There is an asymmetry forming at the top of the insurance value chain that almost no P&C board has yet named. Over the last two to three years, the largest reinsurers have built internal capability to price AI failure modes, packaged that capability into standalone products that underwrite AI model performance, and watched Lloyd's syndicates publish market commentary on the same set of questions. The reinsurance market has, in effect, built a working view on what AI looks like when it goes wrong.
Most cedents have not been part of that conversation. The CRO has not sat down with the reinsurance panel to walk through the ML in the rating stack, the treaty submission documents typically do not describe the model versioning regime that produced last year's loss ratios, and the renewal meeting last June covered exposure shifts and reinstatement premiums without a specific page on the rating model.
The reinsurer is forming an opinion about that cedent's AI anyway. The opinion gets formed in the absence of cedent-led disclosure, which is the worst possible footing for the cedent to be on when it eventually arrives at the renewal meeting.
The Capability Gap, in Concrete Terms
Munich Re is the cleanest example. Since around 2018, the company has been writing performance coverage on AI systems through what is now packaged as aiSure.1 The product underwrites the performance of a customer's AI model directly, agreeing to pay if the model misses a specified accuracy threshold or otherwise underperforms against a defined benchmark.
That product is interesting on its own. The more important point is what it required Munich Re to build internally before it could price the contract. To write coverage on AI performance, the underwriter has to maintain a working model of how AI fails: drift over time, distribution shift in the input data, degradation under adversarial inputs, the rate at which a model trained on 2022 claims experience becomes a poor description of 2026 claims experience. Pricing aiSure required Munich Re to build that internal view across many lines and many model classes, and that view does not stay siloed inside the aiSure product. It informs how the same Munich Re underwriting team thinks about every cedent that comes to them with ML in its rating stack.
A cedent placing a catastrophe XL treaty with Munich Re today is being evaluated by a counterparty that has spent roughly eight years studying how AI fails. Most cedents on the other side of that table have spent considerably less time on the same question, and the gap shows up in how prepared each side is for the conversation about model behavior.
Swiss Re Has Joined the Build
On December 10, 2025, at Abu Dhabi Finance Week, Swiss Re and RIQ signed a Memorandum of Understanding to develop and scale capacity solutions, risk origination opportunities, and AI-enabled capabilities across the UAE.2 RIQ is an AI-native reinsurance platform launched in June 2025 by Abu Dhabi-based IHC together with BlackRock and Lunate, headquartered in the Abu Dhabi Global Market and built around AI-first underwriting infrastructure. The collaboration gives Swiss Re a structured on-ramp into AI-native reinsurance origination, with the platform's analytics being developed alongside Swiss Re's risk expertise rather than retrofitted onto a legacy stack.
The strategic read is straightforward. The world's second-largest reinsurer has decided that AI-native underwriting infrastructure belongs alongside its sheet by 2026, and it has chosen partnership with a purpose-built platform over the slower path of building organically. The MoU is geographically scoped to the UAE today, but the underwriting science Swiss Re is co-developing through it does not stay in one jurisdiction once it is in the firm's analytical bloodstream. For a US cedent placing treaty business with Swiss Re inside the next two renewal cycles, the practical consequence is the same one that exists with Munich Re. The underwriter on the other side of the table has, or is about to have, the analytical machinery to evaluate the AI in the cedent's pricing stack, and whether the cedent walks the underwriter through that AI proactively will determine whose framing the renewal terms reflect.
Lloyd's Has Published Its Position
The London market has been less vocal but no less attentive. The Lloyd's Market Association, with Oxbow Partners, published The Growth of Enhanced Underwriting in the Lloyd's Market: The New Normal? on November 18, 2024, drawing on interviews and a survey of carriers, brokers, MGAs, and capital providers representing 77% of Lloyd's 2023 GWP. The report identifies the specific failure modes the market wanted addressed before algorithmic underwriting could scale further. The top three barriers to adopting algorithmic underwriting cited by survey respondents were loss of control of underwriting decisions (14%), algorithmic bias (12%), and regulatory concerns (10%) — putting model bias second on the list and regulatory uncertainty third.3
The report matters less for the specific rankings and more for what its existence demonstrates. The Lloyd's market does not commission and publish enhanced-underwriting position research until enough of its capital providers have reached internal alignment that the question is worth speaking to externally. The November 2024 publication is evidence that Lloyd's syndicates have been forming a view on algorithmic underwriting for some time, well before the report appeared.
A US carrier with a London-syndicated layer in its catastrophe program is therefore facing, on the London end, the same dynamic that exists with the continental European reinsurers. The underwriting view on AI is being formed by the counterparty regardless of whether the cedent participates. The choice the cedent makes at every renewal is whether to shape that view or discover it after the fact.
The Adjacent Signal: Rating Pressure Is Already Moving
There is a related data point worth holding alongside the reinsurance signal. AM Best took 55 downgrade actions across the US P&C sector in 2023, against 35 upgrades in the same year and 30 downgrades the year prior.4 The cited drivers were reserve adequacy, catastrophe loss volatility, and reinsurance dependency rather than AI specifically, with reinsurance dependency typically operating as the chain of transmission from a tightening reinsurance market to a tightening rating environment.
The reason it matters is that the rating side of the carrier capital structure is tightening alongside the reinsurance side. A cedent whose treaty terms are being repriced at renewal and whose AM Best outlook is being reviewed in the same quarter is operating with simultaneous pressure from two of the three counterparties that decide whether the company is solvent on a forward basis. Rating action tends to follow treaty action with a lag, which is the reason the two signals correlate.
We covered the rating-side mechanics in detail in the canonical post on why reinsurance, not regulation, is the existential AI risk for insurers. The reinsurance signal sits inside a broader tightening of the capital framework around US P&C, and the carriers most exposed to AI-related repricing are the ones that have not yet built a defensible disclosure posture toward either counterparty.
What the Cedent's Silence Costs
Consider the structure of the renewal conversation a CRO at a $400M regional homeowners carrier is going to have with a Munich Re or Swiss Re underwriter in June 2027.
The reinsurance underwriter has read the loss triangles, the cat model output, the schedule of in-force exposures, and the reinstatement structure. Separately, and without telling the cedent they were doing it, the same underwriter has formed an internal view on the AI in the cedent's rating stack. That view was assembled from public filings, the cedent's own prior treaty submissions, conference presentations the cedent's CTO gave, vendor disclosures from the cedent's pricing platform provider, and the underwriter's own pattern recognition across other carriers in the same segment.
If the cedent has not led the conversation, the reinsurer's internal view fills the vacuum. That view is the one the underwriter will reference when they write the renewal pricing memo, when they decide whether to add a disclosure warranty rider, and when they evaluate whether the cedent's representation about its underwriting practices is materially complete. The underwriter is not hostile in this scenario. They are simply doing their job with the information they have, and the cedent's failure to put more accurate information on the table means the underwriter is working from a less generous picture than the cedent could have provided.
The cost of that asymmetry shows up in three places. Treaty pricing comes in higher than the cedent's loss experience would justify, because the underwriter is loading for AI uncertainty the cedent could have priced down with a credible disclosure document. New disclosure warranties or model-change notification provisions appear in the wording, with retroactive bite. At the worst end of the distribution, the renewal becomes a non-renewal, with the reinsurer citing a change in underwriting philosophy the cedent never formally communicated. We unpacked the contractual mechanics of that scenario in the companion post on the disclosure clause your AI strategy lives or dies on.
The Question Worth Asking This Quarter
There is a concept worth naming clearly. The default posture of a US P&C cedent today is what we would call reinsurer-led disclosure: the reinsurer is forming a view on the cedent's AI without any input from the cedent, and that view becomes the basis for renewal decisions. The available alternative is cedent-led disclosure, in which the cedent presents its own view of its AI to the panel and the underwriter responds to that documentation. Almost every US P&C carrier is currently in the first posture, and the strategic move available before the next renewal is to move into the second.
The CEO question that operationalizes the move is short. Does our reinsurer's view of the AI in our pricing stack match ours, and have we asked?
Answering it requires the carrier to do three things most carriers have not done. First, build an internal description of what the rating model is, how it changes, and what the change history looks like, written in language a treaty underwriter can read. Second, sit down with each reinsurer on the panel and walk through that description in advance of the next placement, so the underwriter's internal view is being shaped by the cedent's documentation rather than by inference. Third, capture the disclosure as a written record on both sides, so the conversation is reproducible at the next renewal and at the renewal after that.
None of these three steps is a regulatory deliverable, and no state insurance department is currently asking for them. They are being built for an audience of one: the underwriter on the other side of the treaty placement, who is going to form a view on the cedent's AI either way, and whose view the cedent has the option of shaping rather than discovering after the fact.
The reinsurance market has built the capability to underwrite AI directly, and that capability will be applied to every cedent the market accepts as a counterparty regardless of whether the cedent participates in the analysis. The cedents that walk their underwriters through model documentation in advance of renewal end up with pricing and wording that reflect their own framing alongside the reinsurer's. Silence produces a different result. Where the cedent has not led the disclosure, the analytical capability gets applied through inference from external sources, and the renewal memo lands with framing that the cedent typically discovers only after it has already hardened.
The window to choose which posture you operate inside is open right now, and it tends to close at the renewal meeting where the question first becomes contested rather than hypothetical.
Related
- The Reinsurance Treaty Is Where AI Risk Becomes Existential
- The One Treaty Clause Your AI Strategy Lives or Dies On
- Why AM Best's "Comprehensive Adjustment" Is the Hidden AI Downgrade Lever
- The NAIC Bulletin Is the Floor Your Reinsurer Will Hold You To
Footnotes
-
"aiSure™ is a suite of comprehensive coverage for AI systems designed to address a wide area of AI-related risks for AI providers and corporate adopters caused by AI performance errors… aiSure is designed to reflect the probabilistic nature of AI, where even well-constructed models can produce incorrect outputs." — Munich Re: aiSure™ — More AI Opportunity. Less AI Risk ↩
-
"Swiss Re and RIQ have signed a Memorandum of Understanding (MoU) to develop and scale innovative capacity solutions, risk origination opportunities and AI-enabled capabilities across the UAE… The MoU was signed at Abu Dhabi Finance Week 2025 in the presence of His Excellency Dr Sultan Ahmed Al Jaber, UAE Minister of Industry and Advanced Technology, and Chairman of RIQ, by Mark Wilson, Chief Executive Officer at RIQ, and Andreas Berger, Group Chief Executive Officer at Swiss Re." — Swiss Re: Swiss Re and RIQ partner to advance risk transfer powered by data and AI (December 10, 2025) ↩
-
"In the LMA's report, 'The Growth of Enhanced Underwriting in the Lloyd's Market: The New Normal?', published in November 2024, barriers to adopting algorithmic underwriting were identified by survey respondents as follows: loss of control of underwriting decisions (14%), algorithmic bias (12%) and regulatory concerns (10%)." — Lloyd's Market Association: Over one-third of London market firms now actively using AI and LMA: The Growth of Enhanced Underwriting in the Lloyd's Market — The New Normal? (November 18, 2024) ↩
-
"AM Best's 'US Property/Casualty Downgrades Outpace Upgrades in 2023' report noted there were 55 downgrades last year — a higher total compared to upgrades (35) in 2023, as well as to the number of downgrades (30) in 2022." — Insurance Journal: Twice as Many Personal Lines Insurers Downgraded by AM Best in 2023 (April 2024) ↩
