BRENNEN

Industries / Finance and business services / Insurance

Reinsurance

Reinsurers insure insurance companies, taking on part of their largest or least predictable risks in return for part of the premium.

1846

Cologne Re was founded in Germany. Gen Re, which grew from it, calls it "the world's first independent professional reinsurance company" and says the first professional reinsurance treaty was signed in 1852.

genre.com

34

Milestones achieved across 6 domains.

7

To watch for: announced, not yet confirmed outside.

12

Open deliverables. Nobody has shipped them.

In one minute

Where it started

Companies that did nothing but reinsurance began in Germany in 1846, and in 1877 a Lloyd's underwriter in London began reinsuring fire policies. Most cover is bought through a treaty, a standing agreement that covers a whole class of an insurer's policies for a year, rather than one policy at a time. Failures taught the trade to watch what it passes on: in the 1980s London reinsurers passed the same losses round and round among themselves until claims ran to many times the original loss.

Where it stands

Reinsurers price disasters with computer models, and since the 1990s they have shared risk with investors through catastrophe bonds and similar deals. The broker Aon counted a record $136 billion of such investor capital in 2025, out of about $785 billion of reinsurance capital in all. US catastrophe reinsurance prices fell 16% in 2026 and stood 22% below their 2024 peak, by the broker Guy Carpenter's index. Storms, wildfires and floods, which the first models were not built for, caused 92% of insured disaster losses in 2025, by Swiss Re's count.

What is open

Much of the work still runs on documents that people read one at a time: treaty wordings, submissions from insurers, statements of account. Most catastrophe models are closed, so few outsiders can check the numbers on which prices and bond payouts rest. The open jobs below aim to make those documents and numbers checkable with tools that exist now: open models, public filings and AI models that read documents.

Domains

AchievedWatch forOpen

Advances to build with

1 research note

Open in reinsurance

3 of 12 open deliverables

Open · Treaties and wording

A public library of treaties

A free, searchable collection of the reinsurance contracts that US insurers file with the Securities and Exchange Commission (SEC) as exhibits to their annual reports. Each contract's key terms (retention, limit, reinstatements, hours clause, exclusions) go into a table, and a test scores how well AI models pull out those terms.

Built with what exists

SEC EDGAR, which is free to search; Python; a current AI model, or an open-weight model such as Mistral Large 4 run on one's own computer (Note: "A trillion open parameters"); a second person checking the labels.

Credit · unclaimed

Open · Pricing

From submission to price

A public test of whether AI agents can price a reinsurance layer as a junior analyst would. Each task gives a synthetic submission (loss history, exposure summary, proposed terms); the agent must apply burning cost, the average of past losses that would have hit the layer, and exposure rating, and return a price with its working. Every answer is computed in advance by open actuarial code.

Built with what exists

GEMAct, a Python package for (re)insurance costing described in the Annals of Actuarial Science; reinspy from the Casualty Actuarial Society; the mbbefd R package for exposure curves; synthetic submissions generated by code; any current AI model, or Mistral Large 4 (Note: "A trillion open parameters").

Credit · unclaimed

Open · Catastrophe models

An open flood model for Thailand

An open-source catastrophe model for river flooding of Thailand's industrial estates, run in the open Oasis Loss Modelling Framework, that anyone can inspect and rerun.

Built with what exists

the Oasis Loss Modelling Framework (BSD-3-Clause licence); the European Commission Joint Research Centre's global river flood hazard maps, which give water depth for floods from 1 in 10 to 1 in 500 years; OpenStreetMap outlines of the estates; the JRC's global depth-damage curves; Python.

Credit · unclaimed

Where it sits