Listed insurance companies' computing power "arms race" heats up, and AI is no longer just a cost-reduction tool

Ask AI · How can AI upgrade from cost-reduction tools to an insurer’s core strategy?

21st Century Business Herald reporter Sun Shihui, intern Xu Ruoxuan

Recently, listed insurers have been disclosing annual reports in large numbers. China Life, Ping An, CPIC (China Pacific Insurance), PICC (People’s Insurance Company of China), New China Life, China Taiping, and Sunlight Insurance, among other leading insurers, have successively turned in their 2025 performance reports.

As you review the annual reports, “digital and intelligent transformation” has shifted from a slogan to real operational investment. AI is embedded throughout the strategic layout, infrastructure, business scenarios, and value growth—becoming the core main thread for rebuilding the industry’s competitive logic. However, on the other side of this transformation coin are the real, hard-dollar battles between capital-intensive computing infrastructure investment and long-term returns that have yet to fully materialize. With the AI arms race already in full heat, what every insurer needs to answer is not only how to move faster, but also how to find a sustainable rhythm in the tradeoff between investment and returns.

This is a high-stakes gamble about the future—and the stakes have already been placed on the table.

AI shifts from a “multiple-choice question” to a “required-answer question”

The 2025 annual reports show that multiple leading insurers have increased AI’s strategic weight in their top-level design, viewing it as the core driving force for future growth.

At a performance briefing, Guo Xiaotao, Co-CEO of Ping An, stated clearly: “AI is not a multiple-choice question; it is a required-answer question.” Ping An Group is advancing its “Comprehensive Finance: Jiu Jiu Gui Yi” initiative, aiming to integrate more than 700 million internet-registered users into a unified super entry point driven by AI. This will achieve full aggregation of traffic, entry points, and back-end data, enabling customers to complete a closed loop of medical, elderly care, and comprehensive financial services within a one-stop entry.

CPIC has listed “AI+” as one of its three major strategies for the future. In a low-interest-rate period, traditional profit models that rely on the interest-rate spread urgently need to change. Industry competition is shifting from dividend-driven growth to games in existing market share. At a performance briefing, CPIC Chairman Fu Fan said that during the “15th Five-Year Plan and 16th Five-Year Plan” period, the company will focus on implementing three major strategies: “big healthcare and elderly care, internationalization, and AI+.” The “AI+” strategy will target core business scenario needs, promote large-scale deployment of AI technologies, enhance efficiency and effectiveness, drive the reengineering of business processes, optimize customer experience, and innovate service models.

People’s Insurance (PICC) and China Life have also demonstrated a similar push to secure dominance in AI. At a performance briefing, Ting Xiangqun, Chairman of PICC Group, clearly positioned the technology line as an “accelerator,” proposing that it should “seize opportunities in the development of artificial intelligence more proactively, deepen reform of the technology system and digital construction, accelerate the release of technology’s productive forces, and capture the commanding heights of digital and intelligent transformation.” In 2025, the year-on-year growth rate of AI application scenarios rolled out by PICC Group reached 79%.

Meanwhile, China Life Chairman Cai Xiliang lists “technology dividends” as one of the four major dividends for the next five years, saying it will “forge future-facing digital and intelligent capabilities, and drive upgrades to management, products, and business models through digital and intelligent transformation.” In 2026, China Life has made forward-looking arrangements for 14 reform projects, treating digital and intelligent transformation as a strategic priority.

This strategic shift is also directly reflected in the talent market. During the spring recruitment in 2026, the “AI content” in the insurance industry surged significantly. Among the 4,500 open positions at Ping An, roles in technology and artificial intelligence account for nearly 30%. PICC added a large number of positions for artificial intelligence development and data security offense and defense. Sunlight Insurance also listed robotic engineering and data engineering as recruitment priorities.

Computing infrastructure and data platforms have become the focus of investment

Behind the strategic upgrades is real investment in computing power. The annual reports show that leading institutions are increasing investment in underlying infrastructure such as data centers, hybrid cloud platforms, and large-model training, advancing depth from the application layer to the hardware layer and infrastructure layer.

Ping An has already formed a significant scale in computing power and data accumulation. The annual report shows that Ping An, through an industry-leading database built internally, has cumulatively stored more than 3.2 trillion high-quality text corpora and 8.5 billion image corpora, supporting precise operations for its 251 million individual customers. In 2025, more than 70,000 intelligent agent applications were developed by Ping An Group employees, with a total of 3.65 billion model calls throughout the year.

China Life and PICC have continued to push efforts in data centers and algorithm engineering. Through its “Digital China Life” strategy, China Life has built a digital platform based on hybrid cloud and created a data space featuring “hundreds of millions of data—thousands of features—hundreds of dimensional labels.” PICC has also been optimizing its computing layout in an orderly way, advancing the construction of its western data centers. Its northern information center has obtained national certification for green computing infrastructure. By using a self-developed vertical large model for the insurance domain, “PICC (PICC Chenling),” the scenario-intent understanding accuracy rate has exceeded 99%.

New China Life and China Taiping have also achieved leaps in underlying infrastructure. In 2025, New China Life deployed the first batch of data centers nationwide. The area of its computer rooms increased to 27,000 square meters, and its core supporting capability grew by nearly four times. China Taiping, meanwhile, has basically achieved end-to-end coverage from computing infrastructure to business enablement. It launched a series of “Zhi Xiang AI” products and deployed 50 application scenarios.

Technology investment converted into core performance growth momentum

The ultimate results of technology investment are reflected directly in business growth and value creation. In 2025, multiple listed insurers enabled core business through digital and intelligent capabilities, achieving a quality leap from “scale expansion” to “value creation.”

Against the backdrop of fluctuations in the industry’s manpower scale, insurers have expanded agents’ service radius by pushing tools such as AI consultants and digital employees down to frontline operations.

The “Six-Arm” intelligent application ecosystem developed by CPIC Life uses AI scenario rehearsals to raise agents’ productivity by 15.7%. For users who use “Smart Customer Jing Assistant” frequently, the 30-day conversion rate increased by 1.23 times. In 2025, CPIC Life’s new business value reached 18.609 billion yuan, up 40.1% year over year, and its new business value margin rose by 3.2 percentage points.

Through “human + AI” collaboration, Ping An enables agents to transition into professional elderly-care managers, significantly improving customer retention rates. In 2025, new business value for life insurance and health insurance grew 29.3% year over year, and new business value per agent channel increased by 17.2%.

Sunlight Life’s new business value in 2025 reached 7.64 billion yuan, up 48.2%. By optimizing sales and claims experience through “robotic engineering,” it helped lift total premium scale to a platform of over one trillion.

On the operations side, the efficiency gains brought by digital and intelligent transformation have significantly optimized insurers’ cost structures. In 2025, New China Life achieved attributable net profit of 36.284 billion yuan, up 38.3%. Its intelligent claims review and early warning system went live, and its “rapid completion within 1 minute” rate for preservation services reached 96%. China Life’s proportion of claim cases handled through intelligent claims services is also over 75%, with more than 5.3 million one-stop direct settlement claims.

The other side of AI investment: cost pressure and payback cycle

However, not all investments can deliver immediate results. The annual reports also implicitly reveal tremendous cost pressure—data center construction and the introduction of high-end AI talent are becoming new “profit-eaters” that are consuming earnings.

From the perspective of capital expenditures, the area of New China Life’s computer rooms rose from 7,000 square meters to 27,000 square meters, and its core supporting capability grew by nearly four times. Such infrastructure expansion means a substantial increase in capital spending in the current period. CPIC has stated clearly that this year’s AI investment will be double that of last year, and the compounded growth rate of the investment budget over the next two years will be no less than 40%.

From the perspective of talent costs, among Ping An’s 4,500 spring recruitment positions in 2026, the proportion of technology roles is nearly 30%. PICC added many artificial intelligence development positions—where compensation levels for high-end AI talent are far higher than those for traditional finance practitioners, which will continue to push up labor costs. From depreciation and amortization, large-scale investment in computing power infrastructure will be converted into depreciation expenses over the next several years, creating sustained pressure on profits.

Even more difficult to measure is the investment-output cycle. Building AI capabilities follows the rule of “high input, long cycle, and non-linear returns”—massive upfront investment such as computing platforms, data governance, and model training comes first, while the release of business value often requires years of iteration and fine-tuning. How to balance short-term investment with long-term output will be a common examination question faced by all insurers under their “AI+” strategies.

The transformation of the insurance industry is not a wave that passes overnight, but a set of measures that is gradual, steadily promoted, and far-reaching in impact. In its Deloitte report, “2026 Global Insurance Industry Outlook,” it points out that data quality, system modernization, and security assurance are key factors for AI success. Looking ahead to 2026, insurance companies should further solidify the digital foundation, improve the quality of investment data, advance the transformation of talent strategies, integrate digital touchpoints with human service touchpoints, and conduct proactive risk management. A research report by Citic Securities notes that AI applications are entering a singularity moment, and in the future, AI effectiveness will be concentrated in dimensions such as efficiency improvement, value creation, and deep decision-making.