AI Is Reshaping What Gets Built, What Gets Covered, and How Deals Get Done
AI is reshaping the insurance industry from multiple angles this week, as underwriters confront new physical infrastructure risks tied to AI data centers, the market for dedicated AI liability coverage begins to take shape, and wholesale distributors move to embed AI directly into specialty placement workflows.
54 of 100 — neutral. New property exposure from AI data centers is real and not yet well modeled, but a maturing liability market opportunity and wholesale distribution gains from AI adoption keep this week roughly balanced between emerging risk and emerging opportunity.
The Infrastructure Behind AI Is Becoming Insurance’s Next Big Headache
The physical backbone of the artificial intelligence economy, namely the sprawling data centers being built to process AI workloads, is creating a class of insurance challenges that the industry has not yet fully resolved. A new report from Swiss Re Institute puts the scale of the opportunity and the problem in sharp relief: global premiums tied to data center coverage are projected to more than double, reaching $24.2 billion by 2030, up from $10.6 billion today. Getting there will require underwriters to develop new technical fluency quickly.
Capital spending by the five largest cloud service providers is widely forecast to exceed $600 billion in 2026, with roughly 75% of that figure tied directly to physical AI infrastructure. The construction costs alone for individual facilities can reach $20 billion before equipment is installed, which drives insurers to cover full replacement values even when maximum probable loss scenarios are considerably lower.
The accumulation challenge is compounding the underwriting difficulty. Large data centers are often presented to insurers through separate programs for buildings, equipment, and power plants, making it difficult for carriers to track total capacity exposure. A single physical event could trigger claims across multiple insurance programs simultaneously.
Fire and water present the most significant loss drivers, and emerging technologies are intensifying both exposures. While fire accounts for only about 11% of data center loss events, it drives more than 42% of loss costs. A key emerging concern is the integration of lithium-ion battery backup units into server racks, creating an ignition source that did not previously exist within data processing equipment rooms. On the water side, liquid cooling systems, adopted to manage the significantly higher heat output of modern GPUs, present growing exposure, with liquid-related losses representing nearly 24% of total data center loss costs.
FM’s 2026 loss prevention guidance has responded by increasing recommended fire-resistance wall ratings from one to two hours and introducing more stringent sprinkler expectations.
For risk managers and property underwriters, the message from Swiss Re is clear: passive risk transfer is no longer sufficient. Engaging earlier in design, siting, and power decisions will be essential, as will greater transparency on underlying exposures. With few next-generation facilities fully operational, empirical loss data remains thin, making specialized technical assessment critical to sound underwriting. Practitioners who develop expertise in this space now will be well positioned as the market scales over the next several years.
— R&I Editorial Team at Risk & Insurance
The Cyber Playbook, Revisited: How AI Liability Could Build Its Own Market
Practitioners who watched cyber insurance evolve from a niche endorsement into a multi-billion-dollar standalone line may be watching a similar story unfold with AI liability. That is the view of Andrew Kelly, executive vice president at AJ Wayne & Associates, an E&S lines wholesaler and managing general agent, who sees AI exposure following a familiar arc.
“It would not surprise me if, over the next five to ten years, we see an AI insurance sector develop in the same way cyber did,” Kelly said. He expects that sector to eventually include managing general agents, claims professionals, and specific forms and endorsements designed to address AI exposure, technology exposure, and automation exposure.
For now, the market is handling AI risk indirectly. Many policies address artificial intelligence only indirectly, often through endorsements or silence within existing policy language. That approach carries familiar risks: as with early silent cyber, the gaps between what is covered and what is assumed to be covered can be significant, and those gaps typically surface at claim time.
The analogy to cyber’s early days is instructive in another way. Generative AI-related lawsuits in the U.S. grew 978% between 2021 and 2025, yet standard insurance policies across cyber, tech E&O, product liability, and commercial general liability each leave significant gaps in coverage, according to a Gallagher Re report produced with MIT. Courts are generally treating AI as a tool, placing liability on the organizations deploying it, while vendor contracts typically cap liability at 12 months of fees with no performance warranties.
That liability structure means the exposure lands squarely on the insured, making policy language review and gap analysis a priority for brokers today. Advanced data aggregation is already allowing carriers to identify loss trends within specific sectors rather than exiting industries entirely when results deteriorate, enabling a more focused approach to AI-related risk.
Kelly’s outlook for E&S brokers is constructive. The specialty market has historically been the proving ground for emerging risk lines, and AI liability may follow that same path. The brokers and underwriters who invest now in understanding AI exposure, building relationships with early-mover carriers, and developing clear submission narratives will have a meaningful head start when this market matures.
— Alicja Grzadkowska at Insurance Business Magazine
Placement, Accelerated: CRC Embeds AI Directly Into the Specialty Workflow
While much of the industry conversation around AI focuses on long-term structural change, CRC Group is moving to embed the technology into day-to-day specialty placement workflows right now. The independent wholesale distributor has launched REDY INTEL, an AI-powered engine built into its existing REDY platform, designed to translate submission data into real-time, actionable intelligence across every stage of the placement process.
REDY INTEL has been designed to transform how specialty risks are analyzed, positioned, and placed by turning data into real-time, actionable insight across every workflow, from initial data ingestion and enrichment to embedded analytics. The practical implications span three areas: risk insights, market insights, and placement insights.
On the risk side, the tool instantly extracts and organizes key exposure data, allowing CRC’s teams to engage carriers faster and position risks with greater precision. On the market side, REDY INTEL analyzes CRC’s dataset in real time to identify shifting carrier appetites and the most competitive markets, giving retail partners better coverage options and higher hit ratios. For placement, historical pricing and coverage intelligence guides optimal deal structure from the outset, allowing teams to anticipate likely terms, evaluate alternatives early, and align placements with carrier appetite and client objectives.
The launch reflects a broader trend identified in recent WTW research: P&C insurers that invested more heavily in advanced analytics and AI outperformed slower adopters between 2022 and 2024, achieving combined ratios six points lower and premium growth three points higher. Wholesale distributors that build similar capabilities into their platforms stand to create meaningful competitive separation.
For retail brokers, the more immediate takeaway is practical. Tools like REDY INTEL change what they can reasonably expect from a wholesale partner, including faster turnaround, more granular market guidance, and better-structured submissions. As AI becomes embedded in the infrastructure of specialty placement, retail partners who understand how to leverage those capabilities will be better positioned to secure competitive terms for clients. The distribution technology race is no longer a future-tense story.
— Reinsurance News Editorial Team at Reinsurance News
This Week’s TL;DR
This week’s edition captures AI moving from strategic conversation to operational reality across the insurance and risk management landscape. The physical infrastructure powering the AI economy is generating a new class of property risk that underwriters are still working to fully understand. As data center construction scales rapidly, fire hazards from lithium-ion battery systems, liquid cooling failures, and fragmented multi-program structures are creating accumulation challenges that standard property frameworks were not designed to handle. With empirical loss data still thin and few next-generation facilities fully operational, specialized technical assessment is becoming a prerequisite for sound underwriting in this space.
On the liability side, the market for AI-specific coverage is beginning to take meaningful shape. Courts are consistently treating AI as a tool and placing liability on the organizations deploying it, even as standard policy forms across cyber, tech E&O, and commercial general liability leave significant gaps. The parallels to cyber insurance’s early days are hard to ignore, and the brokers and underwriters who begin developing AI liability expertise now are likely to find themselves well ahead of the market as dedicated forms and specialty programs emerge over the next several years.
In distribution, the shift is already underway. Wholesale platforms are embedding AI directly into placement workflows, changing what retail brokers can reasonably expect in terms of speed, market intelligence, and submission precision. Taken together, this week’s stories reflect an industry that is actively recalibrating across property, liability, and distribution. Practitioners who engage with these developments now, rather than waiting for the market to fully settle, will be better positioned as AI’s influence on risk and placement continues to deepen.