A county board votes, a moratorium takes effect, and a data center project with hundreds of millions in committed capital stops — with no fire, no flood, no physical loss of any kind. That fact pattern is the coverage gap behind a product launched this week: Arbol, a New York-based parametric insurer, announced a Data Center Regulatory Moratorium Cover that pays a predetermined amount when a covered government action halting or suspending data center construction stays in force past an agreed waiting period.
The gap: a construction stop with no physical loss
Standard construction insurance responds to physical loss, and a government moratorium produces none. Builders risk pays to repair or replace damaged work; delay-in-start-up (DSU) cover pays for the financial cost of late completion — but only downstream of insured physical damage. Swiss Re’s DSU guidance states the prerequisite plainly: the insured property must sustain physical damage from an insured peril before DSU responds. A zoning vote damages nothing, so neither policy is triggered.
Business interruption does not fill the gap on a project that has not opened. Business interruption indemnifies lost income from an operating business interrupted by covered physical damage, and a data center still under construction has no operating income to interrupt. The exposure a moratorium creates — debt service, site carrying costs, committed capital sitting idle — accrues before revenue exists, which places it outside the shape of both DSU and business interruption.
Political risk insurance was built for a different map. Traditional political risk cover protects against sovereign acts abroad — expropriation, currency inconvertibility, political violence — rather than land-use votes by domestic county boards and town councils. As Arbol’s CEO Sid Jha put the contrast at launch, mainstream policies require physical property damage to trigger a payout, while traditional political risk coverage usually protects against sovereign expropriation abroad, “rather than suburban county board votes.”
Developers facing a moratorium have historically had three options, none of them insurance. Arbol’s launch materials frame the pre-product reality directly: because a government action produces no physical loss and sits outside standard insurance, developers were left to “lobby it, litigate it or absorb it.” The carrying costs of a suspended project — interest, taxes, security, contractor standby — were simply borne by the developer and its lenders.
What Arbol launched
Arbol announced the Data Center Regulatory Moratorium Cover on October 20, 2026, describing it as a first-of-its-kind product; that “first” is Arbol’s own characterization, carried in its launch materials and trade coverage. The cover is sold to data center developers through brokers, and Arbol positions it as relevant to the investors and capital providers financing the build-out as well.
The cover targets government action against data center construction as a class. Per Artemis’s reporting on the launch, the product is designed to protect against government halting or suspending data center construction as a class, rather than site-specific permit denials — a broker distinction that matters, because an individual project’s permit denial is a different, narrower event than a jurisdictional moratorium and is not what this product describes.
The stated purpose is to protect two specific exposures. Arbol says the cover is designed to protect the capital already committed to a project and the carrying costs that accumulate while construction sits idle. For lenders, Arbol argues, that converts an unhedgeable regulatory exposure into a budgeted cost that can be priced into a credit decision.
How the parametric trigger works
The trigger is an event in the public record, not a loss adjustment. Artemis reports that the trigger and settlement are based on information from the public record: a covered government action takes effect, and if it remains in force beyond the waiting period defined in the policy, the predetermined amount pays. No adjuster measures the developer’s actual loss, and no proof of damage is required, because damage is not the trigger.
The waiting period is the product’s deductible, expressed in time. A moratorium that lifts before the waiting period runs out pays nothing; a moratorium that outlasts it pays the set amount. That structure makes the waiting period the single most important term a broker negotiates against the client’s realistic timeline — local moratoriums commonly run six to twelve months, so a waiting period set near the length of a typical moratorium can quietly empty the cover of value.
The payout is a predetermined amount, agreed when the policy is written. That certainty is the product’s selling point and its limit in one: the amount does not rise if the project’s actual carrying costs run higher than expected, and it does not fall if they run lower. Parametric cover trades indemnity precision for speed and objectivity — Jha’s launch statement was “we pay objectively when a government action halts a project.”
Availability has a hard edge brokers need to know before they pitch it. Reporting by Insurance Journal (Bloomberg) states the product is available nationwide in the US, that premiums vary sharply by geography, permit category, and project exposure — and that Arbol excludes places with active moratoriums, citing prohibitive cost. A client already under a moratorium is, on that reporting, outside the market for this cover; the time to place it is before the vote, not after.
Why this product arrived now
The exposure count is large even on independent measures. Data Center Watch, a monitor run by 10a Labs and cited by Bloomberg, counted at least 106 US data center projects representing roughly $198 billion in investment blocked or delayed amid local opposition in the first half of 2026 alone. Arbol’s launch materials cite a bigger jurisdictional footprint — 45 state-level measures across 36 states and restrictions in 194 local jurisdictions as of August 2026, plus a March 2026 Gallup survey in which 71% of Americans opposed an AI data center in their local area; those figures are Arbol’s, presented in its launch case.
The binding constraints are local, and that is where brokers should look first. State moratorium bills draw the headlines, but many fail; SemiAnalysis, which tracks more than 400 local instruments, counts more than 300 local moratoriums and bans enacted nationwide as of September 2026. Arbol makes the same point in its launch materials: the exposure “has moved from state houses to county boards.”
The honest counterweight is that exposure is not the same as delay. SemiAnalysis’s project-level analysis estimates only about 1,525 MW of US capacity is genuinely delayed by local moratoriums, because most instruments freeze new applications rather than revoke approvals already granted, and entitled projects proceed. A broker quoting moratorium risk off headline counts alone will overstate most clients’ exposure; the risk is parcel-specific and stage-specific, which is precisely what the checklist below is for.
An independent expert voice at launch struck the same note of caution about the product itself. Robert Hartwig, director of the Risk and Uncertainty Management Center at the University of South Carolina, told Bloomberg the cover “could go a long way to calming the nerves” of investors, but that because the product is novel and must be highly tailored to location-specific risk, it will also be “difficult to roll out rapidly.” Brokers should expect bespoke underwriting, not a commoditized form: Arbol says it prices the exposure using localized data including water availability, grid capacity, power rates, and the jurisdiction’s history with data centers.
The broker checklist: a project stopped by government action
1. Builders risk and DSU — confirm the physical-damage wall in writing. The first check is negative confirmation: builders risk responds to physical loss to the work, and DSU responds to delay caused by insured physical damage, so a regulatory stop triggers neither. Getting that answer documented early protects the broker and focuses the client conversation on the actual gap instead of a claim that was never going to pay.
2. Government action exclusions — read the exclusions, not the brochure. Property and construction forms commonly restrict or exclude loss arising from governmental action and ordinance enforcement, and those exclusions are a second reason the standard tower does not respond to a moratorium. The broker check is form-specific: which policies in the client’s program carry governmental-action wording, and does any endorsement carve anything back.
3. Business interruption and contingent covers — check for a non-damage trigger, expect none. A pre-revenue project has no business income to interrupt, and standard BI requires covered physical damage in any event. The check still belongs on the list because clients assume “the policy that pays when we can’t operate” exists; for a construction-stage asset, it does not.
4. Political risk — test the domestic land-use fit before assuming it. Political risk markets write expropriation and sovereign interference, typically outside the insured’s home country. A domestic county moratorium on data centers does not match the classic political risk trigger set, so the broker check is whether any placed political risk wording reaches domestic regulatory action at all — and to record the answer rather than assume the line responds.
5. The moratorium itself — establish whether it actually binds this project. SemiAnalysis’s tracking shows most moratoriums freeze new applications and leave existing approvals intact, which means stage is destiny: a fully entitled project may be untouched, while a project awaiting a permit inside the jurisdiction is fully exposed. The broker’s factual work is parcel-level — is the site inside the boundary (county instruments often stop at city lines), is the needed approval already in hand, does the instrument carry a size threshold the project falls under or over, and are there vested-rights or carve-out arguments.
6. The parametric wording — pin down what “halts or suspends construction” means. Arbol’s cover addresses government action against data center construction as a class, not site-specific permit denials, so the broker must establish how the form treats an applications freeze that prevents a project from starting versus an order stopping work already underway. That single definition decides whether the most common form of moratorium — the applications freeze — is inside or outside the trigger.
7. Waiting period and payout — size both against the client’s real numbers. The waiting period should be set against the moratorium durations actually seen in the client’s jurisdictions, and the predetermined payout should be set against a documented carrying-cost figure: debt service, taxes, site security, and standby costs per month of suspension. A payout chosen as a round number, untested against carrying costs, recreates basis risk inside a product bought to remove uncertainty.
8. Timing and eligibility — place it before there is a headline. On launch reporting, Arbol excludes locations with active moratoriums, so eligibility is a pre-event purchase. The broker check is portfolio-wide: which clients have data center exposure in jurisdictions where opposition is organizing but no instrument has yet passed — those are the accounts where this conversation is still available.
9. Lender requirements — ask what the credit side needs. Arbol’s launch case is aimed at lenders as much as developers: moratorium risk has no established actuarial history or market price, which the company argues makes it hard to underwrite in a credit decision. Brokers serving financed projects should ask lenders directly whether moratorium cover affects covenants, reserves, or pricing — the answer can turn an optional purchase into a financing condition.
The honest limits
This cover carries basis risk by design. The payout is fixed at inception while actual carrying costs are not, so a long suspension on a large project can consume the predetermined amount well before construction resumes. Brokers should present the product as a budgeted partial offset — Arbol’s own framing is a plannable cost — not as indemnity for the full economic loss of a stopped project.
The product is new, untested at claim time, and priced without actuarial history. Arbol’s launch materials concede the underlying risk “has no established actuarial history, market price or transfer mechanism,” and no moratorium claim has yet tested the trigger language in public. Early buyers are purchasing a definition — of a covered government action, of the waiting period, of the public record that settles it — and the broker’s value is in reading those definitions harder than the marketing.
What the launch does establish, independent of any one vendor, is that the gap is real and now has a market answer where none existed. A government stop with no physical loss sat outside builders risk, outside DSU, outside business interruption, and outside classic political risk; as of October 2026 there is at least one purpose-built parametric form aimed at it, sold through brokers. The broker’s job is unchanged in kind and new in content: verify that the standard tower truly does not respond, establish whether a moratorium actually binds the client’s parcel and stage, and then test the parametric wording — trigger, waiting period, payout, exclusions — against the client’s real carrying costs before recommending it.