Today’s CTR
China’s technology contest has moved beyond the familiar question of whether its companies can catch their American rivals. The sharper issue is how quickly technical progress can be converted into durable influence. Beijing is mobilising capital and political authority behind domestic artificial intelligence chips, while Chinese model developers are making open systems increasingly difficult for Silicon Valley to dismiss. Washington, meanwhile, is debating controls that may prove harder to enforce than to announce. The mood is confident but not carefree: China has models, engineers and ambition in abundance, yet advanced manufacturing capacity remains the stubborn bottleneck. Software is sprinting; hardware is still catching its breath.
China Turns Its Artificial Intelligence Chip Drive Into a National Campaign
China is intensifying its effort to build domestic artificial intelligence (AI) chips, with Vice-Premier Ding Xuexiang reportedly overseeing a coordinated push across companies, laboratories and government agencies. Huawei and other Chinese groups have made progress using older deep-ultraviolet lithography equipment and larger clusters of locally produced processors.
The campaign appears to be reducing China’s reliance on imported AI chips, although domestic alternatives still lag Nvidia’s leading products in performance and manufacturing efficiency. The more significant development is organisational: Beijing is treating the semiconductor gap less as a normal commercial problem than as a strategic mobilisation.
That approach can accelerate investment and concentrate engineering talent, but it cannot repeal the physics of advanced lithography. China may narrow the practical gap by using more chips, better software and carefully optimised systems, even before it can manufacture an equivalent individual processor.
The contest is increasingly about usable computing capacity, not whose chip wins the beauty pageant.
Moonshot’s Kimi K3 Gives Washington a New Policy Headache
Moonshot AI’s Kimi K3 model is drawing attention in the United States after posting strong results in coding and reasoning tests. The open-weight model has reinforced the view that Chinese laboratories are closing the performance gap with leading American developers, despite constraints on access to advanced graphics processing units (GPUs).
The policy problem is awkward. Restrictions on chips can slow the construction of computing infrastructure, but they are less effective against software that can be downloaded, modified and deployed across borders. Further controls on Chinese models could also encourage overseas developers to build outside American platforms.
Kimi’s success is therefore commercial as well as geopolitical. Capable, inexpensive and adaptable Chinese models may prove particularly attractive in markets where customers value control over their data and infrastructure more than access to the absolute top benchmark score.
Open models travel lightly, which makes them unusually difficult to contain.
China Packages Artificial Intelligence as an Instrument of Diplomacy
President Xi Jinping is promoting Chinese artificial intelligence technology across the Global South through open models, training programmes and proposed application centres. China has also backed the World Artificial Intelligence Cooperation Organisation, a 29-member initiative intended to give Beijing a larger role in international AI governance.
The initiative extends the logic of the Digital Silk Road. Rather than exporting only telecommunications equipment or cloud infrastructure, China can now offer countries a broader technology package: models, technical training, computing systems and standards.
Lower-cost Chinese systems could gain traction in Southeast Asia, Africa, Latin America and the Middle East, particularly where governments are wary of dependence on a small number of expensive American providers. The tension is that Beijing is simultaneously considering tighter controls on the export of its most advanced technology.
China wants its AI ecosystem to be open enough to spread, but valuable enough to protect.
The White House Struggles to Agree on Chinese Artificial Intelligence Controls
The Trump administration is reportedly divided over how to respond to increasingly capable Chinese AI models. Some officials favour tighter restrictions, while others argue that controlling access to software and open model weights would be impractical and potentially counterproductive.
The disagreement reflects a broader weakness in technology containment policy. Physical products such as advanced processors move through factories, distributors and customs systems; digital models can be copied and distributed at negligible cost.
Overly broad restrictions could also encourage developers in Europe, Asia and emerging markets to adopt Chinese alternatives. A measure intended to preserve American leadership might inadvertently expand the international user base for Chinese technology.
Washington has a formidable chip-control toolkit, but software keeps slipping through the toolbox.
Chinese Open Models Challenge Silicon Valley’s Closed-Garden Strategy
Chinese AI developers are positioning their open models as stable and accessible alternatives to increasingly restricted American systems. Products from Moonshot AI, Alibaba and other laboratories can often be downloaded, customised and operated on a customer’s own infrastructure.
This is more than a philosophical difference over open-source software. It is a distribution strategy. American companies retain advantages in frontier performance and computing resources, but closed application programming interfaces (APIs) require customers to accept changing prices, usage rules and geopolitical exposure.
Chinese developers can exploit those concerns by offering governments and businesses greater technical control. The strategy may sacrifice some recurring cloud revenue, but it can accelerate adoption and make Chinese model architectures part of the global development stack.
Silicon Valley is selling intelligence as a service; China is increasingly giving customers the keys to the engine room.
China’s Chip Companies Crowd the Public-Market Exit
Chinese semiconductor companies are accelerating plans for initial public offerings (IPOs), with memory-chip producers and other strategically important firms seeking capital from domestic markets. At least six major chip companies have reportedly listed or begun listing processes since June.
The rush gives Beijing another mechanism for financing technological self-reliance. Public listings can supplement state funds while giving households and institutional investors direct exposure to the semiconductor campaign.
The risk is that strategic importance may be mistaken for commercial quality. Chipmaking requires enormous and sustained expenditure, and not every company entering the market will possess defensible technology or viable margins. A national priority can still produce very ordinary investments.
China’s semiconductor boom is entering its capital-markets phase; discipline may now matter as much as patriotism.
