🧠 Food for thought

Implications, context, and why it matters.

China’s AI progress depends on easing a hardware bottleneck

• China tops global counts for AI papers, citations, total patents, and industrial robot installations. US-led export controls on advanced chips still limit its goal of becoming the top AI power by 2030 1.
• Beijing has put large public money behind chipmaking, including the National Integrated Circuit Industry Investment Fund. The state-backed vehicle, known as the “Big Fund,” raised US$47 billion in its latest round to push a more self-reliant semiconductor supply chain 1.
• Domestic chips still fall short for training large AI models at scale. DeepSeek, a Chinese AI startup, viewed Huawei’s Ascend 910C AI chip as insufficient for training in early 2025 2.
• Chip limits remain even as more top researchers stay. The share of elite AI scientists working in China rose to 28%, up from 11% three years earlier 2.

The US–China AI race is nudging governments toward direct tech spending

• The US–China race has pushed “AI industrial policy,” where governments spend directly or offer subsidies to support domestic AI ecosystems. Some also back national champion companies 3.
• Other regions have followed. The European Union launched InvestAI to mobilize 200 billion euros (US$217 billion) for AI and supported “AI Factories,” large computing and development hubs, with at least 13 expected to run by 2026 3.
• Big tech has also shifted policies on defense work. OpenAI dropped a ban on using its products for “military and warfare,” and Google revised guidelines to allow some military uses, including weapons and surveillance, according to sources cited in the AI Now analysis 3.
• Some policymakers in the US and UK have framed the race as a values contest. They have urged a focus on democratic principles to build trust and spread AI gains more evenly 4.