Artificial intelligence in China

Artificial intelligence in China is the research, development, deployment, and state governance of artificial intelligence within the People's Republic of China, the world's second largest national AI ecosystem after the United States and its closest competitor at the technical frontier. By 2025 China led the world in AI publications and patent filings, hosted the most widely downloaded open weights large language models (the Qwen and DeepSeek families), and operated a domestic AI chip program built around Huawei's Ascend accelerators to substitute for restricted United States hardware.^[7][26] Stanford's 2026 AI Index reported that the performance gap between the best United States and the best Chinese models on the Chatbot Arena leaderboard had narrowed to 2.7 percent by early 2026, down from 17.5 to 31.6 percentage points in May 2023, even though United States private AI investment outspent China by roughly 23 to 1.^[26] The Chinese government has treated AI as a national strategic priority since at least 2017, when the State Council issued the New Generation Artificial Intelligence Development Plan and set the goal of making China "the world's primary AI innovation center" by 2030, with a core AI industry valued at 1 trillion yuan.^[1] It has matched the rhetoric with industrial policy, procurement, regulation, and large public funds.^[1]

The Chinese AI sector spans the major internet platforms (Baidu, Alibaba, Tencent, ByteDance), telecom and hardware giants (Huawei, Xiaomi), a cohort of well funded large language model startups (the so called "AI tigers" including DeepSeek, Moonshot AI, MiniMax, Zhipu AI, 01.AI, Baichuan), specialist computer vision firms (SenseTime, Megvii, Yitu, CloudWalk), domestic chip designers (Cambricon, Biren, Moore Threads, Huawei's HiSilicon), and major academic centres at Tsinghua University, Peking University, the Chinese Academy of Sciences, and the Beijing Academy of Artificial Intelligence (BAAI). The release of DeepSeek's R1 reasoning model in January 2025 was widely seen as the first time a Chinese open weights model directly challenged the leading United States foundation models on cost and capability.^[10][19]

The ecosystem operates under tight constraints. Since October 2022 the United States has progressively restricted exports of advanced AI chips and the equipment needed to make them, cutting Chinese firms off from the Nvidia H100 and most successor parts.^[11] Chinese labs have responded with a combination of stockpiling, smuggling, software optimisation, and a domestic accelerator program led by Huawei's Ascend line.^[22] Domestically, generative AI is regulated under Cyberspace Administration of China (CAC) measures covering deep synthesis content, recommendation algorithms, and public-facing generative services. The obligations differ by service type; under the 2023 Interim Measures, security assessment and algorithm filing apply when a service has public-opinion attributes or social-mobilization capacity under existing rules.^[3][5]

Government strategy and policy timeline

Background and early plans

AI featured in Chinese national planning before 2017, but mainly as a sub component of broader information technology and robotics initiatives. The State Council's Made in China 2025 plan, issued in May 2015, named ten priority sectors including "new generation information technology" and "high end CNC machine tools and robotics," and committed the state to raising domestic content and self sufficiency in core components. The 13th Five Year Plan (2016 to 2020) repeated AI and machine learning as research priorities. The 2016 "Internet Plus AI Three Year Implementation Plan" jointly issued by the National Development and Reform Commission, Ministry of Science and Technology, Ministry of Industry and Information Technology, and the Cyberspace Administration set initial targets for AI applications in manufacturing and consumer services.

What did the 2017 New Generation AI Development Plan set out to do?

The central document is the State Council's New Generation Artificial Intelligence Development Plan, issued on July 8, 2017 (Guofa No. 35).^[1] The plan set a three step timetable: by 2020, the AI industry should be "in step" with the world leaders; by 2025, China should achieve major breakthroughs in basic AI theory and lead in some technologies; by 2030, China should be "the world's primary AI innovation center," with a core AI industry valued at 1 trillion yuan and related industries valued at 10 trillion yuan.^[1] An intermediate target set the core AI industry above 400 billion yuan by 2025.^[1] The plan described AI as "a new focus of international competition" and "a new engine of economic development," framing it as the main driver of China's industrial upgrading.^[1] It named six key tasks, including basic theory, key technologies (vision, speech, natural language, swarm intelligence), and infrastructure, and called for building open innovation platforms and supporting infrastructure.^[1]

A July 2017 translation by Graham Webster, Rogier Creemers, Paul Triolo, and Elsa Kania, published by New America, became the most widely cited English version of the plan and shaped how Western policy circles read Chinese AI ambitions.^[2] The plan also explicitly identified military civil fusion (junmin ronghe) as a core principle, formalising a channel between civilian AI research and the defence sector.^[2]

Generative AI and algorithm regulations

China was among the first jurisdictions to issue binding rules on generative AI. Three regulations stand out:

1. The Provisions on the Administration of Algorithmic Recommendations of Internet Information Services, issued by the CAC, Ministry of Industry and Information Technology, Ministry of Public Security, and the State Administration for Market Regulation in December 2021 and effective March 1, 2022. Article 24 requires providers whose services have public-opinion attributes or social-mobilization capacity to file with the algorithm registry. The rules also require providers to offer controls over personalized recommendations and prohibit unreasonable differential treatment of users.^[5]
2. The Provisions on the Administration of Deep Synthesis Internet Information Services, issued by the CAC in November 2022 and effective January 10, 2023. These cover deepfakes and other AI generated audio, images, and video, requiring conspicuous labelling, identity verification, and content moderation by providers.^[4]
3. The Interim Measures for the Management of Generative Artificial Intelligence Services, issued July 13, 2023 and effective August 15, 2023. The measures apply to generative services made available to the Chinese public and impose duties concerning training data, personal information, prohibited content, transparency, and user complaints. Security assessment and algorithm filing apply to providers whose services have public-opinion attributes or social-mobilization capacity under existing rules.^[3]

The CAC publishes filing records for algorithms covered by the recommendation and deep-synthesis regimes.^[5] These records should not be treated as a separate universal approval list for every generative model; the 2023 Interim Measures do not establish a separately named large-model filing regime.^[3]

AI Safety Governance Framework (2024)

In September 2024 China's National Technical Committee 260 on Cybersecurity, in coordination with the CAC, released version 1.0 of the AI Safety Governance Framework.^[6] The framework is voluntary in form but signals the official Chinese position on AI risk taxonomy. It groups risks into endogenous safety risks (issues with the models themselves, such as hallucination, bias, and adversarial robustness) and application safety risks (misuse for fraud, disinformation, cyberattacks, or critical infrastructure failure), and lists corresponding technical and governance countermeasures.^[6] China also signed the Bletchley Declaration at the November 2023 UK AI Safety Summit and the Seoul Declaration at the May 2024 follow up summit.

Five Year Plans and local plans

The 14th Five Year Plan (2021 to 2025), adopted at the National People's Congress in March 2021, raised AI to the level of one of seven "frontier sciences and technologies" alongside quantum information, integrated circuits, brain science, genetics and biotechnology, clinical medicine and health, and deep space, deep sea, and polar exploration. Beijing, Shanghai, Shenzhen, Hangzhou, Hefei, and Chengdu issued local AI development plans with their own targets and subsidies, and many established AI industrial parks. Shanghai's Zhangjiang AI Island and Beijing's Zhongguancun district hosted concentrations of AI startups and research centres.

Major industry players

Chinese AI activity is heavily concentrated in a small number of companies and research institutes. The table below lists the most prominent organisations as of 2025.

Scroll sideways for more →

The "AI tiger" label was popularised in Chinese tech media in 2023 to describe the cohort of well funded LLM startups: Zhipu, Baichuan, Moonshot, MiniMax, 01.AI, and (later) Stepfun. A narrower four-company formulation used in 2024 reporting grouped Zhipu, Baichuan, Moonshot, and MiniMax.^[21] DeepSeek, founded by High Flyer hedge fund founder Liang Wenfeng, sat outside this group at first but became the most internationally visible Chinese AI lab in 2025.

Baidu

Baidu was the earliest of the Chinese internet majors to invest seriously in deep learning. It hired Andrew Ng to lead its Silicon Valley AI lab from 2014 to 2017, established the Institute of Deep Learning, and released its first version of the ERNIE (Enhanced Representation through Knowledge Integration) language model in 2019. ERNIE Bot (Wenxin Yiyan in Chinese) launched on March 16, 2023, making Baidu the first Chinese company to publicly release a ChatGPT style assistant. Baidu's Apollo autonomous driving program, launched in April 2017, operates a fleet of robotaxis under the Apollo Go brand in cities including Wuhan, Beijing, and Chongqing. Baidu also designs the Kunlun AI accelerator, manufactured initially by Samsung on 14 nm and later by SMIC on more advanced nodes.

Alibaba and Qwen

Alibaba's DAMO Academy, established in 2017, became the parent of much of the company's AI research. The Qwen (Tongyi Qianwen) family of large language models launched in April 2023 and rapidly became the most widely used open weights model series of Chinese origin. Alibaba released a series of versions including Qwen, Qwen 1.5, Qwen 2, Qwen 2.5, and Qwen 3, with sizes ranging from 0.5 billion to several hundred billion parameters and dense and mixture of experts variants. Qwen 2.5 Coder topped open source coding leaderboards in late 2024. Qwen models are released under the Apache 2.0 or Tongyi licence and are available on Hugging Face and ModelScope, Alibaba's own model hub. Alibaba Cloud is the largest Chinese cloud provider and the primary serving infrastructure for Qwen.

Tencent and ByteDance

Tencent's Hunyuan family includes text, image, and 3D generation models. Hunyuan Large, released in November 2024, is a 389 billion parameter mixture of experts model with 52 billion active parameters. Hunyuan Video, an open source text to video model released in December 2024, was widely benchmarked against Sora and Kling. Tencent uses Hunyuan internally to power features in WeChat, QQ, and Tencent Meeting.

ByteDance, the parent of Douyin and TikTok, launched its Doubao consumer chatbot in August 2023 and quickly became the largest Chinese consumer assistant by monthly active users, surpassing 70 million MAU by late 2024 according to Aicpb. ByteDance's research lab, formerly Skylark and now branded as ByteDance Seed, develops the Doubao model series. Volcano Engine is the company's enterprise cloud arm and offers Doubao through API.

Huawei

Huawei plays a dual role as both a model and chip maker. The Pangu model series, first announced in 2021, includes Pangu Alpha (Chinese language LLM), Pangu CV, Pangu Sigma (a sparse model), and a Pangu Weather model that the company published in Nature in July 2023, claiming superior medium range forecasting performance versus the European Centre for Medium Range Weather Forecasts.^[17] Huawei's Ascend AI accelerators (310, 910, 910B, 910C) anchor Chinese efforts to substitute for Nvidia GPUs and are paired with the open source MindSpore deep learning framework, an alternative to PyTorch.

LLM startups

Zhipu AI, spun out of Tsinghua University's Knowledge Engineering Lab in 2019, develops the GLM and ChatGLM models, including GLM 4 (released January 2024) and the multimodal CogVLM. Zhipu raised significant funding from Alibaba, Tencent, Saudi Aramco's Prosperity7 Ventures, and others. The United States Department of Commerce added Zhipu and several subsidiaries to the Entity List in January 2025, citing concerns over support for Chinese military modernisation.^[12]

Moonshot AI, founded by Yang Zhilin in March 2023, became known for the long context Kimi chat assistant, which marketed a 200,000 character context window in October 2023 and expanded to 2 million characters in 2024. Moonshot raised about 1 billion USD from Alibaba and others in early 2024. The Kimi K1.5 model, released January 2025, focused on reasoning and was benchmarked against OpenAI's o1 series.

MiniMax, founded by Yan Junjie in late 2021, released the abab model series and the MiniMax M1 reasoning model in 2025. The company also operates Talkie, a global character chat app.

01.AI, founded by Kai Fu Lee in March 2023, released the Yi model family. Yi 34B, released November 2023, was briefly the highest scoring open weights non Chinese language model on the Hugging Face Open LLM Leaderboard. The company restructured its pre training operations into a joint venture with Alibaba in late 2024 amid the cost pressures of frontier training.

Baichuan AI, founded by former Sogou CEO Wang Xiaochuan in April 2023, released the Baichuan series and focused increasingly on medical applications.

DeepSeek

DeepSeek is the AI lab spun out of High Flyer, a quantitative hedge fund founded by Liang Wenfeng in 2015. High Flyer purchased thousands of Nvidia A100 GPUs in 2021 to 2022, before the United States export controls cut off direct access, giving DeepSeek an unusually large compute base for an independent lab. DeepSeek released DeepSeek LLM in late 2023, DeepSeek V2 in May 2024 with a novel multi head latent attention architecture, DeepSeek V3 in December 2024 (a 671 billion parameter mixture of experts model with 37 billion active parameters that the company reported was trained for about 5.6 million USD in compute), and the DeepSeek R1 reasoning model in January 2025.^[9][10] R1's release, with weights and a permissive licence, was widely covered in international media in late January 2025 and was associated with a sharp single day decline in Nvidia's share price on January 27, 2025.^[19]

Hardware ecosystem

Chinese AI hardware splits into three layers: GPU and AI accelerator design, fabrication, and high bandwidth memory.

Accelerator designers

• Huawei's HiSilicon designs the Ascend family. CSET found that the second-generation Ascend 910B series appeared in technical documentation in 2022 and was designed for domestic fabrication at SMIC. Its reduced active-core count probably reflected SMIC yield or capacity constraints.^[18] Huawei pairs Ascend with the open source CANN compute architecture and the MindSpore framework as alternatives to CUDA and PyTorch.
• Cambricon Technologies (Zhongke Cambricon), spun out of the Chinese Academy of Sciences in 2016, designs the Siyuan series of training and inference chips. Cambricon was the first pure play AI chip company to list on Shanghai's STAR Market, in July 2020.
• Biren Technology designs the BR series, originally targeted at training. Biren was added to the Entity List in October 2023.^[12]
• Moore Threads, founded by former Nvidia China general manager Zhang Jianzhong in 2020, designs the MTT series of GPUs aimed at AI training and rendering. The company was added to the Entity List in October 2023.^[12]
• Enflame Technology (Suiyuan) designs the i20 and similar inference chips and is backed by Tencent.
• Baidu's Kunlun chip line is now a separate subsidiary, Kunlunxin, and supplies internal Baidu AI workloads.
• Alibaba's T head subsidiary designs the Hanguang inference chip and the Yitian 710 server CPU.

Fabrication

Semiconductor Manufacturing International Corporation (SMIC) is the largest Chinese contract chip maker. SMIC achieved volume 14 nm production in 2019. TechInsights identified the Huawei Mate 60 Pro's Kirin 9000S system on chip as using SMIC's N+2 7 nm process.^[24] ASML has not shipped extreme ultraviolet systems to China. A September 2024 Dutch measure expanded the national export-license requirement to ASML's TWINSCAN NXT:1970i and NXT:1980i deep ultraviolet immersion systems.^[24] Other Chinese foundries include Hua Hong Semiconductor and Yangtze Memory Technologies; Yangtze Memory was added to the Entity List in December 2022.

Memory

High bandwidth memory is a binding constraint on Chinese AI accelerators. As of 2025 the world's HBM supply was effectively controlled by SK Hynix, Samsung, and Micron. Chinese efforts to indigenise HBM are led by ChangXin Memory Technologies (CXMT) and Wuhan Xinxin (XMC, a Yangtze Memory subsidiary), but Chinese HBM remained one or two generations behind the international leaders.

Software

The dominant Chinese deep learning frameworks are Baidu's PaddlePaddle (open sourced 2016), Huawei's MindSpore (open sourced 2020), and Alibaba's MNN. Chinese researchers also use international frameworks: PyTorch is the standard at the major LLM labs.

Academia and talent

The leading Chinese university AI programs are at Tsinghua University, Peking University, the Chinese Academy of Sciences (especially the Institute of Automation and the Institute of Computing Technology), Shanghai Jiao Tong University, the University of Science and Technology of China, Zhejiang University, Fudan University, Nanjing University, Westlake University (founded 2018 in Hangzhou), and the Hong Kong University of Science and Technology. Tsinghua's THUDM lab developed the GLM and CogVLM models that anchor Zhipu's product line, and many of the founders of the LLM startups are Tsinghua alumni: Yang Zhilin (Moonshot), Zhang Yaqin and Tang Jie (Zhipu), Wang Xiaochuan (Baichuan), and others.

The Beijing Academy of Artificial Intelligence (BAAI), established in November 2018 with backing from the Beijing municipal government, was the first non corporate Chinese AI institute to train a frontier scale model: the Wudao 1.0 (March 2021) and Wudao 2.0 (June 2021) models. Wudao 2.0 was reported to have 1.75 trillion parameters, although the architecture was a sparse mixture of experts and the model was never made widely available outside BAAI's collaborators.

How does China compare to the United States on research output and talent?

The Stanford AI Index reports show consistent Chinese leadership on quantitative measures of AI research. The 2025 AI Index, citing OpenAlex data, reported that China produced 23.2 percent of AI publications in 2023, Europe 15.2 percent, and the United States 9.2 percent. The United States led the share of the 100 most-cited AI publications, while China led AI patent filings.^[7] The 2026 AI Index reported that China accounted for 69.7 percent of global AI patent filings and that its industrial robot installation rate ran roughly nine times that of the United States.^[26] On talent flow, however, the same reports show that the United States retained the largest share of AI researchers working at frontier labs and the largest gross inflow of senior AI researchers, with China a net exporter of AI doctoral graduates.^[7] The 2026 AI Index reported that net AI talent migration to the United States had fallen sharply from its 2017 peak as more Chinese researchers stayed home.^[26]

The Center for Security and Emerging Technology (CSET) at Georgetown has tracked AI talent and workforce patterns since 2019.^[14] Its study of United States AI doctorates found that more than 90 percent of Chinese-national graduates stayed in the United States for at least five years, while a separate 2023 report mapped the geography and sectoral demand of China's AI workforce.^[14]

The ChinAI Newsletter, written since 2018 by Jeffrey Ding, then a doctoral student at Oxford and now an assistant professor at George Washington University, has translated Chinese language AI policy and industry writing into English and is widely cited as a primary source on Chinese AI debates.^[16]

Models and benchmarks

Chinese labs have produced a continuous stream of large language models since 2021. The list below covers the most prominent post 2023 releases.

Scroll sideways for more →

June to August 2026 release sequence

An August 14 post summarized the interval as "8 Chinese companies released 10 AI models in 75 days."^[29] June 1 through August 14 contains 75 calendar dates only when both endpoints are counted, while 74 days elapse between them. Primary sources support eight developer organizations and ten named events, but not ten equivalent releases. The sequence mixes hosted launches, later weight publication, a preview-to-official transition, dated checkpoint updates, a staged video-product rollout, and future release promises. It is a snapshot of release cadence, not a uniform measure of model quality, training speed, or open-weight output.

Scroll sideways for more →

Most Chinese open weights models are distributed via Hugging Face and Alibaba's ModelScope, but their licenses differ. The original DeepSeek V3 weights use the separate DeepSeek Model License, which permits commercial use, while DeepSeek R1 and DeepSeek V3.2-Exp use MIT.^[27] Through 2025 DeepSeek continued the lineage with DeepSeek V3.1 and the V3.2-Exp release (September 2025), which kept the V3 backbone but added a sparse attention mechanism to cut the cost of long context inference.^[27]

On standard benchmarks (MMLU, GSM8K, MATH, HumanEval, SWE Bench, GPQA), the strongest Chinese models in 2024 to 2025 ranged from comparable to within a few months of the leading United States closed models, depending on the benchmark. Contemporaneous Epoch AI model tracking and Chatbot Arena results placed Qwen 2.5 and DeepSeek V3 or R1 among the leading Chinese or open-weight models, but rankings varied by date, prompt mix, and evaluation design.^[25] By early 2026 Stanford's AI Index measured the Arena gap between the best United States and best Chinese model at just 2.7 percent.^[26]

Regulatory framework

China's AI regulation operates within the broader internet governance regime built around the 2017 Cybersecurity Law, the 2021 Data Security Law, and the 2021 Personal Information Protection Law. Generative AI services made available to the Chinese public must comply with the 2023 Interim Measures, including duties concerning lawful training data, personal information, prohibited content, transparency, and complaint handling.^[3] Providers whose services have public-opinion attributes or social-mobilization capacity must also complete the security-assessment and algorithm-filing procedures required by the existing rules.^[3][5] Deep-synthesis providers have separate identity-verification, content-governance, and labelling duties.^[4]

The Interim Measures prohibit providers from generating categories of content already barred under Chinese internet-content rules and require remedial steps when prohibited content is found.^[3] They do not impose one identical pre-release approval procedure on every model or service.

The measures apply to generative AI services offered to the public in the People's Republic of China. Their scope provision excludes research, development, and internal use by industry, education, research, cultural institutions, and public bodies when those activities do not provide generative AI services to the domestic public.^[3]

In March 2025 the Cyberspace Administration of China and three other departments issued measures on labelling AI-generated and synthetic content, effective September 1, 2025. The measures distinguish visible labels from metadata-based implicit labels and apply through the existing algorithm-recommendation, deep-synthesis, and generative-AI regimes.^[3]

As of August 15, 2026, China had not adopted a comprehensive Artificial Intelligence Law. A March 2026 National People's Congress report recorded representatives' proposals to formulate such a law.^[40]

Export controls and the chip war

The most significant external constraint on Chinese AI is United States export control of advanced semiconductors and the equipment to make them. The Bureau of Industry and Security at the Department of Commerce (export controls) has issued a series of measures:

• October 7, 2022: A package of new rules placed performance based controls on advanced AI chips (defined by total processing performance and interconnect bandwidth thresholds), restricted exports of semiconductor manufacturing equipment for sub 14 nm logic, sub 18 nm DRAM, and 128 layer NAND, and added a foreign direct product rule covering integrated circuits incorporating United States technology destined for designated Chinese facilities.^[11]
• October 17, 2023: An update tightened the chip thresholds (capturing the Nvidia A800 and H800 designed for the Chinese market), expanded the geographic scope to additional countries, and added new entity listings.^[12]
• December 2, 2024: BIS added controls on high bandwidth memory, semiconductor-manufacturing equipment and software tools, and expanded Foreign-Direct Product Rule coverage.^[42]
• January 13, 2025: The Biden administration's "AI Diffusion" Interim Final Rule established a tiered country framework dividing the world into a top tier of close allies with broad access to advanced AI chips, a middle tier subject to country level compute caps and validated end user requirements, and a bottom tier (China and embargoed countries) facing comprehensive denial.^[13]
• May 13, 2025: The Trump administration's Bureau of Industry and Security formally rescinded the AI Diffusion rule two days before its compliance date, on the basis that it was overly complex and undermined United States diplomacy. Separate guidance warned that use of listed PRC advanced-computing chips, including Huawei's Ascend 910B, 910C, and 910D, risks violating United States export controls and may expose companies to enforcement when the Export Administration Regulations apply.^[28]

The practical effect on Chinese AI compute has been substantial. From October 2023, exports, reexports, and in-country transfers to China of advanced-computing chips meeting the controlled performance thresholds, including the H100, required BIS authorization and were generally reviewed under a presumption of denial.^[12] Nvidia's H20, a chip designed to fall just below the United States thresholds, became the dominant Nvidia part shipped to China in 2024. The United States imposed a license requirement on H20 exports in April 2025; Nvidia said on July 14 that it had been assured licenses would be granted, and later reported receiving licenses in August 2025.^[28] Chinese firms responded with: (1) accelerated procurement of domestic accelerators, especially Huawei Ascend; (2) algorithmic and systems optimisation, exemplified by DeepSeek's reported low cost training runs;^[9] (3) rental of overseas compute, including some grey market access to United States cloud regions; and (4) reported smuggling, documented in Reuters and Wall Street Journal reporting through 2024 and 2025.^[22]

The Entity List, maintained by BIS, has expanded to include Chinese AI hardware companies and model developers. The October 2023 rule added Moore Threads, Biren, and related entities, while January 2025 revisions added Beijing Zhipu Huazhang Technology and subsidiaries.^[12]

Secondary effects of the controls include: the bifurcation of the global AI software stack (CUDA dominant outside China, an emerging Ascend and CANN stack inside China); the growth of Chinese model exports to third countries through open weights distribution; and a domestic policy push to accelerate self sufficiency in semiconductor manufacturing, organised around the China Integrated Circuit Industry Investment Fund (the so called "Big Fund"), whose third phase was announced in 2024 at roughly 340 billion yuan.^[15]

Investment and funding landscape

Private investment in Chinese AI grew rapidly from 2018 to 2021, contracted in 2022 and 2023 with the broader Chinese tech downturn, and rebounded in 2024 alongside the global generative AI boom. According to the Stanford AI Index 2025 report, China attracted about 9.3 billion USD of private AI investment in 2024, second to the United States at 109.1 billion USD.^[7] The 2026 AI Index put China's 2025 private AI investment at about 12.4 billion USD, against roughly 23 times more in the United States, underlining that China's near parity in model quality has been achieved at a fraction of the capital outlay.^[26] Earlier years were closer: the 2024 AI Index reported 2021 private AI investment of 52.88 billion USD in the United States and 17.21 billion USD in China.^[8] The CB Insights and PitchBook trackers show similar gaps with different absolute numbers. Industry analysts attribute the relative slowdown in Chinese private AI funding after 2021 to a combination of regulatory tightening on internet platforms (the 2020 to 2022 platform rectification campaign), domestic capital market headwinds, and the chilling effect of United States outbound investment screening on cross border capital flows.

Government guidance funds and state owned investors play a larger role in Chinese AI investment than in the United States. The most prominent are the China Internet Investment Fund (operated by the Cyberspace Administration), the Big Fund (Integrated Circuit Industry Investment Fund), and various municipal funds in Beijing, Shanghai, Shenzhen, and Hefei. State participation is not uniform across firms. A 2025 CETaS mapping identified state-backed funds in some newer AI ventures and golden shares in major internet companies, while private and corporate investors remained prominent across the AI startup ecosystem.^[20]

Mergers and acquisitions of independent Chinese AI labs accelerated in 2024 and 2025 as the cost of frontier training rose. The 01.AI pre training joint venture with Alibaba (announced December 2024 and finalised in early 2025), the Baichuan partnership with Tencent, and several smaller transactions consolidated frontier scale training inside the major cloud providers. Bloomberg reported in August 2024 that Tencent joined a financing round of more than 300 million USD for Moonshot AI alongside existing backers Alibaba and Gaorong Capital.^[23]

DeepSeek and the open weights story

The release of DeepSeek V3 on December 26, 2024 and DeepSeek R1 on January 20, 2025 marked the first widely recognised moment when an independent Chinese lab matched the leading United States closed models on a major capability dimension while making weights and methodology public.^[9][10] DeepSeek V3 reported a training compute cost of 2.788 million GPU hours on Nvidia H800 chips, equivalent to about 5.576 million USD at the lab's stated 2 USD per GPU hour rate.^[9] As the technical report put it, "the full training of DeepSeek-V3 costs only 2.788M H800 GPU hours," a figure the report stressed "excludes the costs associated with prior research and ablation experiments on architectures, algorithms, or data."^[9] The accompanying technical report described an architecture combining mixture of experts with multi head latent attention, FP8 mixed precision training, and a custom load balancing strategy.^[9] The R1 reasoning model used reinforcement learning from rule based rewards on top of V3 to elicit chain of thought style reasoning.^[10]

The global response in late January 2025 was unusually visible. The DeepSeek consumer app reached the top of the Apple App Store free charts in the United States on January 27, 2025.^[19] Nvidia's share price fell about 17 percent that day, wiping nearly 600 billion USD of market capitalisation.^[19] Coverage in the Financial Times, Reuters, Wall Street Journal, and Bloomberg framed R1 as a turning point in the assumed United States lead. Analysts noted, however, that DeepSeek's reported 5.576 million USD described the final V3 pretraining run rather than prior experiments, personnel, data, or owned infrastructure. Epoch AI's reconstruction treated the disclosed chip-hour total as plausible but not as total model-development spending.^[25]

DeepSeek's broader contribution to the Chinese AI sector was twofold. It demonstrated that an independent lab could compete at the frontier without the resources of a major internet platform, encouraging a wave of new Chinese AI startups in 2025. It also entrenched the Chinese sector's preference for open weights distribution, a pattern that distinguishes the Chinese frontier from the largely closed United States frontier and that continues with Qwen 3, Hunyuan, and others. A 2025 Stanford HAI and DigiChina issue brief characterised the result as a "diverse open-weight AI ecosystem" in which open release, rather than closed productisation, had become the default Chinese strategy.^[26]

Surveillance, public sector, and other applications

Chinese AI capacity has been deployed in domains beyond consumer chatbots. Computer vision and facial recognition systems from SenseTime, Megvii, Yitu, CloudWalk, Hikvision, and Dahua are used widely by Chinese police and municipal governments. Independent research, including the work of the Australian Strategic Policy Institute, Human Rights Watch, and academics such as Darren Byler, has documented the use of these systems in mass surveillance programs in Xinjiang. United States Entity List rules added Dahua, Hikvision, Megvii, SenseTime, and Yitu in October 2019, and CloudWalk in June 2020, citing their roles in human rights violations or abuses in Xinjiang.^[41]

Autonomous driving is a major Chinese AI application. Baidu Apollo, Pony.ai, WeRide, AutoX, Didi Autonomous Driving, and Huawei's intelligent driving solutions all operate testing fleets, and several have launched commercial robotaxi services in restricted operational design domains. Chinese electric vehicle makers (BYD, Nio, XPeng, Li Auto, Zeekr) deploy advanced driver assistance systems with increasingly large neural networks; XPeng's XNGP and Huawei's ADS 3.0 use end to end learned planners derived from Tesla style architectures.

In robotics, Unitree, EngineAI, Robot Era, Booster Robotics, AgiBot, and other firms produce humanoid robots and quadrupeds. Unitree's H1 humanoid (announced August 2023) and G1 humanoid (announced May 2024) are sold to research and industrial customers. Chinese central and provincial governments announced humanoid robot industrial development plans in late 2023 and 2024 setting targets for 2025 and 2027. According to Stanford's 2026 AI Index, China's industrial robot installation rate ran roughly nine times that of the United States.^[26]

In drug discovery, the AlphaFold family from Google DeepMind has been complemented by Chinese efforts at companies including XtalPi and Insilico Medicine, with public sector work at the Beijing Genomics Institute and the Chinese Academy of Sciences.

Geopolitical context

Chinese AI development is embedded in a larger competitive relationship between China and the United States. The 2017 development plan was widely interpreted in Washington as evidence of strategic intent, and prompted a series of United States measures including the 2018 Foreign Investment Risk Review Modernization Act, the establishment of the Joint AI Center at the Department of Defense, the 2020 American AI Initiative executive order, the export controls described above, and the October 30, 2023 Executive Order on AI by the Biden administration. The Trump administration's January 23, 2025 Executive Order on Removing Barriers to American Leadership in Artificial Intelligence revoked parts of the 2023 order while continuing the export control framework.

At the multilateral level China signed the 2023 Bletchley Declaration, the 2024 Seoul Declaration, and supported the 2024 United Nations General Assembly resolution on AI. China has also pursued bilateral AI dialogues with the United States (formal track 1 talks were held in Geneva in May 2024 and the autumn of 2024) and the European Union, alongside its own outreach through the Belt and Road Initiative and the Global AI Governance Initiative announced by Xi Jinping in October 2023.

The Chinese position in international AI governance forums has emphasised state sovereignty over data and algorithms, opposition to United States led restrictions on technology transfer, and support for inclusive development for the Global South. United States officials and several Western governments have argued that Chinese AI is enabling state surveillance and military modernisation. Both framings shape policy decisions, and the underlying technical trajectory continues largely independent of either rhetorical position.

See also

• DeepSeek
• Qwen
• ERNIE
• Hunyuan
• Doubao
• Kimi
• Pangu
• Ascend
• SMIC
• Cambricon
• Tsinghua University
• BAAI
• Wudao
• Made in China 2025
• Export controls
• Stanford AI Index

References

1. ^State Council of the People's Republic of China. Notice of the State Council Issuing the New Generation Artificial Intelligence Development Plan. Guofa No. 35, July 2017.
2. ^Graham Webster, Rogier Creemers, Paul Triolo, and Elsa Kania. Full Translation: China's New Generation Artificial Intelligence Development Plan (2017). New America, August 1, 2017.
3. ^Cyberspace Administration of China. Interim Measures for the Management of Generative Artificial Intelligence Services and Measures for Labelling AI-Generated and Synthetic Content. 2023 and 2025.
4. ^Cyberspace Administration of China. Provisions on the Administration of Deep Synthesis Internet Information Services. December 11, 2022.
5. ^Cyberspace Administration of China. Provisions on the Administration of Algorithmic Recommendations of Internet Information Services and Summary of the provisions. January to March 2022.
6. ^National Technical Committee 260 on Cybersecurity. AI Safety Governance Framework Version 1.0. September 2024.
7. ^Stanford Institute for Human-Centered Artificial Intelligence. Artificial Intelligence Index Report 2025. April 2025.
8. ^Stanford Institute for Human-Centered Artificial Intelligence. Artificial Intelligence Index Report 2024. April 2024.
9. ^DeepSeek AI. DeepSeek-V3 Technical Report. arXiv:2412.19437, December 2024.
10. ^DeepSeek AI. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning. arXiv:2501.12948, January 2025.
11. ^Bureau of Industry and Security, US Department of Commerce. Implementation of Additional Export Controls: Certain Advanced Computing and Semiconductor Manufacturing Items. October 2022.
12. ^Bureau of Industry and Security, US Department of Commerce. Implementation of Additional Export Controls Update and Commerce Makes Revisions to the Entity List to Strengthen U.S. National Security. October 2023 and January 2025.
13. ^Bureau of Industry and Security, US Department of Commerce. Framework for Artificial Intelligence Diffusion. January 2025.
14. ^Center for Security and Emerging Technology. Keeping Top AI Talent in the United States and Assessing China's AI Workforce. 2019 and 2023.
15. ^Wendy Chang, Rebecca Arcesati, and Antonia Hmaidi. China's Drive Toward Self-Reliance in Artificial Intelligence. MERICS, July 2025.
16. ^Jeffrey Ding. ChinAI Newsletter archive. 2018 to present.
17. ^Kaifeng Bi et al. Accurate medium-range global weather forecasting with 3D neural networks. Nature, July 2023.
18. ^Jacob Feldgoise and Hanna Dohmen. Pushing the Limits: Huawei's AI Chip Tests U.S. Export Controls. CSET, June 17, 2024.
19. ^Reuters. China's DeepSeek sets off AI market rout. January 27, 2025.
20. ^Centre for Emerging Technology and Security. Seeking Deeper: Assessing China's AI Security Ecosystem. 2025.
21. ^South China Morning Post. China's four new AI tigers emerge as investor favourites. April 19, 2024.
22. ^Wall Street Journal. The Underground Network Sneaking Nvidia Chips Into China. July 3, 2024.
23. ^Bloomberg. Tencent Joins $300 Million Financing for China's AI Unicorn. August 5, 2024.
24. ^TechInsights. SMIC N+2 7nm, Huawei Mate 60 Pro; ASML. Statement regarding Dutch government's updated export license requirement. 2023 and 2024.
25. ^Epoch AI. What went into training DeepSeek-R1? and DeepSeek-V3 model record; MERICS. China's Drive Toward Self-Reliance in Artificial Intelligence. 2025.
26. ^Stanford Institute for Human-Centered Artificial Intelligence. Artificial Intelligence Index Report 2026 and Technical performance chapter; Stanford HAI and DigiChina. Beyond DeepSeek: China's Diverse Open-Weight AI Ecosystem and Its Policy Implications. 2025 and 2026.
27. ^DeepSeek AI. Model cards for DeepSeek-V3, DeepSeek-R1, and DeepSeek-V3.2-Exp. 2024 to 2025.
28. ^Bureau of Industry and Security. AI Diffusion Rule rescission and chip-related controls and General Prohibition 10 guidance; NVIDIA. NVIDIA CEO Promotes Global Access to American AI in Washington and China and 2026 annual report. 2025 to 2026.
29. ^Rui Ma. 8 Chinese companies released 10 AI models in 75 days. X, August 14, 2026.
30. ^MiniMax. MiniMax M3: Frontier Coding, 1M Context, Native Multimodality and MiniMax-M3 commit history. June 1, 2026; repository history accessed August 15, 2026.
31. ^Jie Tang. GLM-5.2 Coding Plan availability announcement; Z.ai. GLM-5.2 model card and GLM-5.2 technical article. June 13, June 16, and June 17, 2026.
32. ^Meituan. LongCat-2.0 model and API announcement, LongCat-2.0 open-source announcement, and LongCat-2.0 model card. June 30 and July 6, 2026.
33. ^Tencent. Tencent Hunyuan Officially Releases Hy3 and Hy3 model card. July 6, 2026.
34. ^Moonshot AI. Kimi K3 technical blog and Kimi K3 model card. July 16 and July 27, 2026.
35. ^Dreamina. Giving creators more control with Dreamina Seedance 2.5. July 31, 2026.
36. ^DeepSeek AI. Release update log and DeepSeek-V4-Flash-0731 model card. April 24 and July 31, 2026.
37. ^Qwen. Qwen3.8-Max-Preview announcement, Qwen3.8-Max: A New Bar for Coding and Cowork, and Qwen3.8-2.4T-A95B model card. July 19, August 3, and August 12, 2026.
38. ^DeepSeek AI. Release update log and DeepSeek-V4-Pro-0813 model card. April 24 and August 13, 2026.
39. ^Z.ai. GLM-5.3 technical post and GLM-5.3 developer documentation. August 14, 2026.
40. ^National People's Congress of China. Report on handling representatives' legislative proposals. Republished by the National Office for Philosophy and Social Sciences, March 2026.
41. ^Bureau of Industry and Security, US Department of Commerce. Addition of Certain Entities to the Entity List and Addition of Entities to the Entity List, Revision of Certain Entries, and Removal of Entities from the Entity List. October 2019 and June 2020.
42. ^Bureau of Industry and Security, US Department of Commerce. Commerce Strengthens Export Controls to Restrict China's Capability to Produce Advanced Semiconductors for Military Applications. December 2, 2024.