> ## Content Index > Fetch the complete content index at: https://icopilots.com/llms.txt > Use this file to discover other available public pages before exploring further. # China's AI flywheel effect is massive - URL: https://icopilots.com/chinas-ai-flywheel-effect-is-massive/ - Published: 2026-07-30T07:54:00.000Z - Updated: 2026-07-30T07:54:00.000Z - Author: Philippe MEDA - Tags: AI, #ai open source, #flywheel, #china There's a very compact and clear explanation of how China is starting to get the edge on AI. It's published by the U.S.-China Economic and Security Review Commission in Washington (which is delightful in itself). It clearly summarizes a point I was making just a few weeks ago: the [U.S. and China are chasing AI supremacy in a very different way](https://icopilots.com/china-and-us-are-chasing-ai-supremacy-in-very-different-ways/). I quote the paper summary: 1️⃣ **China has opted to go all in on an open-source approach to AI.** Most Chinese labs publish model source code and weights. They also charge far less to use high-end products than their global competitors. This has resulted in the acceleration of global uptake of Chinese AI and created *a feedback loop where widespread adoption drives iteration, then further adoption.* As of publication, Alibaba’s Qwen models accounted for the largest model ecosystem on Hugging Face, with over 100,000 derivatives. 2️⃣ **This open ecosystem enables China to innovate close to the frontier despite significant compute constraints.** Chinese labs have narrowed performance gaps with top Western large language models. They have also developed key architectural and training advances that are now *industry standards.* 4️⃣ **Open model proliferation creates alternative pathways to AI leadership.** China’s strategy prioritizes data curation and refinement through the deployment of *embodied AI* in manufacturing, robotics, and research where *specialized, real-world data from widespread use may compound into advantages that proprietary U.S. models cannot easily replicate*, even if they maintain technical superiority on benchmarks. 5️⃣ **China’s open AI model strategy and its manufacturing dominance are mutually reinforcing.** As the Commission’s 2025 Annual Report documented, China’s industrial base generates “interlocking innovation flywheels” across adjacent sectors. Open models accelerate this dynamic by enabling low-cost AI deployment across factories, factories, logistics networks, and robotics—generating *real world data that feeds back into model improvement.* Beijing has built the institutional infrastructure to exploit this advantage, designating data as a formal factor of production and permitting enterprises to carry data assets on their balance sheets. 6️⃣ **U.S. export controls primarily target the digital loop**—restricting access to advanced chips used for frontier model training—**but are not well suited to addressing the physical loop of deployment-driven data creation and accumulation across China’s manufacturing base.** As open models reduce the compute required for effective deployment, China’s ability to generate proprietary industrial data at pace and scale becomes increasingly independent of access to cutting-edge hardware. This gap in the U.S. policy framework means that *even successful controls on training compute may not prevent China from building AI advantages rooted in its physical economy.* *One of my own takes at the time was exactly this:* > (...) when you build monopolies around model access and subscription revenue, you optimize for a different outcome than when you integrate AI into the physical and administrative fabric of a society at scale and iterate from there. And with AI, the huge payoff of real-life applications at scale is an immense data trove. For now, **I'd suggest you keep the notion of *Embodied AI* in mind. And if you're an industrial or an AI startup, clearly focus on this part of the market as a solid move forward.** --- The full paper is here: [Two Loops: How China’s Open AI Strategy Reinforces Its Industrial DominanceOpen vs. Closed Approach![](https://storage.ghost.io/c/8f/a3/8fa3eb46-8b94-4e7c-85af-c9fe45614239/content/images/icon/favicon-4dfc6eb4-afb4-4101-b357-70283c438824.ico)U.S.-China Economic and Security Review CommissionNgor Luong![](https://storage.ghost.io/c/8f/a3/8fa3eb46-8b94-4e7c-85af-c9fe45614239/content/images/thumbnail/two-loops-AI-graphic-3ae65951-ac42-48ba-b9ac-fd13ac0e72c0.png)](https://www.uscc.gov/research/two-loops-how-chinas-open-ai-strategy-reinforces-its-industrial-dominance?ref=icopilots.com) And my recent series on the importance of humanoid robots: [The humanoid robot form factor makes sense once you look at the dataPhilippe Méda argues the humanoid robot form factor wins on imitation-learning data, the resource AI training actually needs.![](https://storage.ghost.io/c/8f/a3/8fa3eb46-8b94-4e7c-85af-c9fe45614239/content/images/icon/android-chrome-512x512-95156ec9-4962-4f94-bfc7-822265beb575.png)innovation copilotsPhilippe MEDA![](https://storage.ghost.io/c/8f/a3/8fa3eb46-8b94-4e7c-85af-c9fe45614239/content/images/thumbnail/photo-1601132359864-c974e79890ac-7ba0b22d-dc4b-4979-afd1-7770eea870e1)](https://icopilots.com/humanoid-robot-form-factor-imitation-learning/)