The pitch is not that any single coverage is novel. It is that a founder can assemble the whole program without a broker, at 11pm, in the ten minutes before signing an enterprise contract that demands a certificate of insurance.

Distribution is direct and ecosystem-native: the YC network, founder word-of-mouth, and embedded integrations. Named customers include Deel and Artisan. Product leadership said in June 2026 that Corgi intends to open its infrastructure to rival carriers — the pivot from insurer to insurance infrastructure company, which is also the pivot from a 3x revenue multiple to a software multiple.

Business model and pricing

Corgi collects premium and keeps what it does not pay out. That is the model, and it is why the valuation math is contested. Its published guidance (Corgi’s own blog, 2026): a pre-seed or seed company pays roughly $2,000–$4,000 a year for a core CGL, tech E&O and cyber bundle at $1M limits; a Series A adding D&O and EPLI runs $10,000–$25,000; growth-stage runs $30,000–$100,000-plus.

Now hold that against the traction. Sacra puts average revenue per customer at roughly $1,000 a year across 40,000-plus customers (December 2025) — below the bottom of Corgi’s own published pre-seed range. Either the book is overwhelmingly the smallest, cheapest policies, or many of those 40,000 accounts are minimal-premium certificates bought to clear a contract requirement. Both readings fit the testimonials that Corgi quoted at a fraction of a broker’s price. The unanswered question is whether that price is an efficiency dividend or a rate that has not yet met its losses.

Traction over time

What is absent from that table is everything an insurance analyst would ask for first: loss ratio, combined ratio, reserves, how much risk is ceded. None of it is public.

Market analysis

US commercial insurance ran roughly $310B in premium in 2025 (The Report Cubes), liability lines about a third of it; cyber alone was around $16B globally in 2025 with credible paths to $40B by 2030 (MarketsandMarkets). Corgi’s serviceable slice — venture-backed technology companies — is a small fraction of that, but it is the fraction where AI liability, model errors and algorithmic-bias exposure are being invented in real time and no incumbent has a credible policy form yet. That is genuine greenfield.

The cyclical picture is less friendly. Management liability entered 2025 oversupplied, with D&O rates down about 5% and renewals repricing lower without competition (Aon, 2025). A soft market is the worst environment in which to prove a pricing edge is real: incumbents can simply match you, and the margin your technology is supposed to earn gets competed away.

Competitive intel

See the competitor set above. The sharpest fight is with Vouch, which took the opposite structural bet — divesting its carrier to become a broker placing risk with AM Best-rated paper — and now runs a comparison page arguing that Corgi’s RRG is unrated, outside state guaranty funds, legally confined to third-party liability (no property, crime, workers comp), and may not satisfy the insurance clauses in enterprise contracts and leases. That is the single most effective objection any competitor can raise, and it gets stronger as Corgi’s customers grow into contracts with rating requirements.

History and evolution

• 2024 — Founded; Y Combinator; operates as a startup insurance brokerage.
• 2024–2025 — Shuts the profitable brokerage, goes into stealth to build a carrier.
• Jul 2025 — Receives full approval; begins writing through Technology Risk Retention Group across 49 states.
• Dec 2025 — Affiliated renters-insurance platform Beagle sues AppFolio (N.D. Cal.), alleging it tried to destroy Beagle’s business after losing insurance customers to it; AppFolio countersues.
• Jan 2026 — Emerges from stealth with $108M at roughly $630M.
• Apr 2026 — Sues Vouch and its chief legal officer in Delaware, alleging she formed a shell LLC (Augmenta Advisory) to submit a fake application and extract Corgi’s underwriting logic. Vouch calls the claims false and a PR campaign, and later says a Corgi co-founder probed Vouch’s platform two years earlier.
• Apr–May 2026 — Sues Matcha, a rival founded by its own former chief of staff and engineers, over trade secrets, then over an alleged anonymous Wikipedia campaign to delete Corgi’s page. Matcha calls it bullying and forum-shopping out of California, where the non-competes would be void.
• May 2026 — $160M at $1.3B; three weeks later $106M at $2.6B.
• Jun 2026 — Launches Corgi Claims, the in-house AI TPA.

What people say

The case for. The customer feedback is unusually good, and specific rather than vague. Founders on Corgi’s customer page and in Product Hunt commentary describe getting a quote and certificate fast enough that they double-checked it was real, clearing an insurance requirement in minutes to close a seven-figure enterprise contract, and — repeatedly — pricing far below a broker’s, in one case roughly a tenth. Sacra’s December 2025 read puts churn under 1%, strong for an SMB-adjacent insurance product. The praise is consistent: it works, and it is cheap.

The complaints. They are structural, and they come from the industry rather than the customers — which is what you would expect for a book of business too young to have been tested by claims. Vouch’s public attack (June 2026) is the sharpest: the RRG has no AM Best rating, no state guaranty fund behind it, cannot write anything but third-party liability, and may fail the insurance clauses in enterprise contracts and leases; RRG policies must by law carry a disclosure that they are not regulated like standard policies, and Vouch’s contention is that most founders never read it. Fintech Blueprint (2 June 2026) went further on the finance: $2.6B on $40M of premium is roughly 65x, and premium is not software revenue — most of it is meant to be paid back out as losses. It notes that both B rounds were led by substantially the same investors, that Corgi discloses no loss ratio, no combined ratio and no reinsurance retention, and that an exit below $500M would be a wipeout for common holders given the preference stack. Industry commentators have made the reserving point bluntly: Corgi wrote long-tail lines — D&O, E&O, cyber, EPLI, AI liability, where claims surface years after the policy sells — on barely a year of loss history, and only recently began recruiting the in-house actuary who would price them and set reserves. And the litigation record is now four cases deep in under a year. Whatever the merits, a company suing a competitor, a spinout by its own former staff, and a software platform simultaneously is spending founder attention it does not have to spare.

Outlook: the open question

For the bull case to be true, the 2025–2026 accident years have to develop inside Corgi’s pricing assumptions — because there is no rated balance sheet and no guaranty fund standing behind them if they do not. Everything else about this company is genuinely impressive and almost beside the point.

The bull case is coherent. Insurance is words; the words are now machine-readable; the traditional chain of broker, wholesaler, carrier and TPA is mostly a coordination tax. Corgi collapsed it and owns the customer, the quote, the paper and now the claims desk. If the loss ratio holds, it earns underwriting profit plus float, and selling its rails to other carriers converts a 3x revenue business into a software multiple. That is what TCV paid for.

The bear case lives entirely in the tail. An RRG is the cheapest way to get carrier economics without carrier capital, and that cheapness is why the objection has teeth: no rating, no guaranty fund, member-owned, reinsured by a captive in the same family. In a short-tail line you would learn quickly whether the pricing works. In D&O and tech E&O you learn in year three or four, by which point the book is 10x bigger. And in a soft management-liability market, Corgi’s low price may simply be the market price rather than an edge.

Three things to watch, none of them premium growth. Whether Corgi ever publishes a loss ratio or takes an AM Best rating — the moment it does, the argument ends one way or the other. Whether it retains customers into Series B and beyond, where enterprise contracts demand rated paper and Vouch’s objection stops being marketing. And whether the 2025-2026 reserves develop cleanly. Premium is the easiest number in insurance to grow and the last one you should trust.

Sources and further reading

• TechCrunch — Insurance startup Corgi hits $1.3B valuation 4 months after its Series A (Dominic-Madori Davis, May 6, 2026)
• TechCrunch — Corgi announces $106M raise at $2.6B valuation, double what it was worth 3 weeks ago (May 28, 2026)
• PR Newswire — Corgi Insurance Raises $108 Million, Receives Regulatory Approval to Launch the First Full-Stack Insurance Carrier for Startups (January 2026)
• Sacra — Corgi revenue, funding and business model breakdown (2026)
• Fintech Blueprint (Lex Sokolin / Luke Spill) — Analysis: How an AI Insurer reached a $2.6B Valuation on $40MM of Premium (June 2, 2026)
• Vouch — Vouch vs Corgi comparison page, the RRG objection in full (as of June 1, 2026)
• Insurance Business — Insurtech sues rival, alleges exec created sham firm to steal secrets (April 2026)
• The Insurer — Matcha accuses Corgi of bullying tactics in Wikipedia suit filed on same day as $160M raise (May 28, 2026)
• PR Newswire — Corgi Insurance Launches Corgi Claims, the AI-Native TPA (June 29, 2026)
• Aon — Management Liability Insurance Market in 2025: Stability Amid Evolving Risks (2025)