China's state nuclear sector says it has started large-scale production of silicon-28 at 99.99 percent purity, a small materials claim with real weight for silicon-based quantum computing.
The announcement is not just another line in the long list of Chinese quantum milestones. If CNNC's claim holds up in production, China has moved one of the more awkward inputs for silicon quantum chips from an overseas sourcing problem into a domestic industrial process.
The June 15 announcement came from the Research Institute of Physical and Chemical Engineering of Nuclear Industry, a Tianjin-based lab under state-owned China National Nuclear Corporation. According to Chinese state media and CNNC's own account, the institute achieved China's first independent, large-scale production of silicon-28 isotope at isotopic abundance above 99.99 percent. Before this, high-purity silicon-28 supply was concentrated outside China, including Russian and European channels, which made it exactly the sort of niche dependency that becomes strategically important only when a country can no longer assume access.
Silicon-28 has zero nuclear spin. That is the whole point. Natural silicon is mostly silicon-28, but it also contains silicon-29 and silicon-30. Silicon-29 has a non-zero nuclear spin, and that magnetic noise is the problem for silicon spin qubits because it can shorten coherence time and make qubit states harder to preserve. Silicon-30, like silicon-28, has zero nuclear spin, so the issue is not that all heavier silicon isotopes create interference. The issue is the silicon-29 left in the material.
As Xinhua reported, CNNC described the material as the purest silicon and an indispensable core material for silicon-based quantum chips. That sounds like promotional language, but in this case the physics gives the phrase some substance. Silicon quantum computing researchers have spent years trying to reduce the amount of silicon-29 around qubits because cleaner isotope composition gives the device one less source of noise to fight.
The strategic weight is easy to miss because silicon-28 is not a headline chip like a GPU, a lithography tool, or a finished quantum processor. Washington has spent years restricting China's access to advanced semiconductors and manufacturing equipment, betting that hard-to-replace supply chain chokepoints would slow Beijing in critical technologies. A material bottleneck is less visible than an export ban on a machine from ASML, but it can matter in the same way if a lab or company cannot buy what it needs at the right purity.
China's answer has been to close as many of those gaps as it can through state labs, industrial ministries, and companies that sit outside Western regulatory reach. CNNC's announcement fits that pattern. It is a nuclear research institute, not a venture-backed quantum startup, saying it can now make a quantum computing input at scale. That is how a national technology program looks when it is built around supply security as much as scientific prestige.
The race is not one-sided. ASP Isotopes, which operates isotope enrichment facilities in South Africa, has said it is building silicon-28 supply for quantum computing and next-generation semiconductor customers, including through commercial supply agreements. The important point is not that one side has finished the race. It is that both sides now understand isotope supply as part of the quantum stack, not as a specialist chemical procurement detail left for later.
China's broader quantum program already had stronger proof points than this materials announcement. Origin Quantum's 72-qubit Origin Wukong machine went online in January 2024 and, according to China Science and Technology Daily reporting cited by Investor's Business Daily, had drawn more than 20 million remote visits by February 2025. USTC's Zuchongzhi 3.0, a 105-qubit superconducting processor described in a 2025 Physical Review Letters paper, posted random circuit sampling results designed to compete in the same benchmark territory as Google's 105-qubit Willow processor.
Those machines do not all depend on silicon-28 in the same way. Superconducting processors are a different hardware route from silicon spin qubits. But that distinction actually sharpens the point. China is not placing one bet on one quantum architecture. It is building across hardware approaches while trying to make the supporting materials less vulnerable to outside pressure.
CNNC also pointed to uses beyond quantum computing, including advanced semiconductor manufacturing, high-end navigation systems, and precision measurement. Those applications are plausible, but they are not why this announcement deserves attention now. The nearer question is whether silicon-based quantum computing can take advantage of existing semiconductor manufacturing knowledge while still meeting the extreme materials requirements that qubits impose. A domestic silicon-28 source at 99.99 percent purity would remove one practical obstacle from that path.
The harder part has not disappeared. Producing a purified isotope is not the same as building a fault-tolerant quantum computer, and no country has solved that problem. But supply chains decide which labs can keep iterating when the politics turn hostile. Monday's silicon-28 claim is a reminder that China's technology competition with the US is not only being fought at the level of finished chips. It is also being fought in the quiet materials layers underneath them.
Also read: NewCore launches with $66 million to give AI agents proper enterprise identities • Fox Is Buying Roku for $22 Billion Because the Home Screen Is Now Television • Nvidia sells $20 billion in bonds as AI demand outpaces its own cash generation
