JASON HSU2026年7月24日 Mikey BurtonThe Chinese government has spent three years giving away its most capable artificial intelligence to anyone to use. That generosity was always temporary. Chinese authorities have reportedly been meeting with Alibaba, ByteDance and the start-up Z.ai to discuss restricting overseas access to some of China’s most advanced models — including systems that have not yet been released.三年来，中国政府一直把它最强大的人工智能免费向所有人开放。然而，这样的慷慨从来都是不会持久的。据报道，中国当局一直在与阿里巴巴、字节跳动以及初创公司智谱AI会面，讨论限制境外用户访问中国部分最先进模型的问题——包括一些目前尚未发布的系统。The U.S. government’s instinct will likely be to answer with a ban of its own. That would be a mistake. The right response is the opposite: America and its allies should build and deploy their own so-called open-weight models, which users can download, run and modify on their machines. They should do so at scale, and commit to releasing new and better versions on a dependable schedule — so that anyone building on an American model can count on the next one arriving. China’s move is an opening for the United States to take the lead again on A.I. It will not stay open for long.美国政府的本能反应很可能是以自己的禁令作为回应。那将是一个错误。正确的做法恰恰相反：美国及其盟友应该构建并部署自己的所谓“开放权重”模型，使用户能够将模型下载到本地，在自己的设备上运行并进行修改。而且，这项工作应当大规模推进，并承诺按照可靠的周期发布更为强大的新版本。让所有基于美国模型进行开发的人都能确信，下一代模型会如期推出。中国的这一举动正是美国重夺人工智能主导权的一个机会窗口。但这个窗口不会开太久。An A.I. model’s “weights” are the billions of individual settings acquired and fine-tuned during training that help determine how a model responds to prompts. When a company keeps its weights secret, as OpenAI does with GPT, other companies and developers can rent access through its servers — but they cannot look inside or take the model with them. When a company publishes these weights, as the Chinese company DeepSeek has for its A.I. model, any user can download and run the weights on its own hardware, inspect them and retrain them for a specialized task.人工智能模型的“权重”是指训练过程中获得并微调好的数十亿个独立参数，它们有助于决定模型如何对提示词作出回应。当一家公司对其权重保密时——比如OpenAI对GPT的做法——其他公司和开发者只能通过其服务器租用访问权限，但无法查看模型内部，也无法将模型带走。当一家公司公开这些权重时——比如中国公司DeepSeek，任何用户都可以下载这些权重，在自己的硬件上运行、检查它们，并针对特定任务进行重新训练。That is how China is arguably winning the A.I. race. Last week, the Chinese start-up Moonshot A.I. unveiled Kimi K3, which the company says is the world’s largest open-weight A.I. model and rivals the performance of Anthropic and OpenAI’s best models to date. Alibaba’s Qwen is now the most downloaded A.I. model family on earth. Developers have used it to build more than 100,000 specialized models of their own. On OpenRouter, a service that lets developers send their work to whichever A.I. model they choose, Chinese open-weight models accounted for more than 20 percent of the units of text processed over a number of weeks in 2025 and more than 40 percent over a few weeks this spring; in 2024, such models’ weekly processing of text was as low as about 1 percent.可以说，这正是中国在人工智能竞赛中占据优势的原因。上周，中国初创公司月之暗面推出了Kimi K3，该公司称，这是目前全球规模最大的开放权重人工智能模型，其性能足以媲美Anthropic和OpenAI迄今最优秀的模型。阿里巴巴的通义千问如今已成为全球下载量最高的人工智能模型家族。开发者已利用它构建了超过10万个各类专用模型。在OpenRouter——一个允许开发者将任务发送给任选人工智能模型的服务平台——上，来自中国的开放权重模型在2025年连续数周处理的文本量已占全部处理量的20%以上，而在今年春季的连续数周时间里，这一比例更是超过40%；相比之下，2024年，这类中国模型每周处理的文本量占比最低时仅约为1%。Much of that traffic is American. Airbnb’s customer service agent runs substantially on Qwen; Airbnb’s chief executive, Brian Chesky, has called the model fast and cheap. Cursor, among the most valuable A.I. coding start-ups in the country, built its Composer 2 model on a base licensed from Moonshot. DoorDash now routes the “lower-level work” of its code review system to Kimi, DoorDash’s co-founder Andy Fang reports, reserving an American model for the hardest tasks. More American businesses, attracted by lower costs, are joining these start-ups in using these models.这些流量中，很大一部分来自美国。Airbnb的客服系统大量运行于通义千问之上，其首席执行官布赖恩·切斯基称赞该模型速度快、成本低。美国最具价值的人工智能编程初创企业之一Cursor则基于月之暗面授权的基础模型开发了自己的Composer 2模型。外送平台DoorDash联合创始人方安迪（音）透露，如今公司已将代码审查系统中的“较低层级工作”交由Kimi完成，只将最困难的任务交由美国模型来做。受成本优势吸引，越来越多的美国企业正在跟随这些初创公司的脚步，转向使用这些中国模型。This had been China’s plan since 2017: push Chinese models into every market, shape the standards others adopt and pull the world toward the Chinese hardware the models run best on. The country ran the same sequence with solar and telecom equipment and, most notably, critical minerals — building the world’s dependence on cheap rare earths before restricting exports once that dependence was real.这正是中国自2017年以来一直奉行的战略：将中国的人工智能模型推向全球各个市场，塑造其他国家所采用的技术标准，并进一步推动世界转向那些最适合运行中国模型的中国硬件。中国此前在太阳能产业、通信设备领域，尤其是在关键矿产领域，都采取过同样的路径——先利用廉价稀土建立起全球依赖，再在这种依赖真正形成之后限制出口。If Chinese labs restrict access now, won’t everyone just switch to a better American model? No, because switching an A.I. model is not like canceling a subscription. It means tearing out a foundation. A European manufacturer running a Chinese model on its own premises, trained on confidential records that cannot legally leave the building, cannot simply swap in a superior American model that costs more to use and runs on someone else’s servers. That dependence is key to China’s strategy. A country that can’t switch off a foreign model has handed China a grip on the systems it runs on.如果中国实验室现在限制访问，难道大家不会干脆转向一个性能更好的美国模型吗？不会，因为切换人工智能模型不像取消订阅那么简单。这意味着要拆除整个底层基础。一家欧洲制造商如果在本地运行一个中国模型，并用依法不得出境的机密数据对其进行了训练，就无法简单地换入一个使用成本更高、运行在他人服务器上的美国优质模型。这种依赖正是中国战略的关键所在。一个无法摆脱外国模型的国家，就等于把其所运行系统的控制权拱手交给了中国。Developers outside China should be concerned about what’s next. Everyone building on a Chinese model assumes a better one is coming — free, as it always has been. If China keeps its next generation of models at home, the rest of the world’s builders who rely on something freely modifiable are stuck with an antiquated version, while Chinese firms advance with something more powerful. The gap widens with every release.中国以外的开发者应当对接下来的走势保持警觉。每一个基于中国模型进行开发的人都默认——未来会有更好的版本推出，并且像以往一样免费。如果中国将下一代模型留在国内，那么世界上那些依赖免费可修改模型的开发者将被困在一个过时的版本里，而中国企业则在更强大版本的加持下不断向前。双方之间的差距将随着每一次新版本发布而不断扩大。Two U.S. congressional committees have opened investigations into Airbnb and Cursor’s parent company over their use of Chinese models. The Trump administration is reportedly considering effectively barring Americans from using Chinese A.I. models. But banning those models would impede American developers and hand China the claim that its technology is too good for America to compete with fairly. The United States dominates in tech by making its platforms the ones the world’s developers choose to build on, not by banning foreign use.两个国会委员会已就Airbnb和Cursor的母公司使用中国模型一事展开调查。据报道，特朗普政府正在考虑有效禁止美国人使用中国人工智能模型的措施。但是，禁止这些模型将给美国开发者带来不便，也会给中国提供口实，宣称它的技术先进到美国都无法通过公平竞争来应对。美国在科技领域的主导地位，靠的是让全球开发者都愿意基于美国的平台进行开发，而不是通过禁止外国使用来达成的。The U.S. government should fund a contract to train an advanced open-weight model. It should stipulate those weights remain permanently open. That contract would give companies in the United States and allied nations a foundation to build on and compete in this arena of A.I. development. OpenAI and Google have released open-weight models, but only occasionally, because nothing rewards them for doing it consistently. The buyers of such a model would be the firms already reaching for Qwen and Kimi as well as the international partners that will not put their data on Chinese servers.美国政府应当出资建立一个训练先进开放权重模型的项目，并规定这些权重必须永久保持开放。这将为美国及其盟友国家的企业提供一个共同的基础平台，使它们能够在人工智能发展的这一领域展开创新与竞争。OpenAI和Google都发布过开放权重模型，但只是偶尔为之，因为没有任何机制能激励它们持续这样做。这类模型的潜在用户不仅包括那些已经开始采用通义千问和Kimi的企业，也包括那些不愿把数据存放在中国服务器上的国际合作伙伴Customers that adopt an open-weight American model also need the chips to run it, somewhere to store data and the computing power to put the model to work. That is what the Development Finance Corporation and the Export-Import Bank — the government agencies that back U.S. business abroad — should be financing all at once. Chinese models have come to dominate in places like Nairobi and São Paulo because China made them free, packaging the model, the hardware and the financing together. America has the capital and the technology to offer the same package, on better terms, with security standards that even newcomers can meet.采用美国开放权重模型的用户同样需要运行模型所需的芯片、存储数据的基础设施，以及让模型真正投入使用所需的算力。这正是美国国际开发金融公司和美国进出口银行——这两个负责支持美国企业开展海外业务的政府机构——应当同时提供融资支持的方向。中国的模型之所以能够在内罗毕、圣保罗等地占据主导地位，是因为中国免费提供模型，并将模型、硬件和融资打包在一起。美国拥有足够的资金和技术，完全能够提供同样的整体方案，而且条件更优、安全标准更高，即使是刚起步的国家也能够达到这些标准。OpenAI and Anthropic’s business is the “frontier”: the hardest, highest-value work, offered to customers that want the best model available and will pay for it. That is not the business an open-weight model takes, where finances or data-sensitivity factor in more heavily. For many U.S. customers, the frontier model keeps the hard work moving, and something cheaper — and easier to tinker with — sits underneath it. The frontier lab gets paid either way.OpenAI和Anthropic的商业模式立足于“前沿”，也就是承担最困难、价值最高工作的模型，服务对象是那些希望获得当前最优秀模型、并愿意为此付费的客户。开放权重模型面向的并非同一市场——在那里，财务成本或数据敏感性的考量权重更大。对许多美国用户而言，前沿模型负责完成最复杂的任务，而底层则运行着一种更便宜、也更容易进行定制修改的模型。无论哪种情况，前沿模型开发企业都能够获得收益。The only question is whose model is underneath. It should be an American one.唯一的问题是：那个底层模型来自哪里。答案应该是美国。Jason Hsu是哈德逊研究所的高级研究员。翻译：纽约时报中文网点击查看本文英文版。