Telecom Operators Report Collective Q1 Profit Decline as Token Operations Accelerate

Following their announcements of transitioning to token-based operations at the start of the year, the performance trends of China's three major telecom operators have drawn significant industry attention. Financial results for the first quarter show that reported revenue of 266.5 billion yuan, a 1.0% year-on-year increase, but its net profit attributable to shareholders declined by 4.2% to 29.3 billion yuan. achieved revenue of 131.4 billion yuan, up 2.3% year-on-year, while its net profit dropped 17.1% to 7.35 billion yuan. saw revenue decrease by 0.5% to 102.8 billion yuan, with net profit falling 17.6% to 4.885 billion yuan.

A comparison of the operators' disclosed data reveals that profit performance has notably weakened compared to the same period in 2025. Growth is increasingly concentrated in new business areas such as computing power, cloud services, and intelligent solutions. 's computing power business generated 15.4 billion yuan in revenue, growing 8.3% year-on-year. 's Tianyi Cloud revenue increased by 6.8%, while its intelligent service revenue surged 39.4%. continues to position communication services, computing services, and intelligent services as its three core businesses, emphasizing its transformation into a "world-class technology service enterprise."

Notably, market expectations had already factored in first-quarter profit pressure due to two key factors: a policy change increasing the value-added tax rate for telecom services from 6% to 9% effective January 1, 2026, and slowing growth in traditional connectivity businesses.

If evaluated solely by revenue, the three operators' quarterly performance appears reasonable. However, a longer-term perspective reveals clear limitations in their traditional connectivity operations. Although 's other business revenue grew 12.7%, its core service revenue declined 1.1% to 219.9 billion yuan, indicating plateauing growth in basic communication services. 's service revenue increased merely 0.7% to 122.7 billion yuan, significantly slower than previous growth rates driven by connection scale. 's service revenue fell 0.9% to 90.058 billion yuan, with total operating revenue declining 0.5%. Despite adding over 3.58 million mobile and broadband users in Q1, 's blended ARPU remained stagnant, reflecting limited profitability from new subscribers.

Cost and profit metrics more directly highlight the pressure. 's EBITDA decreased 5.0% to 76.7 billion yuan, with a net profit margin of 11.0%. 's pre-tax profit reached 6.1 billion yuan, but net profit fell 17.6%. still recorded a 17.1% profit decline despite a 0.5% reduction in operating costs.

This indicates that operators must confront both peak connectivity revenue and compressed profit margins due to tax burdens, depreciation, labor, and network support costs. Following the telecom tax adjustment implementation early this year, industry predictions of operators entering a "profit pressure preceding revenue adjustment" phase are materializing.

However, the slowdown in traditional businesses doesn't imply instability in core operations. currently serves 1.009 billion mobile customers, 333 million broadband subscribers, and maintains 1.504 billion IoT connections. has reached 310 million 5G network users, representing 71.3% penetration. exceeds 750 million IoT connections.

This demonstrates that connectivity businesses remain cash flow and customer base anchors for operators, though they increasingly function as "steady-state operations" rather than core valuation drivers. Future profit trends will increasingly depend on emerging businesses like computing power, cloud services, and AI applications.

The most closely watched aspect of recent financial reports is operators' prominently announced token operation strategy. Tokens represent the smallest data units in text processing. AI-based token operations focus on consumption, optimization, cost control, and value conversion, aiming to transform token usage into tangible business value.

became the first operator to elevate token operations to strategic importance. Its 2025 annual report established token services as its primary operational focus, integrating technological, talent, and channel capabilities to transition from traditional traffic management to token value management. During the Spring Festival, its intelligent bell service attracted over 4 million AI production users, with daily token consumption increasing 14-fold. Since March, cloud computer additions exceeded 60,000 units, driving 10-fold growth in daily token usage.

's Jiangsu branch has developed a token operation system using digital intelligence platforms, addressing challenges like slow fault localization, difficult troubleshooting, and unclear cost calculations. The branch's monthly large model calls now exceed 20 million, consuming over 30 billion tokens monthly. Through precision flow control systems, it has blocked over one million abnormal calls, reducing fault localization time from 5 minutes to 2 minutes—a 60% efficiency improvement.

While hasn't detailed specific token operation progress, it emphasizes "chain integration." The company aims to integrate global premium models into Mobile Cloud, establishing a service chain where "agents use tokens, tokens drive computing power, and computing power consumes electricity." It has built 92.5 EFLOPS intelligent computing clusters domestically, connecting with over 1,300 overseas data centers to create a global computing network.

At the 2026 World Mobile Congress, unveiled its Cloud AI's "Agent+Token+AI Cloud" framework, launching 200 AI applications and agents serving over 400,000 government, healthcare, and financial clients. Its Q1 report showed computing power revenue growing 8.3% to 15.4 billion yuan, with intelligent computing center revenue up 11.7%. Unicom Cloud serves nearly 440,000 enterprise clients, with cloud intelligence products exceeding 110 million paid users.

The first quarter is too brief to fully assess operators' transformation outcomes. However, current information suggests that token operations are shifting operators from mere "pipeline providers" to integrators of model calls, inference computing, industry applications, and infrastructure—creating billable, deliverable, and replicable AI service chains. This transforms computing resources into directly manageable services.

The critical question remains whether token operations can genuinely convert to profits, depending on scale effects and pricing capabilities. Q1 results haven't yet demonstrated strong convincing power: 's profit buffer still primarily comes from Tianyi Cloud and intelligent services; 's computing revenue growth outpaces traditional connectivity but coincides with overall profit decline; and 's integrated agent-token-computing chain hasn't prevented a 4.2% net profit drop. Token operations haven't immediately reflected in financial data.

Nevertheless, industry consensus holds that collective transition to token operations aligns with the AI industry's shift from training to inference. IDC predicts that by 2027, over 85% of Chinese organizations will transform traditional cloud environments into AI-native platforms, migrating value focus to AI-native architectures. Multiple studies indicate exploding inference-side demand, positioning token calls and inference computing as new growth engines.

Operators—with networks, data centers, edge nodes, enterprise clients, and localized delivery capabilities—are naturally positioned to assume infrastructure and application roles in the inference era. Thus, Q1 2026 financial reports represent a watershed moment: traditional connectivity businesses continue providing stable cash flow while emerging operations assume dual roles as strategic narratives and growth engines.

Token operations will further propel operators from "traffic managers" to "AI service providers." The key observation for China's telecom industry isn't whether individual operators can quickly reverse profit declines, but whether they can develop sustainable monetization paths through industry models, inference computing, edge nodes, enterprise privatization deployments, and overseas services. Only when tokens evolve from concept to revenue will this transformation cycle gain market recognition.