Guidewire Software said on Tuesday it will buy Hyperexponential, a London-based maker of AI-driven underwriting and pricing software for commercial insurers, for $1.7 billion in cash. The total rises to about $2.1 billion if $400 million in earnouts tied to revenue targets are paid in full. The deal is expected to close next year; Guidewire said it leaves the fiscal 2027 guidance it issued in September unchanged.
Hyperexponential was founded in 2017 and had raised close to $90 million, according to PitchBook, with Andreessen Horowitz and Battery Ventures among its investors — an outsized return for a company selling into one of software's slowest-moving verticals. It says nearly 600 insurers in 40 countries use its platform, which models and prices commercial property and casualty risk.
The strategic logic is about where the money sits in a carrier's stack. Guidewire sells the core: policy administration, claims, billing. Hyperexponential sells the pricing and actuarial layer where underwriters decide what a risk is worth. Insurers have historically bought those separately — a core system from one vendor, a pricing platform from another, spreadsheets and actuaries in between. Owning the pricing layer lets Guidewire pitch one stack in which agentic software drafts submissions, runs scenarios and prices risks next to a human underwriter who signs off.
That "next to a human" is both the sales pitch and the regulatory constraint. Underwriting is audited, filed with regulators and, in most markets, requires a named individual to stand behind a price. Vendors therefore sell AI as a tool that shortens the mechanical work — data entry, submission triage, scenario runs — rather than as an autonomous decision-maker. Hyperexponential's own positioning follows the same line: faster iteration on pricing models, with the underwriter still choosing.
The earnout is the tell about risk. Paying $400 million only if revenue targets are met shifts part of the bet onto Hyperexponential's growth, and keeps the headline multiple down until the combined product is proven to sell. Insurance carriers are cautious buyers with long procurement cycles, and consolidation in insurtech has been slow for exactly that reason.
For the wider enterprise-AI market, the deal is a data point that vertical software with real workflow lock-in still commands a premium. $1.7 billion in cash for a company that raised roughly $90 million is a strong outcome at a moment when many AI application startups are valued on growth rather than revenue — and a bet that the durable money in "agentic AI" sits in owning a regulated workflow, not in selling a model.
