
From short-video browsing and AI tool usage to large-scale model training, industrial simulation, and scientific research breakthroughs, computing power serves as a critical enabler. As the computing network accelerates its deployment, it faces challenges such as physical environment interference, failed computing resource scheduling, and data exposure during transmission. In response, many financial institutions have launched innovative products—including "Computing Power Insurance," "Compute Coverage," and "Token Protection"—tailored for AI and computing industry enterprises, with computing-related scenarios serving as the insured scope.
Why can computing power be insured?
Reporters learned that computing power insurance currently available primarily focuses on property insurance for data center operations and liability insurance covering the provision of computing services.
Wang Guojun, a professor at the School of Insurance of University of International Business and Economics, stated that computing power is a core productive asset in the AI era with significant economic value. However, it faces risks such as data center failures and cyberattacks, which are identifiable and quantifiable—meeting the fundamental criteria for insurability under property insurance. Consequently, financial institutions treat it as an insurable subject.
For example, PICC Property and Casualty's "Computing Power Insurance" provides dual security assurance of "comprehensive hardware asset protection + full operational liability coverage." The comprehensive cybersecurity insurance for computing power, "Kunshu Shield · Token Insurance," covers three layers: data security, computing power service assurance, and hardware equipment. China Life Property & Casualty Guangdong Branch led the pilot implementation of a computing hub cybersecurity insurance program, focusing on scenarios such as cyber attacks, data breaches, system failures, and business interruptions to provide financial risk protection for the Shaoguan data center cluster.
"Computing power insurance is not a simple expansion of traditional property insurance. Instead, it integrates physical assets, service continuity, and digital rights into a unified risk pool, marking a theoretical shift from protecting assets to safeguarding operations," said Tian Lihui, Dean of the School of Finance at Nankai University. Insurance underwriting requires three conditions: the subject must have monetizable economic value, face identifiable pure risks, and allow for loss assessment and compensation. Computing power precisely meets these criteria. Especially now, as computing power evolves into a newly defined, priced, and tradable asset, its transactional nature provides institutional support for insurable interest.
How can enterprises obtain compute insurance services?
According to financial institutions such as PICC Property and Life Insurance and China Life Property & Casualty, the computing power insurance products currently available are tailored for enterprises within the local computing power industry chain.
For example, computing power center operators—including owners, co-owners, lessees, and users of the center—as well as computing power service providers, such as enterprises, public institutions, social organizations, and other entities offering computing power services, may all apply for "Computing Power Protection" services.
Tian Lihui stated that computing power is a portfolio of expensive physical assets—servers, accelerators, storage, and power/cooling infrastructure—all with clear market prices and depreciation schedules; equipment damage is compensable. Once deployed, computing power becomes a continuous service delivery capability. Interruptions in training, data loss, or information breaches result in quantifiable revenue losses and liability costs. This shifts the scope of coverage from hardware to operational liability.
However, some experts argue that the scarcity of historical claims data and the rapid evolution of risk profiles make traditional large-number laws and static actuarial models unsuitable. When risks materialize, underwriting requires simultaneous coordination across hardware, networks, data, and algorithms. With a shortage of multi-skilled professionals, unified industry standards and contract terms still lacking, liability boundaries often remain ambiguous. To establish a sustainable commercial model, regulatory guidance is needed to enable data sharing, dynamic actuarial pricing, standardized contracts, reinsurance support, and risk mitigation subsidies.
From pilot to full-scale rollout: refining the service ecosystem
Most compute insurance products currently on the market are still in early pilot phases. In the future, compute insurance won't just be about selling more policies; it will become a connector within the tech finance ecosystem.
According to Tian Lihui, as digital assets like tokens become measurable and tradable, insurance coverage will expand from hardware and services to digital rights. This shift will elevate insurers from mere risk bearers to risk managers and credit enhancers, while transforming computing power from a technical resource into standardized, investable, tradable, and securitized assets—providing systemic support for tech finance.
"The hallmark of a mature ecosystem is not the number of policies issued, but whether we can ensure insurance benefits are genuine, losses are properly compensated, and speculative risks remain off the books," said Tian Lihui.
Wang Guojun believes that insurance serves both loss compensation and financing enhancement functions. It should collaborate with computing power service providers, banks, reinsurers, and third-party adjusters to build an integrated tech-finance ecosystem encompassing pre-event risk control and early warning, post-event claims settlement, and supporting financing solutions.
Industry insiders also note that computing power companies must explore financing, leasing, collateralization, and asset securitization, while financial institutions need trusted risk pricing and credit enhancement tools. Insurance can provide a pricing anchor for computing assets through actuarial analysis, help data centers secure credit via policy-backed credit enhancement, compensate for payment delays through credit insurance, and cover operational disruptions and data breaches with liability insurance. In the future, there is significant potential for collaboration between insurance and banks, leasing firms, trusts, securities firms, and industrial investors—enabling full-cycle coverage of computing infrastructure construction, circulation, and usage.



