--- title: China’s new AI bottleneck isn’t chips. It’s running out... description: China's AI development faces a growing bottleneck beyond chip shortages: a severe lack of high-quality Chinese-language training data. Chinese accounts for... canonical: https://daily.dev/posts/china-s-new-ai-bottleneck-isn-t-chips-it-s-running-out-of-chinese-language-training-data--n28xgzf7e twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: China’s new AI bottleneck isn’t chips. It’s running out of Chinese-language training data. | daily.dev og:description: China's AI development faces a growing bottleneck beyond chip shortages: a severe lack of high-quality Chinese-language training data. Chinese accounts for... og:url: https://daily.dev/posts/china-s-new-ai-bottleneck-isn-t-chips-it-s-running-out-of-chinese-language-training-data--n28xgzf7e og:image: https://api.daily.dev/og/posts/N28xgZf7E.png og:image:alt: China’s new AI bottleneck isn’t chips. It’s running out of Chinese-language training data. og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # China’s new AI bottleneck isn’t chips. It’s running out of Chinese-language training data. **[The Next Web](https://daily.dev/sources/tnw)** · 2 min read · 0 upvotes · 0 comments ## Summary China's AI development faces a growing bottleneck beyond chip shortages: a severe lack of high-quality Chinese-language training data. Chinese accounts for only 1.3% of global web content compared to nearly 50% for English. Walled-garden platforms like WeChat and Douyin refuse to share data externally, forcing AI labs to rely on lower-quality sources. Epoch AI estimates global high-quality public text could be exhausted within six years. Beijing is responding with a national plan to build validated AI training datasets by 2028 across sectors including healthcare, finance, and manufacturing. Meanwhile, Chinese publishers are beginning to ban AI training use of their content, further tightening supply. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What percentage of global web content is in Chinese vs English? Chinese accounts for just 1.3% of global web content, compared to nearly 50% for English, 6% for Spanish, and 5% for Japanese, according to internet tracker W3Techs. This disparity means Chinese AI developers must train models with far less native-language material, paying more per useful token than Western counterparts because models must work harder with limited data. _Developers building multilingual models track data availability challenges like this on daily.dev._ ### When does Epoch AI predict global high-quality public training data will be exhausted? Epoch AI estimates the worldwide supply of high-quality, publicly available text could be fully exhausted within six years. OpenAI co-founder Andrej Karpathy has similarly warned of a 'data wall' by decade's end, a constraint that hits Chinese AI development harder given the language's small share of web content. _Teams working on LLM training pipelines follow data supply research like this on daily.dev._ ### What is China's national plan to address the AI training data shortage? China's National Data Administration unveiled a nationwide plan in June to build validated AI training datasets by 2028, covering sectors including manufacturing, energy, healthcare, finance, agriculture, autonomous driving, and embodied AI. The initiative treats data as strategic infrastructure, with officials stating that AI competition depends on high-quality data supply systems, not just models and compute. _Engineers and researchers tracking AI policy and data infrastructure developments follow stories like this on daily.dev._ ## Similar posts on daily.dev - [China is closing in on US tech lead despite constraints](https://daily.dev/posts/china-is-closing-in-on-us-tech-lead-despite-constraints-h3iqbb7db) · TechCentral · 1 upvotes · 0 comments - [The next AI breakthrough won’t come from bigger models, but from better data](https://daily.dev/posts/the-next-ai-breakthrough-won-t-come-from-bigger-models-but-from-better-data-8ruccmjsi) · InfoWorld · 1 upvotes · 1 comments - [China narrows the AI gap with the US to a record-low 6%](https://daily.dev/posts/china-narrows-the-ai-gap-with-the-us-to-a-record-low-6--8sdtzu8as) · The Next Web · 1 upvotes · 0 comments - [China is increasingly keeping its best AI talent to itself](https://daily.dev/posts/china-is-increasingly-keeping-its-best-ai-talent-to-itself-rhfqhd11v) · TechCrunch · 0 upvotes · 0 comments --- Tags: [#ai](https://daily.dev/tags/ai), [#llm](https://daily.dev/tags/llm) [View this post on daily.dev](https://daily.dev/posts/china-s-new-ai-bottleneck-isn-t-chips-it-s-running-out-of-chinese-language-training-data--n28xgzf7e) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"China’s new AI bottleneck isn’t chips. It’s running out of Chinese-language training data.","url":"https://daily.dev/posts/china-s-new-ai-bottleneck-isn-t-chips-it-s-running-out-of-chinese-language-training-data--n28xgzf7e","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/china-s-new-ai-bottleneck-isn-t-chips-it-s-running-out-of-chinese-language-training-data--n28xgzf7e"},"datePublished":"2026-08-09T08:20:04.500Z","dateModified":"2026-08-09T08:20:31.580Z","description":"China's AI development faces a growing bottleneck beyond chip shortages: a severe lack of high-quality Chinese-language training data. Chinese accounts for...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/465b587fdbdeca6e86513e01fcce4bbf?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/465b587fdbdeca6e86513e01fcce4bbf?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The Next Web","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"The Next Web","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/tnw","url":"https://daily.dev/sources/tnw"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/china-s-new-ai-bottleneck-isn-t-chips-it-s-running-out-of-chinese-language-training-data--n28xgzf7e","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai,llm","timeRequired":"PT2M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The Next Web","item":"https://daily.dev/sources/tnw"},{"@type":"ListItem","position":3,"name":"China’s new AI bottleneck isn’t chips. It’s running out of Chinese-language training data."}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/china-s-new-ai-bottleneck-isn-t-chips-it-s-running-out-of-chinese-language-training-data--n28xgzf7e#faq","mainEntity":[{"@type":"Question","name":"What percentage of global web content is in Chinese vs English?","acceptedAnswer":{"@type":"Answer","text":"Chinese accounts for just 1.3% of global web content, compared to nearly 50% for English, 6% for Spanish, and 5% for Japanese, according to internet tracker W3Techs. This disparity means Chinese AI developers must train models with far less native-language material, paying more per useful token than Western counterparts because models must work harder with limited data. Developers building multilingual models track data availability challenges like this on daily.dev."}},{"@type":"Question","name":"When does Epoch AI predict global high-quality public training data will be exhausted?","acceptedAnswer":{"@type":"Answer","text":"Epoch AI estimates the worldwide supply of high-quality, publicly available text could be fully exhausted within six years. OpenAI co-founder Andrej Karpathy has similarly warned of a 'data wall' by decade's end, a constraint that hits Chinese AI development harder given the language's small share of web content. Teams working on LLM training pipelines follow data supply research like this on daily.dev."}},{"@type":"Question","name":"What is China's national plan to address the AI training data shortage?","acceptedAnswer":{"@type":"Answer","text":"China's National Data Administration unveiled a nationwide plan in June to build validated AI training datasets by 2028, covering sectors including manufacturing, energy, healthcare, finance, agriculture, autonomous driving, and embodied AI. The initiative treats data as strategic infrastructure, with officials stating that AI competition depends on high-quality data supply systems, not just models and compute. Engineers and researchers tracking AI policy and data infrastructure developments follow stories like this on daily.dev."}}]} ```