What it covers
- GenAI output errors and hallucinations causing financial loss
- IP infringement and defamation from AI generated content
- Unauthorised data disclosure in AI outputs
- Bodily injury from reliance on AI output
- Property damage from AI driven decisions
- Regulatory violations, within the form
Watch out for
- Excludes companies selling AI products
- Excludes standard cyber attacks and intentional misuse
- Defence costs erode the limit
- Published capacity differs across 2026 sources ($9.25M in March 2026 materials, $10M aggregate in later trackers), so the current limit must be confirmed at submission
Who it fits
- Non AI companies deploying GenAI features or agents at scale
- SaaS, fintech, and healthtech deployers facing enterprise procurement questions
Who it does not fit
- AI model builders selling AI as the product
Frequently asked questions
Who distributes Testudo?
Brokers. It is the same product that Baldwin's Innovation Practice (the former FounderShield) distributes, which means it is a market any appointed broker can reach, not a platform exclusive.
What revenue size does it target?
The distributor materials describe US companies up to $10B in revenue that deploy GenAI, so this is not a startups only product.
Source: Program details from Testudo and distributor published materials, as compiled in our AI risk research checked November 14, 2026. Limit figures conflict across sources and are flagged on this page.
Talk through your stack before you sign the next contract
Send us the term sheet, the customer MSA, or the renewal you are worried about. We will tell you plainly what it requires, what your current programme does about it, and what we would place differently.