BY HASAN MUHAMMAD

The number that should command attention this week is not a geopolitical headline. It is 140 trillion. That is the daily AI token consumption across China as of March 2026, up from roughly 100 billion in early 2024, a thousandfold increase in two years, according to Liu Liehong, head of the National Data Administration. MWC Shanghai 2026, which closed on June 26, served as the stage for announcing what comes next.

China has already priced AI tokens at 9.9 yuan a month and deployed that pricing across three carriers simultaneously. China Mobile, China Telecom, and China Unicom are repositioning themselves as full-stack AI providers. China Mobile cut total 2026 capital expenditure by 9.5% while increasing investment in computing power networks by 62.4%. China Telecom lowered overall capital spending by 9.2%, allocating 35% of total capex toward computing infrastructure. China Unicom cut total expenditure by nearly 8%, directing more than 35% of spending toward AI and data centers.

Bloomberg reported Beijing is preparing to spend around $295 billion over the next five years on building data centers across the country, with the National Development and Reform Commission drafting a blueprint for a network of interconnected computing hubs. China's 15th Five-Year Plan formally designates the construction of six major national networks, including a next-generation communications network and a computing network, as a top-level priority. The NDRC has announced that implementation plans for all six networks are being released in sequence.

At MWC Shanghai, David Wang, Huawei's Deputy Chairman, argued that the boom in AI agents will drive rapid growth in token services requiring ultra-broadband networks with high uplink, high reliability, and low latency. The optical infrastructure, the AI compute stack, and the consumer billing layer are being engineered together. That vertical integration is the competitive advantage, not any single component within it. MWC Shanghai also made visible a less-discussed dimension of China's digital strategy: the relationship between deployment scale and standards power.

China's 5G-A deployment across 330 cities is a data-generation machine feeding Chinese standards proposals to 3GPP with field evidence that competing proposals from carriers who have not yet deployed cannot match. The 6G standard-setting process runs through 3GPP and the ITU on a timeline extending to 2028. The architectural decisions being made now, which spectrum bands to standardize, whether integrated sensing and communication becomes a core feature, and how AI is embedded in the radio access network, will determine which countries' equipment can interoperate and whose cannot.

The window to shape 6G architecture with comparable field evidence is narrowing. An Accenture study cited at MWC Shanghai found that 79% of global telecom operators remain at Level 0 or Level 1 on the TM Forum's six-level network-autonomy scale, meaning operations are largely manual or rely only on basic monitoring. China Mobile has already reported measurable operational gains from Level 4 deployments. That gap in real-world deployment is a gap in standards influence.

China has demonstrated that AI and telecom integration, when executed at national scale with coordinated state and carrier investment, produces a different kind of digital economy, one where AI services reach mass-market pricing years before Western counterparts achieve comparable deployment. China's ability to site large-scale data centers in its western regions under the East-West Computing Resources Transmission Project converts renewable energy overcapacity into a durable AI infrastructure cost advantage. That advantage compounds over time.