University of Illinois professor Rakesh Kumar argues that US export controls have pushed China to build domestic chip and AI alternatives, rather than stopping its progress. The case matters for technology policy: China’s growing ability to run AI on homegrown chips complicates the idea that restricting advanced hardware alone can preserve a lasting lead.
Watch Desk analysis
What happened
In a Fortune interview published on 16 October, Kumar says the controls have encouraged investment in a domestic semiconductor ecosystem. He points to Huawei’s statement that it had shipped more than 100 million devices powered by its Kirin processors since 2023, and to Chinese AI companies adapting models to Huawei’s Ascend chips. Meituan has also said it trained a 1.6-trillion-parameter model entirely on domestically made chips.
Kumar’s argument is not that China has matched the US across the board. He distinguishes strategic progress from commercial competitiveness, saying Chinese chipmakers are finding workarounds but still face disadvantages against Nvidia and AMD. The article also cites IDC figures, via Reuters, showing Nvidia’s share of China’s data-centre chip market fell to 34% in the first half of 2026, with shipments of its AI accelerators down 44% year on year.
Why it matters
Export controls are meant to limit access to advanced technology. Kumar’s contention is that they can also make domestic substitutes a national priority, shifting the race towards how efficiently a country can use the hardware it can obtain. That matters beyond frontier-model training: the article notes that robotics and smart cameras can use less advanced chips.
The figures and examples do not, on their own, establish that controls have failed. Kumar is making a policy argument, and the commercial gap he describes is an important qualification: building alternatives for strategic uses is not the same as matching rivals on cost, performance or scale.
Our read
The strongest point here is the distinction between slowing access to leading-edge chips and preventing useful AI development. Those are not the same policy objective, however often they are squeezed into the same sentence. Kumar makes a case that restrictions have helped spur substitutes; the evidence he cites also leaves open how competitive those substitutes will be commercially. The sensible test is not whether China can build around constraints, but whether it can do so at scale and at a cost that changes the market.
What to watch
- Whether Huawei’s chip shipments and China’s AI workloads continue to grow.
- Whether Chinese chips prove competitive for commercial inference, not only strategic deployments.
- How US export controls evolve as officials weigh slowing China’s progress against encouraging domestic alternatives.
Discussion spark: Should chip export controls be judged by whether they slow access to the best hardware, or by whether they prevent rivals from building competitive alternatives?
Sources and evidence
- Source update (16 October 2026, 07:00 UTC)
- Source update (16 October 2026, 07:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.