China's two largest telecom operators are restructuring around "token operations" as artificial intelligence rewrites the economics of connectivity.
China's two largest telecom operators are restructuring around "token operations" as artificial intelligence rewrites the economics of connectivity.

China Mobile Ltd. established a Token Office to unify AI token creation, transmission and application, joining China Telecom Corp. in a strategic pivot from traffic-based to token-based operations that could reshape how China's $200 billion telecom industry monetizes its networks.
"AI is profoundly reshaping all aspects of human production and daily life, and telecom operators must reconstruct their service models and network architecture," China Mobile Chairman Chen Zhongyue said.
The Token Office follows the creation of China Mobile's Digital Intelligence Business Unit and Computing Power Office, consolidating functions previously scattered across departments including mobile cloud, digital intelligence, marketing and government-enterprise divisions. China Telecom Chairman Ke Ruiwen separately described the shift as moving from "traffic operations" to "token operations," from connection-driven growth to intelligence-driven growth, and from cybersecurity to system security. At a June 25 conference in Shanghai, China Telecom launched the AI Token Global Service Ecosystem Alliance and its XINGCHEN TokenHub platform, which connects more than 300 mainstream large language models across 230 global cloud nodes.
The restructuring shows that China's state-owned telecom giants see AI inference — the process of running trained models — as the next revenue driver after years of slowing consumer growth. China Mobile shares fell 1.23% in Hong Kong trading, with short selling accounting for 21.4% of volume, as investors weigh the costs and timeline of the transition.
The "token" in this context refers to the fundamental unit of data processed by AI models — each query, image generation or document analysis consumes a measurable number of tokens. By positioning themselves as token operators, China Mobile and China Telecom aim to charge for AI compute consumption rather than traditional data throughput, a model that could generate higher revenue per unit of network traffic.
Token Operations Reshape Telecom Economics
China Telecom's OneGrowth 2026 conference, held under the theme "Shared AI Token Era Ahead," laid out the infrastructure underpinning this transition. The carrier's "2+5+X" global AIDC layout is accelerating, with the ALC international submarine cable landing in Hong Kong and total international cable capacity exceeding 304 terabits across 185 cables. GSMA Chief Executive John Hoffman praised China Telecom's "forward-looking deployment of globalized computing power infrastructure" as a benchmark for the industry's digital transformation.
The XINGCHEN TokenHub platform serves as the operational backbone, offering a one-stop service for token production, orchestration and application. Its TeleAgent feature provides one-click access to major global large language models, targeting enterprise customers with customized AI packages. For small and medium businesses, the platform offers standardized token packages, creating a tiered pricing model that mirrors how cloud computing evolved from custom contracts to pay-as-you-go services. China Telecom also launched four lightweight quantum products, including eSurfing Quantum Secret, which landed first in the Asia-Pacific region. The company's satellite direct-to-phone service went live in Hong Kong and Laos, while its global Internet of Vehicles "One Card" capability now covers more than 230 countries and regions.
Samsung's $200 Billion Bet Adds to AI Infrastructure Race
The telecom token push coincides with a broader surge in AI infrastructure investment across Asia. Samsung Electronics Co. announced a 265.5 trillion won ($200 billion) investment plan, with 203 trillion won directed toward semiconductor clusters in Yongin and Pyeongtaek, South Korea, and AI computing power data centers. The plan highlights the scale of capital required to support the AI inference workloads that China Mobile and China Telecom aim to monetize through their token platforms. Jefferies analysts project storage chip prices will rise 40% to 50% quarter-on-quarter in the third quarter of 2026, driven by supply tightness and AI demand, with a 30% to 40% sequential increase expected in the fourth quarter.
For investors, the question is whether token operations can generate returns commensurate with the infrastructure buildout. China Telecom's platform already connects 300-plus models, but the revenue contribution from token services remains undisclosed. China Mobile's stock trades at roughly 10 times forward earnings, a discount to global telecom peers such as AT&T and Verizon, reflecting market skepticism about the pace of the AI-driven revenue transition. The Bank for International Settlements warned in its latest annual report that if AI investment returns fall short of expectations, the current capital expenditure boom could trigger "significant macroeconomic turmoil," citing complex circular financing models in the AI industry chain as a particular risk.
The success of this strategy depends on whether enterprise demand for AI inference grows fast enough to offset the decline in traditional voice and messaging revenue. China's AI market is projected to reach $80 billion by 2028, according to industry estimates, providing a potential revenue pool that neither carrier can afford to ignore.
This article is for informational purposes only and does not constitute investment advice.