Insurance · Deep dive

Honeycomb Insurance

The habitational-property MGA that never sends an inspector — aerial imagery and computer vision underwrite condo, HOA and small-multifamily buildings the admitted market keeps mispricing.

emerging

The question that decides it: As Honeycomb pushes past $275M of GWP into non-admitted E&S risks — older buildings, lapsed coverage, sub-par roofs — does the no-inspection AI loss ratio hold through a full underwriting cycle, or does the reinsurance panel that actually bears the risk force Honeycomb to send inspectors and give back the cost edge that is the entire pitch?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Honeycomb is a US managing general agent writing commercial property insurance for condo and homeowner associations, landlords and small-to-mid multifamily buildings. It underwrites remotely — aerial imagery, computer vision, hundreds of geospatial and building-level data points, no human site visit — and cedes the risk to reinsurers, with SiriusPoint as its lead capacity provider since 2022. First policy bound June 2021; exited 2025 at roughly $275M of annualised gross written premium across 20 states and about 30,000 customers, up 70% YoY (company disclosure, June 2026). A $40M Series B extension in June 2026 (Zeev Ventures) took total funding to about $95M.

Founding story

Itai Ben-Zaken is a Wharton MBA and Hebrew University computer engineer, ex-BCG, who ran QuinStreet’s digital insurance division — the parent of Insurance.com and CarInsurance.com. That was a lead-generation business, and it taught him where the money actually sits: not in the click, but in the underwriting decision the carrier makes after it. Between QuinStreet and Honeycomb he ran Comprendi, a digital ad-recommendation start-up he wound down in 2018.

He and co-founder Adi Ben-Zaken started Honeycomb in 2019 aimed at a specific slice of the market: habitational commercial property — condo associations, HOAs, small landlord portfolios, the two-to-fifty-unit multifamily buildings that sit between homeowners and true middle-market real estate. His public diagnosis was mundane: the segment was underwritten by broad territory averages and static NAIC-code rating factors, quotes took weeks, and physical inspections cost hundreds of dollars per building. The team spent two years, largely in Tel Aviv, building the computer-vision stack before writing anything. First US policy bound in June 2021, in Illinois.

How it works

Honeycomb is not the insurance carrier. It is an MGA — a licensed producer that handles underwriting, distribution, policy administration and first-notice-of-loss on behalf of an admitted or non-admitted carrier that owns the paper, with the underwriting risk itself ceded to a panel of reinsurers. SiriusPoint has been the lead capacity provider since 2022 and is also an equity investor via its Series A participation. That structure means Honeycomb’s P&L is fee-based — MGA commission on the premium it writes plus profit commission if the book runs well — not a spread on premium and losses. The reinsurers own the loss ratio outcome; Honeycomb owns the loss ratio expectation.

The technology is what has to work. When a broker runs a quote, Honeycomb ingests hundreds of signals per building: aerial and satellite imagery (rooftop condition, geometry, footprint, adjacent vegetation), street-level photography, parcel and permit records, geospatial hazard layers, and historical loss data from public sources and prior policies. Computer-vision models estimate roof age, material and damage; other models score fire and water risk against the property’s specific characteristics rather than a ZIP-code composite. Output: a bindable price in minutes, and — critically — the underwriting file is built without dispatching an inspector. Ben-Zaken told Fortune in June 2026 that on well-maintained buildings, Honeycomb quotes up to 40% below the incumbent.

The trade-off is structural. A carrier that never sends a human relies on the model to spot the deferred-maintenance building the applicant did not disclose. When the model is right, Honeycomb’s expense ratio undercuts any traditional habitational underwriter. When it is wrong, the loss lands on the reinsurance panel — which is why the panel’s willingness to keep providing capacity at Honeycomb’s target price is the mechanism that decides whether the business scales.

Product and business overview

• Condo & HOA / COA insurance. Master policy for condo and homeowner associations — building coverage, common-area liability, D&O for the board. Founding product; the segment where the incumbents’ NAIC-code broad-brush is widest.
• Landlord and small-multifamily habitational. Single-building landlord policies and small-portfolio multifamily, with per-risk TIV limits Honeycomb has been raising as the book seasons.
• Honeycomb Specialty (non-admitted / E&S). Launched in 2024 for risks outside admitted appetite: non-standard electrical panels, below-average roofs, lapsed coverage, older frame construction. July 2025: rolled to every state Honeycomb writes; TIV limits lifted from $10M to $25M per risk.
• Broker-facing agent portal. Distribution is dominantly retail brokers, not DTC. The portal is the actual product for the broker: fast quote, bindable output, appetite transparency.

Business model and pricing

Revenue comes in three layers: MGA commission (a percentage of GWP written on behalf of the carrier partner), profit commission (a share of the underwriting margin when the book beats loss-ratio thresholds), and policy fees where regulation allows. Honeycomb does not disclose its schedule; specialty habitational MGAs in this segment typically retain a mid-teens percentage of premium plus a profit share on top.

Honeycomb does not publish per-policy pricing. The public pricing anchor is the “up to 40% cheaper for well-maintained buildings” claim Ben-Zaken has repeated to trade press (Fortune, June 2026; Insurance Business America, 2024) — worth reading as directional, not as a rate table. At $275M of annualised GWP exiting 2025, a mid-teens fee take would imply GAAP revenue in the $35-50M zone plus profit commission on the reinsurance treaties, split undisclosed.

Traction over time

Market analysis

The US commercial habitational property market — HOAs, condo associations, landlord and small-multifamily — is highly fragmented. Travelers, the largest writer in the landlord/condo association line, reportedly holds roughly 7% share; the balance sits across roughly a hundred regional and specialty carriers. Segment premium runs into the low tens of billions annually; Honeycomb’s launch materials cited $22B of multifamily property in play. The structural forces are the same ones moving the rest of US property: hardening reinsurance, retreating admitted carriers on older or coastal risks, rising litigation cost in weather-exposed states, and a widening gap between what buildings actually cost to insure and what NAIC-code averaged rates say they should. That gap is Honeycomb’s opportunity — and the reason three well-funded MGAs (Honeycomb, Steadily, Obie) are chasing it in parallel.

Notably absent from Honeycomb’s map: Florida. Focusing on inland, lower-cat states deliberately excludes the market that has consumed most insurtech obituaries — Kin, Slide, HCI — and keeps the reinsurance panel’s cat exposure containable. Whether Honeycomb holds that discipline while chasing scale is a separate question.

Competitive intel

The set breaks into three layers, all in the sidebar. Insurtech peers — Steadily, Obie, CoverTree — pitch the same broker on the same “modern underwriting” narrative and are broadly one funding stage behind. Traditional admitted carriers — Travelers, Nationwide, Liberty Mutual, plus regional habitational specialists — still hold the incumbency and the balance sheet. The public E&S comp — Kinsale — is where Honeycomb Specialty is starting to look. The read: Honeycomb’s edge is not that it invented AI underwriting for buildings; it is that it built the operational stack (broker portal, remote-inspection ML, reinsurance relationships) to attack admitted and E&S off one platform. That’s a distribution and integration moat, not an unassailable technology one.

History and evolution

• 2019 — Founded by Itai and Adi Ben-Zaken. Two years of quiet R&D in Tel Aviv building the computer-vision stack.
• June 2021 — Public launch: $3.3M seed and first US policy in Illinois (initial states IL, AZ, MI, OH).
• January 2022 — $15.4M Series A led by Ibex; SiriusPoint joins as investor and lead capacity provider — the critical carrier relationship that lets the MGA actually write paper.
• 2022-23 — Deliberate state expansion. Ben-Zaken publicly framed the pace as choosing loss-ratio evidence over premium growth so reinsurers would renew capacity.
• May 2024 — $36M Series B (Zeev Ventures). Revenue reported more than doubled YoY. Honeycomb Specialty (non-admitted / E&S) launches.
• July 2025 — Non-admitted product rolled to every state Honeycomb writes; per-risk TIV limit raised from $10M to $25M. New Chicago HQ opens.
• August 2025 — Minnesota launch announced as first of five planned expansion states (with Colorado, Oregon, Kansas, Missouri).
• June 2026 — $40M Series B extension (Zeev) at a higher price than 2024; total funding ~$95M. Fortune: 30,000 customers (+70% YoY), $275M annualised GWP.

What people say

The case for. Trustpilot runs around 4.7/5 across roughly 165 reviews (August 2026), with the recurring theme the buying experience: brokers and small-multifamily owners describe a quote-and-bind flow that returns a real price in minutes on properties incumbents wouldn’t quote in weeks. Trade sites like SimplyInsurance and LogicalLandlord praise the pricing on well-maintained buildings and the willingness to write older frame construction the admitted market surcharges. On the capital side, SiriusPoint has stayed on paper four straight years, Ibex has added money in every round since 2022, and Zeev led twice — 2024 and 2026 at a higher price — the market vote that the loss ratio behind the growth is real, at least as reinsurers can currently measure it.

The complaints. The recurring complaint themes on the BBB Honeycomb Programs profile and on Trustpilot’s lower-star reviews are what you’d expect from a fast-scaling habitational MGA: claims handoff to third-party adjusters with poor communication and long resolution timelines; renewal shock in the form of premium jumps and mid-term policy changes; and — the most operationally telling — “last-minute declination of quotes,” where a bindable-looking price is pulled once an underwriter (or the model) surfaces a red flag late in the process. Multiple BBB narratives describe cancellation notices issued over minor infractions or renewal-process failures, with the company refunding after complaint but leaving the policyholder unbound in the meantime. Not unusual for a property MGA growing 70% a year; a franchise problem if it keeps growing at the same rate on top of a bigger base.

The two sides say what they usually say in scaling MGAs: the front end is better than the incumbents; the claims machine has not caught up.

Outlook: the open question

The answer resolves in the reinsurance treaty renewals of 2027 and 2028, not in the marketing. Honeycomb has done the hard part first: built a computer-vision underwriting stack, convinced a rated carrier (SiriusPoint) to put paper behind it, and grown to $275M of GWP in segments the admitted market treats as an afterthought. An MGA that never sends an inspector is structurally cheaper than a traditional habitational carrier — and Zeev doubling down at a higher price in June 2026, plus Ibex following every round, suggests the private underwriting numbers are defensible.

The mechanism is the reinsurance panel. Honeycomb does not carry its own losses. If the admitted-book loss ratio creeps as it scales, or — more likely — if the new Honeycomb Specialty prices deferred-maintenance E&S risks too aggressively, the panel will re-rate the treaty. Two failure modes: reinsurers force Honeycomb to (a) send inspectors on a growing share of risks, giving back the cost edge that is the whole pitch, or (b) cut the target loss ratio, pushing premiums up and handing the segment back to Travelers and the regionals. What has to be true for the model to work at scale: no-inspection underwriting holds within a few points of the incumbent loss ratio through a full weather cycle, and Specialty proves out its own loss ratio on E&S risks fast enough that reinsurers keep supplying capacity for the more profitable admitted book. What has to be true for it to break: a bad two-year loss run on Specialty forces the panel to reprice both books together, leaving Honeycomb either running a smaller admitted-only book or subsidising growth with the equity it just raised.

The claims-handoff complaints are the near-term tell. The reinsurance treaty conversations are the long-term one.

How a challenger would attack it

Attack the back office and the bind reliability. Honeycomb’s documented failure modes are precise: claims handed to third-party adjusters with poor communication and long timelines, renewal shock with mid-term policy changes, and “last-minute declination of quotes” — a bindable-looking price pulled when the model surfaces a late red flag. For a broker, an unreliable bind is worse than a slow one; a challenger that guarantees its quotes (price-lock at bind, no post-quote pull) and runs claims in-house wins the brokers Honeycomb’s growth is straining, because the broker eats the E&O exposure when a client is left unbound. The second wedge is the technology’s replicability: aerial imagery, roof-condition CV and geospatial hazard layers are now available from vendors (the same stack Steadily and Obie draw on), so the moat is the reinsurance relationship, not the model — and that’s attackable by a challenger with a stronger balance-sheet story, such as a hybrid carrier retaining a slice of its own risk, which signals skin-in-the-game SiriusPoint’s fee-based MGA cannot. Third, hit the E&S flank early: Honeycomb Specialty is pricing deferred-maintenance buildings on no-inspection AI before that loss ratio has seasoned; a rival that pairs remote underwriting with cheap drone-based physical verification on exactly those risks can quote nearly as fast, price more accurately, and let Honeycomb adverse-select itself on the buildings its model can’t see inside.

Same playbook, new buyer

Point the no-inspection stack at habitational niches and geographies Honeycomb has deliberately fenced off. The obvious first: coastal and cat-exposed markets. Honeycomb’s inland-only discipline is rational for its reinsurance panel, but it leaves the hardest-hit, worst-served condo and HOA buyers — the Gulf and Southeast associations facing non-renewal — with no AI-underwriting option at all; a challenger built cat-first, with parametric wind riders and a panel priced for it, faces no Honeycomb competition by Honeycomb’s own design. Second, the asset classes adjacent to its models: CoverTree proved manufactured housing works as a standalone niche; student housing, senior living and mixed-use small commercial share the same aerial-imagery underwriting surface with even fewer specialist writers. Third, export the model: the two-country structure (US distribution, Tel Aviv ML) is replicable for the UK and EU block-of-flats markets, where post-Grenfell remediation has made habitational cover scarce and expensive, and where no US-focused MGA — least of all one whose 20-state expansion is still unfinished — will arrive this decade. Honeycomb can’t follow into any of these quickly: every new risk class or cat zone requires renegotiating the reinsurance treaty that is already the binding constraint on its core book.

Sources and further reading

• Fortune — Investors bet $40M on Honeycomb’s no-inspection landlord insurance AI (June 4, 2026)
• FinTech Global — Honeycomb Insurance raises $40M in Zeev Ventures-led round (June 5, 2026)
• TechCrunch — Honeycomb Insurance grabs $36M Series B from solo VC-led Zeev Ventures (May 7, 2024)
• Insurance Journal — MGA Honeycomb launches with bet on multi-family property insurance (June 22, 2021)
• PR Newswire — Honeycomb $15.4M Series A led by Ibex Investors (January 20, 2022)
• SiriusPoint — SiriusPoint invests in Honeycomb and becomes lead underwriting capacity provider (2022)
• The Insurer — Honeycomb expands non-admitted program to 18 states; ups TIV limit to $25M (July 16, 2025)
• Insurance Journal — Honeycomb launches in Minnesota, first of five planned expansion states (August 21, 2025)
• Insurance Business America — The insurer that never sends an inspector (2024)
• BBB — Honeycomb Programs Inc. customer complaints profile (accessed August 2026)
• Trustpilot — Honeycomb Insurance customer reviews (accessed August 2026)