小 i 导 读
2026年7月17日至20日，世界人工智能大会（WAIC）在上海召开，超1100家全球企业参展，共带来4400余项展品。同期，中国“月之暗面”公司发布Kimi K3开源模型，部分性能领先或持平先进模型，引发美股震荡。对此，北京大学国际关系学院教授、北京大学中外人文交流研究基地执行主任王栋与北京大学中外人文交流研究基地研究助理张学玉在文章中指出：中国人工智能竞争力已不再仅仅体现为单项技术突破，而是在技术研发、产业应用和治理体系相互支撑下，逐渐形成一种具有持续竞争优势的“体系能力”。文章从技术创新、大规模应用和治理能力三个层面分析，并提出三个值得关注的趋势。
本文约3400字，读完约11分钟
In mid-July, the World Artificial Intelligence Conference took place in Shanghai, China, under the theme “Intelligent Partnership for a Shared Future.” More than 1,100 companies from around the world took part, showcasing some 4,400 exhibits, including 108 Chinese-developed AI chips and 261 large language models.
2026年7月17日至20日，世界人工智能大会（WAIC）在中国上海召开。大会以“智能伙伴 共创未来”为题，超1100家全球企业参展，共带来4400余项展品，其中展示了108款中国国产芯片与261款大模型。
Over the past few years, international attention to China’s AI development has largely centered on a single question: Can China catch up with the United States in foundational models? As home-grown models such as DeepSeek and Kimi have advanced with remarkable speed and AI technologies have increasingly found commercial and industrial applications, a more consequential shift has begun to emerge. China’s competitiveness in artificial intelligence is no longer defined solely by breakthroughs in individual technologies. Instead, it is increasingly rooted in a self-reinforcing ecosystem in which technological innovation, industrial deployment and governance mutually reinforce one another, giving rise to a durable systemic advantage.
过去几年，国际社会观察中国人工智能发展，往往聚焦于一个问题：中国能否追赶美国最先进的大模型。伴随DeepSeek、Kimi等国产模型快速迭代，以及人工智能技术不断走向产业应用，一个更加值得关注的变化正在出现：中国人工智能竞争力已不再仅仅体现为单项技术突破，而是在技术研发、产业应用和治理体系相互支撑下，逐渐形成一种具有持续竞争优势的“体系能力”。
This systemic strength is underpinned by China’s steadily expanding capacity for technological innovation. In recent years, the country’s development of foundational models has entered a period of rapid acceleration. Homegrown models such as DeepSeek, Kimi, Tongyi Qianwen and ERNIE have continued to narrow the gap with the world’s leading AI systems, demonstrating strong competitiveness in reasoning performance, development efficiency and cost optimization. In July, Moonshot AI unveiled its open-source Kimi K3 model. It outperformed Anthropic's Claude Fable 5 in long-context coding tasks and online information retrieval and matched OpenAI's flagship GPT-5.6 Sol in scientific reasoning. It sent ripples through U.S. equity markets.
这种体系能力建立在不断增强的技术创新能力之上。近年来，中国大模型研发进入快速发展阶段，DeepSeek、Kimi、通义千问、文心一言等模型不断缩小与国际领先模型间的差距，在推理能力、开发效率和成本控制等方面展现出较强竞争力。2026年7月17日，中国“月之暗面”公司发布Kimi K3开源模型，在长程序编写任务和在线信息检索能力上领先于Anthropic模型Claude Fable 5，在科学推理能力方面与OpenAI顶尖模型GPT-5.6 Sol持平，引发美股震荡。
2026世界人工智能大会暨人工智能全球治理高级别会议于7月17日至20日在上海举行。
Meanwhile, successive rounds of U.S. export controls on advanced chips have failed to derail China’s AI development; they have only accelerated the country’s drive for technological self-reliance. From the development of indigenous AI chips and the expansion of national computing hubs to the rollout of new digital infrastructure projects—for example, the “East Data, West Computing” initiative—China is steadily building an AI ecosystem that integrates semiconductors, computing power, electricity and data resources. In the first half of this year, China’s output of integrated circuits by major industrial enterprises rose by 23.1 percent year-on-year to an average daily production of more than 1.5 billion chips. At the 2026 WAIC in Shanghai, Huawei unveiled its Ascend 950 Super Node, which links 1,024 Ascend processors through a high-bandwidth connection to deliver substantially greater computing power. The system has already reached performance parity with some of Nvidia’s leading AI platforms.
与此同时，美国持续升级的芯片出口管制政策并未阻断中国人工智能发展，反而推动了中国技术自研能力不断提升。从国产AI芯片研发，到全国智能算力中心布局，再到“东数西算”等新型基础设施建设，中国正在逐步完善覆盖芯片、算力、电力和数据资源的人工智能基础设施。2026年上半年，中国规模以上工业集成电路产量同比增长23.1%，平均每天生产超15亿块。2026年上海世界人工智能大会上，华为公司发布昇腾950超节点，采用宽带集合1024颗昇腾芯片以实现算力提升，性能方面已经赶上部分英伟达领先芯片。
If technological innovation determines the upper limits of a country’s AI capabilities, large-scale deployment determines how much economic value those capabilities can ultimately create. In this respect, one of China’s distinctive strengths lies in its ability to translate technological advances into widespread industrial applications.
如果说技术创新决定了一国人工智能的发展上限，大规模应用则决定着人工智能能够创造出多少真实价值，而这正是中国当前人工智能最具特色的优势之一。
In recent years, China’s foundational models have been deployed across manufacturing, finance, healthcare, education and transportation, accelerating the digital and intelligent transformation of traditional industries. The country’s vast digital economy and extensive internet ecosystem have also provided fertile ground for AI experimentation and commercialization. From AI-generated video and intelligent office software to smart devices, embodied AI and humanoid robots, a number of innovative companies are building vertical applications on top of the models, steadily expanding the commercial frontiers of AI. Meanwhile, the rapid improvement of open-source models is lowering barriers to adoption and enabling more small and medium-sized enterprises to develop industry-specific solutions based on mature AI platforms. As a result, AI innovation is spreading beyond a handful of technology giants into a much broader industrial ecosystem. According to the International Data Corporation, China’s AI investment is projected to reach $26.69 billion this year, with spending on AI services expected to exceed $4 billion.
近年来，中国大模型正广泛赋能制造、金融、医疗、教育、交通等领域，推动传统产业数字化、智能化升级。中国庞大的数字经济市场和丰富的互联网应用生态，为人工智能技术提供了广阔的实验场。从视频生成、智能办公，到智能终端、具身智能、人形机器人等新兴产业，大量创新企业围绕基础模型开发垂直应用，拓宽人工智能商业边界。开源模型能力持续提升，也使越来越多中小企业能够基于成熟模型构建行业解决方案，使人工智能创新从少数头部企业扩散到更广泛的产业生态中。国际数据中心（IDC）预测，中国2026年AI投资规模有望达到266.9亿美元，服务市场总投资预计规模将超过40亿美元。
Finally, China's evolving AI ecosystem is underpinned by a steadily improving governance framework. While artificial intelligence is an increasingly important driver of economic growth, it also brings new challenges, including algorithmic bias, privacy breaches, misleading content, malicious use and disruptions to labor markets. Addressing these risks has become an integral part of China’s AI strategy. At the WAIC event this year, President Xi Jinping noted that artificial intelligence “carries within it great opportunities as well as challenges to governance” and stressed the need to “take seriously the various types of inherent and secondary risks that AI may trigger.” In recent years, China has consistently pursued an approach that gives equal weight to development and security. It has gradually strengthened the regulatory framework governing generative AI, deep synthesis technologies and recommendation algorithms, while building a more comprehensive governance system spanning technological innovation, industrial deployment and risk management. At the same time, the country continues to expand investment in talent development, scientific research and long-term industrial planning. The outline of its 15th Five-Year Plan identifies artificial intelligence as a key area for seizing the opportunities created by the latest technological and industrial transformation. It calls for deepening the “AI Plus” initiative, strengthening systemic innovation capabilities, reinforcing the role of enterprises as the principal drivers of technological innovation and advancing the integrated development of education, sci-tech and talent cultivation. Together, these measures are intended to provide a stable policy environment, a stronger talent pipeline and sustained institutional support for the long-term development of AI.
最后，支撑中国人工智能体系能力持续建设的，还有不断完善的治理能力。人工智能既是推动经济增长的重要动力，也会带来算法偏见、隐私泄露、虚假内容生成、恶意使用与滥用、就业市场冲击等新风险。在本次WAIC峰会上，习主席强调人工智能“既蕴含巨大机遇，也面临治理挑战”，“要高度重视人工智能引发的各类内生与衍生风险”。近年来，中国始终强调人工智能“发展与安全并重”，持续完善生成式人工智能、深度合成、算法推荐等领域监管规则，逐步健全覆盖技术研发、产业应用到风险治理的综合框架。同时，中国不断加强人才培养、科研投入和产业规划，在“十五五”规划纲要中提出要抓住科技革命与产业变革历史机遇，深化拓展人工智能+，提高体系化创新能力，强化企业科技创新主体地位，并一体化推进教育科技人才发展，为人工智能长期发展提供稳定环境、政策、人才支撑。
7月17日，参观者体验与机器人踢点球。
China is also playing an increasingly active role in shaping global AI governance. As both a rising technological power and a member of the Global South, it has sought to balance two priorities: managing strategic competition and AI-related risks with other leading technology powers, particularly the United States, while addressing the widening AI divide between developed and developing economies and safeguarding the technological sovereignty of emerging countries. At the opening ceremony of WAIC, President Xi announced that China would provide 5,000 AI training opportunities for participants from developing countries over the next five years and establish international AI cooperation centers with partners including the Association of Southeast Asian Nations, the League of Arab States and the African Union. Since unveiling the Global AI Governance Initiative in 2023 and promoting international AI cooperation under the framework of the United Nations, China has steadily expanded technology training, capacity-building programs and digital infrastructure cooperation for developing countries. It launched the AI Capacity-Building Action Plan and this year established the World AI Cooperation Organization (WAICO). Through these initiatives, China aims to make the benefits of AI development more broadly accessible to developing economies.
中国也越来越积极参与全球人工智能治理，基于“技术强国”与“全球南方国家”双重身份，一方面注重与美国等技术强国做好竞争管控与风险防范，另一方面关注人工智能“南北鸿沟”与技术主权问题。在WAIC开幕式现场，习主席郑重宣布，未来五年中国将面向发展中国家提供5000个人工智能专题研修名额，并面向东盟、阿盟、非盟等建设国际人工智能合作中心。从2023年发布《全球人工智能治理倡议》，推动联合国框架下人工智能国际合作，到加强面向全球南方的技术培训、能力建设和基础设施合作，提出并实施《人工智能能力建设普惠计划》，再到于2026年正式成立世界人工智能合作组织（WAICO），中国正努力推动人工智能发展成果惠及更多发展中国家。
Looking ahead, three trends are likely to shape the next stage of China’s AI development:
展望未来，中国人工智能发展将呈现三个值得关注的趋势。
First, innovation will increasingly shift from advances in foundational models to deeper industrial deployment. Embodied AI, humanoid robots and other emerging applications are likely to become new engines of growth, while governments, investors and industry will continue to accelerate the integration of AI with the real economy. Second, competition between China and the United States will enter a more complex phase. As Chinese models continue to improve, particularly with the rapid expansion of the open-source ecosystem, Washington’s strategy of building a small yard with a high fence around frontier AI technologies—together with its efforts to establish a coalition of leading AI powers—is likely to come under growing pressure. Competition will increasingly extend beyond model performance to encompass global markets, technical standards, developer ecosystems and international rules governing AI. Third, AI will reshape not only the way economies produce but also employment patterns, labor relations and the broader social division of labor. Strengthening education and workforce training, adapting social protection systems and ensuring that the benefits of AI-driven growth are more widely shared will become central policy challenges as China moves toward a more intelligent society.
第一，人工智能创新将进一步向应用深化，具身智能、人形机器人等领域有望成为新增长点，资本、产业和政府也将持续推动人工智能与实体经济深度融合。第二，中美人工智能竞争将进入更加复杂的新阶段。伴随中国模型能力不断提升，特别是开源生态快速壮大，美国试图围绕先进人工智能建立“小院高墙”和高能力人工智能联盟的战略将面临更大挑战。未来双方竞争将不仅围绕模型性能展开，更将延伸至全球市场、技术标准、开发者生态以及国际治理规则。第三，人工智能不仅改变生产方式，也将深刻重塑就业结构、劳动关系和社会分工，如何完善教育培训体系、优化社会保障制度，使更多人能够分享人工智能带来的发展红利，将成为中国迈向智能社会必须回答的重要课题。
Ultimately, global competition in artificial intelligence will not be set by technological leadership alone. It will increasingly depend on which countries can build a complete AI ecosystem that combines technological innovation, industrial dynamism and governance resilience. China’s AI development is evolving from a strategy centered on catching up in individual technological breakthroughs to one focused on building enduring systemic capabilities. That may well prove to be the most consequential shift in China’s role in the next phase of global technological competition.
未来全球人工智能竞争，比拼的不只是技术领先，更是谁能够构建起兼具创新能力、产业活力与治理韧性的完整生态。中国人工智能的发展，正在从追赶技术突破迈向塑造体系能力，而这或许才是中国在新一轮全球科技发展中最值得关注的变化。
来源：中美聚焦（China-US Focus），原标题“China Builds a Systemic Edge in AI”
订阅