China's Bold Leap: The Zhengzhou 60,000-Chip AI Supercomputing Hub Goes Live

China has marked a pivotal moment in its technological self-reliance journey with the full activation of its largest scientific artificial intelligence computing cluster at the Zhengzhou core node of the national supercomputing network. Announced on April 14, 2026, this infrastructure now boasts 60,000 domestically developed AI accelerator chips, doubling its capacity from the initial 30,000 chips used during trial operations that began on February 5. Located in Zhengzhou, Henan Province, this hub stands as the nation's most powerful platform for AI-driven scientific research, often referred to as AI for Science (AI4S), enabling breakthroughs in fields like drug discovery, new materials development, and clean energy solutions.

The rapid expansion—from trial to full scale in just two months—highlights China's accelerated push to build sovereign computing capabilities amid ongoing US export restrictions on advanced semiconductors. By relying entirely on homegrown technology, the cluster circumvents dependencies on foreign suppliers like NVIDIA, positioning Chinese universities and research institutions at the forefront of global AI innovation.

Technical Foundations: Sugon's scaleX and Domestic Chip Ecosystem

Developed by Dawning Information Industry, commonly known as Sugon, the cluster leverages the innovative scaleX platform, comprising multiple scaleX640 supernodes interconnected via the proprietary scaleFabric network—a fully domestic Remote Direct Memory Access (RDMA) solution ensuring ultra-low latency data transfer. Each supernode supports up to 640 AI cards in a single cabinet, a world-first design that maximizes density and efficiency for trillion-parameter AI models.

The 60,000 chips represent six types of self-developed processors, delivering performance on par with leading global systems while integrating over 1,000 open-source large language models (LLMs). Connected to China's national supercomputing network, it provides access to more than 3 million CPU cores and 200,000 GPUs, democratizing high-performance computing for smaller entities. This full-stack ecosystem—from data aggregation to application deployment—eliminates the need for users to manage complex software configurations.

Defying US Curbs: A Milestone in Technological Sovereignty

Facing stringent US export controls since 2022, China has invested heavily in indigenous alternatives like Huawei's Ascend series and Phytium CPUs, which power Sugon's systems. Chen Jing, vice president of the Technology and Strategy Research Institute, emphasized that the cluster's scale rivals Meta and Google's facilities but uses no NVIDIA ecosystem, marking a 'key breakthrough in large-scale commercial deployment of domestic AI chips.'

This self-reliance is crucial for higher education, as universities previously reliant on imported hardware can now scale AI simulations without supply chain vulnerabilities. The Zhengzhou hub exemplifies China's 15th Five-Year Plan (2026-2030) priorities, coordinating computing power with energy infrastructure to fuel the 'new quality productive forces' agenda.

Empowering AI for Science: Key Applications in Research

Designed primarily for AI4S, the hub accelerates complex simulations unattainable on traditional systems. In drug discovery, it enables rapid molecular modeling to identify novel compounds; for materials science, it simulates quantum properties for superconductors or batteries; and in clean energy, it optimizes fusion reactor designs or photovoltaic efficiencies.

• Drug discovery: Protein folding predictions in hours vs. weeks.
• New materials: High-throughput screening of 10^6 candidates daily.
• Clean energy: Climate modeling with unprecedented resolution.

The 'super scientific computing agent' revolutionizes workflows: Researchers submit natural language queries (e.g., 'Simulate protein-ligand interactions for COVID variants'), and the system decomposes tasks, selects models, allocates resources, and delivers results autonomously.

Boosting Henan Universities: Zhengzhou University Leads the Charge

Zhengzhou University (ZZU), a 'Double First-Class' institution overseeing the National Supercomputing Center since its 2020 inception, stands to gain immensely. ZZU's School of Computer Science and Artificial Intelligence integrates supercomputing for smart city governance, environmental monitoring, chemistry, materials, biology, and medicine research. Collaborations like the China-Russia Supercomputing Research Institute further amplify its global reach.

Other Henan powerhouses—Henan University of Technology and Zhengzhou University of Light Industry—benefit from shared access, enabling faculty and students to tackle grand challenges without prohibitive costs. The center's location in the Zhengzhou High-Tech Zone, home to these universities, fosters seamless academia-industry synergy.

For instance, ZZU researchers have developed the Kunshu geological language model using supercomputing resources, showcasing potential for earth sciences acceleration.

National Network Integration: Democratizing Access for All Institutions

As part of China's seven national supercomputing centers, Zhengzhou links to a vast pool of resources, allowing universities nationwide— from Tsinghua to smaller provincials—to submit jobs remotely. This levels the playing field, enabling PhD students at Henan Agricultural University to run large-scale genomic analyses alongside elite labs.

In higher education, this translates to enhanced graduate training: Master's theses on climate modeling or theses on quantum chemistry become feasible, boosting publication rates and international rankings.

Real-World Impacts: Case Studies and Early Wins

Though newly operational, precedents from trial phases include optimized material designs for batteries, reducing simulation times by 90%. Universities like PLA Information Engineering University in Zhengzhou, a military academy focused on AI, will leverage it for cybersecurity and intelligent systems research.

Stakeholder perspectives: Henan provincial officials hail it as a catalyst for 'industrial transformation,' while academics note it shortens research cycles from months to days, fostering innovation hubs.

Expert Insights: Self-Reliance Meets Global Ambition

Chen Jing underscores the disruptive user interface: 'No more IT hurdles—natural language drives end-to-end science.' This agentic paradigm aligns with global trends but prioritizes accessibility for Chinese academia.

Challenges include energy demands (addressed via 'computing-power coordination') and model fine-tuning for domain-specific tasks. Solutions: Nationwide data sharing and open-source LLMs tailored for Chinese research.

Future Outlook: Scaling Nationwide and Global Leadership

Plans include further expansions and replication across centers, per the 15th FYP. For higher ed, expect surges in AI-specialized PhDs, interdisciplinary centers at unis like ZZU, and international collaborations (e.g., Belt and Road AI exchanges).

Implications: Positions Chinese universities as AI research powerhouses, attracting talent amid global competition. Actionable for academics: Apply via national portals for compute time, focusing on AI4S proposals.

For more on research opportunities in China, explore research jobs.

Photo by Jon Geng on Unsplash