Merry Christmas in China

How China’s AI Super‑Network Quietly Rewired the Global Technology Race

On December 3, 2025, China activated a technological system unlike anything the world has seen: a 34,175‑mile (55,000‑km) optical AI network linking 40 cities and allowing the country’s top computing centers to operate as a single, unified supercomputer. The Future Network Test Facility (FNTF) achieves 98 percent of the efficiency of a single data center, despite spanning a distance longer than the Earth’s circumference. It can run 128 networks simultaneously, support 4,096 concurrent service trials, and move data at speeds that collapse days into minutes. In one early test, it transferred 72 terabytes of radio telescope data in 1.6 hours, a task that would have taken nearly two years on the public internet.¹²³ The world barely noticed, but China’s leadership understood exactly what it had built.

The FNTF is not merely a network. It is the operational core of a national strategy that has been accelerating all year; one that integrates breakthroughs in chip manufacturing, renewable energy, AI architecture, and national‑scale simulation into a single, coherent system. The December 3 activation was the moment these pieces snapped together, and President Xi Jinping and the CCP ensured they retain the mandate of heaven for generations.

Throughout 2025, China made rapid progress across the semiconductor stack, pioneering domestic Extreme Ultraviolet Lithography (EUL), refining Deep‑Ultraviolet Lithography (DUL), and using cryogenic defect‑reduction and Cryo‑Electron Tomography (cryo‑ET) to dramatically improve chip yields. These advances gave China the hardware independence required to scale compute clusters without relying on foreign suppliers.⁴ After all, export restrictions to the largest and most advanced manufacturing hub in human history could only serve as a delay tactic.

At the same time, China expanded its renewable energy footprint at a pace unmatched globally. A solar fleet approaching one terawatt, combined with under‑sea, wind‑powered data centers, created the energy foundation for a compute system that can route workloads to wherever power is cheapest and cleanest.⁵ This is the essence of the country’s “East Data, West Computing” strategy: western regions generate abundant renewable energy; eastern regions generate dense AI demand; the FNTF stitches them together into a single computational organism.⁶

China also invested heavily in national‑scale simulation, modeling everything from energy flows to cooling efficiency to the 2023 ring‑CLOs infrastructure redesign. These simulations gave China the ability to orchestrate compute placement with extraordinary precision. The FNTF is the real‑world expression of those models, a network that behaves less like the public internet and more like a deterministic, scheduled system.

And then there is CALM, China’s new Continuous Autoregressive Language Model. Built on vector‑to‑vector prediction, CALM is optimized for distributed training. It benefits from the FNTF’s parallelism, and the FNTF benefits from CALM’s efficiency. CALM is the first frontier‑class model designed for a national super‑network, not a single hyperscale cluster.⁷

Taken together, these developments form a self‑reinforcing loop: domestic chips enable cheap compute; cheap compute accelerates AI development; AI accelerates robotics and automation; automation strengthens manufacturing; manufacturing strengthens the supply chain; and the supply chain strengthens China’s geopolitical leverage. The December 3 activation didn’t start this cycle, it completed it, and the Dragon woke up.

China now enters 2026 with a vertically integrated, nationally orchestrated compute‑energy‑manufacturing system. It is a model fundamentally different from the decentralized, market‑driven approaches of the United States, Europe, and India. The implications are profound. China can offer compute access the way nations once offered oil or security guarantees. It can tie access to manufacturing commitments, data localization, or standards alignment. It can accelerate robotics deployment at a pace that reshapes global industrial competitiveness. And it can cultivate an open‑source ecosystem that grows faster simply because training is cheaper and faster on its network.

The world is only beginning to recognize what happened on December 3, 2025, let alone understand it. What is most auspicious is that in February 2026, we enter the year of the Fire Horse; what was a Chinese cultural revolution will likely become a Chinese technological one.

Closing Perspective for Policymakers and Business Leaders

For policymakers and business executives, the activation of China’s national AI super‑network is not simply a technological milestone; it is a structural shift in global power back to a multi-polar world. In the Age of AI, compute is a geopolitical asset, and China has built the first system that treats it as such. The countries and companies that thrive in this new environment will be those that recognize compute dependency as a strategic vulnerability, diversify their supply chains, and invest in interoperable architectures capable of operating across multiple jurisdictions. The December 3 activation signals the end of the era of fragmented, ad‑hoc digital infrastructure and the beginning of nationally coordinated compute systems. Those who fail to adapt will find themselves negotiating from a position of weakness in a world where access to compute is as consequential as access to energy or capital.

References

1. Gizmodo. “China Launches 34,175-Mile AI Network That Acts Like One Massive Supercomputer.” Gizmodo, December 3, 2025.
2. Future Network Test Facility. “FNTF Operational Overview and Performance Metrics.” Ministry of Industry and Information Technology, December 2025.
3. China Science Daily. “FNTF Transfers 72TB of FAST Telescope Data in Record Time.” China Science Daily, December 4, 2025.
4. China Semiconductor Journal. “2025 Breakthroughs in Domestic Lithography and Cryogenic Chip Manufacturing.” CSJ, October 2025.
5. International Energy Agency. “China’s Renewable Energy Statistics: 2025 Update.” IEA, November 2025.
6. Xinhua News Agency. “China’s East Data, West Computing Strategy Gains Momentum.” Xinhua, August 2025.
7. Institute of Automation, Chinese Academy of Sciences. “CALM: A New Frontier in Autoregressive Language Modeling.” CAS Technical Reports, September 2025.