CyberCube

CyberCube is the leading AI-native cyber risk analytics and accumulation-modelling platform for the global (re)insurance industry. The company was incubated inside Symantec in 2015 under then-General Manager Pascal Millaire, then spun out as a standalone, VC-backed business in March 2018 with a $15M Series A led by Forgepoint Capital. Cumulative equity funding now exceeds $285M across Series A through growth rounds, culminating in an October 2025 $180M+ cornerstone investment from Spectrum Equity, joining existing investors Forgepoint Capital, Hudson Structured Capital Management and MTech Capital. Former Verisk chairman and CEO Scott G. Stephenson chairs the board.

Commercial model: data and models, not capacity. This is the critical differentiation against Coalition and At-Bay. Coalition and At-Bay are cyber MGAs — they distribute policies, automate underwriting decisions and depend on reinsurer capacity (Allianz, Swiss Re, Arch, Lloyd's syndicates) to absorb loss. CyberCube takes no underwriting risk. It licenses software, models and data to (re)insurers, reinsurers and brokers, who use it to price, accumulate, reserve, reinsure and securitise cyber risk on their own balance sheets. Closer analogues are Moody's RMS and Verisk — third-party cat-modelling infrastructure — not the MGAs that consume that infrastructure.

Product suite. Three flagship products anchor the platform. Portfolio Manager is a scenario-based stochastic cyber catastrophe model used by (re)insurers and ILS investors to quantify accumulation risk and tail loss across books of cyber business; PMv6 shipped in July 2025 with an expanded Enterprise Intelligence Layer covering Germany, France, Australia, Spain, Canada, UK and Japan. Account Manager scores individual-account cyber risk for underwriters at point of quote. Exposure Manager, launched in 2025, is a portfolio-wide audit and benchmarking tool letting (re)insurers evaluate cyber exposure across the whole book independently of underwriting systems. Underneath sits a proprietary data layer assembled from cybersecurity telemetry partners (originally Symantec's threat data), firmographic and technographic enrichment, and loss/claims data contributed by clients.

Traction. CyberCube states its solutions are used by 75% of the top 40 U.S. and European cyber insurers by gross written premium, and by a majority of the top 20 global brokers. The client roster includes Munich Re (which co-publishes systemic cyber research with CyberCube), Guy Carpenter (formal licensee of the stochastic cat model since 2018), Hannover Re (whose Cumulus Re 2025-1 cloud-outage cat bond uses CyberCube as third-party modeller), Chubb, Gallagher, Aon, MAPFRE RE, and 130+ total institutions. Portfolio Manager has been the third-party model behind the majority of cyber catastrophe bonds issued since Beazley's landmark 2023 issuance — Artemis routinely refers to CyberCube as "the cyber cat bond modeller."

Where it sits in old → new → AI. Cyber is structurally different from nat-cat: no 30-year loss history, an adversarial rather than physical hazard, and dependency structures (cloud providers, shared software stacks) that shift faster than the annual model release cycle of the incumbents. This is why cyber modelling is one of the few cat-modelling categories where an AI-native entrant (CyberCube) has out-penetrated the incumbents (Moody's RMS, Verisk) rather than being rolled up by them. The category tailwind is the maturing cyber ILS market and regulators (PRA, NAIC) increasingly pressing (re)insurers to demonstrate defensible cyber accumulation views — both push demand for a third-party, auditable model that a carrier's internal team cannot produce credibly on its own. The risk on the other side is commoditisation: if Verisk and Moody's close the gap on cyber models, or if large reinsurers accept in-house views as sufficient, CyberCube's pricing power erodes. The 2025 Spectrum Equity round is positioned against that window — scaling the data moat and the broker/ILS footprint before the incumbents catch up.