去年中国1.2万亿美元的商品贸易顺差使人们关注电动汽车、电池、太阳能电池板和工业产能过剩。随着贸易伙伴对中国制造规模做出反应，关税、反补贴调查和本地化要求随之而来。
China’s US$1.2 trillion goods trade surplus last year has focused attention on electric vehicles, batteries, solar panels and industrial overcapacity. Tariffs, anti-subsidy investigations and localisation requirements have followed as trading partners respond to China’s manufacturing scale.
然而，服务业正在发生这种不太明显的转变。2026年前7个月，服务业出口同比增长17.1%，个人文化娱乐服务出口激增57.8%。中国服务业贸易逆差仍远未取代制造业。但数字服务可以在没有工厂、集装箱或港口的情况下在全球范围内扩展。
Yet the less visible shift is happening in services. Services exports rose by 17.1 per cent year on year in the first seven months of 2026, while exports of personal cultural and entertainment services surged by 57.8 per cent. China still runs a trade deficit in services, which are nowhere close to replacing manufacturing. But digital services can scale up globally without factories, containers or ports.
人工智能（AI）生成的微剧提供了一个早期的例子。中国在国内实现了这种形式的工业化：一到三分钟的剧集、快速的情节曲折以及围绕智能手机观看和支付而设计的悬念。该行业的收入从2020年的不到10亿元人民币增长到去年的估计1000亿元人民币（149.1亿美元）。ReelShort等中国支持的平台已经拥有大量海外受众。
Microdramas generated by artificial intelligence (AI) offer an early example. China industrialised the format domestically: one-to-three-minute episodes, rapid plot twists and cliffhangers designed around smartphone viewing and payment. The sector’s revenue rose from less than 1 billion yuan in 2020 to an estimated 100 billion yuan (US$14.91 billion) last year. Chinese-backed platforms such as ReelShort already have sizeable overseas audiences.
这些公司不是简单地出口中国电视，而是出口围绕观众数据、快速本地化、表演营销以及越来越多的人工智能构建的制作系统。
Rather than simply exporting Chinese television, these companies are exporting a production system built around audience data, rapid localisation, performance marketing and, increasingly, AI.
中国工厂通过缩短产品周期、压缩成本、生产多种变体以及快速扩大畅销产品的规模而变得强大。人工智能允许内容制作者将相同的逻辑应用于娱乐。2026年前8个月，中国发布了约43万部微剧，是2025年全年的13倍，其中人工智能占比超过90%。
Chinese factories became formidable by shortening product cycles, compressing costs, producing many variants and quickly scaling up whatever sold best. AI allows content producers to apply the same logic to entertainment. China released about 430,000 microdramas in the first eight months of 2026, as much as 13 times all of 2025, with AI accounting for over 90 per cent.
生产力的提高解释了这种快速采用的原因。韩国生产商Vigloo使用人工智能将产量从三个月削减至一个月，并将成本削减至约五分之一。使用字节跳动Seedance的深圳工作室也实现了类似的压缩。但优势不仅仅在于更便宜的视频生成。中国拥有一个能够快速、大规模地将人工智能输出转化为商业内容的生态系统。这一点很重要，因为微剧是为快速实验而设计的。
The productivity gains explain the rapid adoption. South Korean producer Vigloo cut production from three months to one and costs to about one-fifth using AI. Shenzhen studios using ByteDance’s Seedance have achieved a similar compression. But the advantage is not just cheaper video generation. China has an ecosystem capable of turning AI output into commercial content quickly and at scale. That matters because microdramas are built for rapid experimentation.
制片人可以在开场节目中看到观众是否留下来，然后将营销支出转向获得吸引力的标题，并放弃那些没有吸引力的标题。更低的成本也意味着更多的赌注。虽然传统工作室可能会花几个月的时间开发几个昂贵的项目，但人工智能微剧制作人可以推出许多作品，快速阅读观众行为并只扩大获胜者的规模。
Producers can see within the opening episodes whether viewers are staying, then redirect marketing spending towards titles gaining traction and abandon those that do not. Lower costs also mean more bets. While a traditional studio may spend months developing several expensive projects, an AI microdrama producer can launch many titles, read audience behaviour quickly and scale up only the winners.
突破性的成功很少--2026年上半年抖音上发布的人工智能微剧中，只有0.48%的观看量超过1亿次。尽管如此，微剧可能是比投入数百万美元制作长篇作品更好的风险模型：微剧让失败更容易被发现。
Breakout successes are few and far between – only 0.48 per cent of AI microdramas released on Douyin in the first half of 2026 exceeded 100 million views. Still, microdramas may be a better risk model than committing millions to a long-form production: microdramas make failures far easier to discover.
然而，规模本身也带来了问题。北京目前正在推动该行业提高质量、原创性和版权保护，有效地告诉这个为速度而建的行业放慢脚步。
Yet scale is creating its own problems. Beijing is now pushing the industry towards greater quality, originality and copyright protection, effectively telling a sector built for speed to slow down.
爱奇艺展示了这不再是边缘格式。围绕优质长篇剧打造的爱奇艺，其短篇剧在今年第二季度首次占据国内市场份额，尽管集团收入同比下降了5%。它的回应是深入推进人工智能和简短生产。微剧可能无法解释其更广泛的收入压力，但它们正在重塑其经济学。
iQiyi shows how this is no longer a fringe format. Built around premium long-form drama, iQiyi saw its short-form titles take the top domestic market share for the first time in the second quarter this year, even as group revenue fell by 5 per cent year on year. Its response has been to push deeper into AI and short-form production. Microdramas may not explain its wider revenue pressure, but they are reshaping the economics of it.
微剧不仅仅是一种时尚，更是一个测试案例。它们展示了中国的工业优势如何应用于数字服务，在没有定义实体贸易的工厂、航线和关税风险的情况下创造全球规模的出口。一部微剧通过应用程序跨越国界。即使人工智能降低了翻译、配音和本地化的成本，但没有容器需要检查，边际发行成本几乎为零。
More than a fad, microdramas are a test case. They show how China’s industrial strengths can be applied to digital services, creating exports that scale globally without the factories, shipping lanes and tariff exposure that define physical trade. A microdrama crosses borders through an app. There is no container to inspect and nearly zero marginal distribution cost, even as AI lowers the cost of translation, dubbing and localisation.
数字产品仍然可能面临限制。TikTok的案例展示了国家安全和平台监管的效果如何替代关税的效果。但当出口产品是娱乐、软件、知识产权或注意力时，为制造业产能过剩而设计的政策就不那么有效了。
Digital products can still face restrictions. TikTok’s case shows how the effect of national security and platform regulation can substitute for those of tariffs. But policies designed for manufacturing overcapacity work less neatly when the export is entertainment, software, intellectual property or attention.
中国正在将熟悉的制造能力应用于服务业：速度、规模、成本控制、快速实验和无情淘汰弱势产品。只有输出发生了变化。
China is applying familiar manufacturing capabilities to services: speed, scale, cost control, rapid experimentation and a ruthless elimination of weak products. Only the output has changed.
娱乐增加了另一个维度，因为故事塑造了人们的看法。人工智能不仅放大了输出，还放大了其中所包含的假设和刻板印象，特别是当算法用流量奖励这些叙述时。我看过中国的微剧将美国塑造成恶霸，并用一个虚构的东南亚小国作为洗钱和腐败的代名词。重复仍然可以使这种描绘正常化。
Entertainment adds another dimension because stories shape perceptions. AI amplifies not only output, but also the assumptions and stereotypes embedded in it, especially when algorithms reward those narratives with traffic. I have seen Chinese microdramas cast the United States as a bully and use a fictional small Southeast Asian country as shorthand for money laundering and corruption. Repetition can still normalise such portrayals.
中国花了40年时间围绕实物商品建立了一个出口机器。如果同样的经济学通过游戏、在线文学、娱乐、应用程序和人工智能服务传播，中国将获得另一个全球化渠道，该渠道对工厂的依赖程度更低，更难通过传统关税解决。
China spent four decades building an export machine around physical goods. If the same economics spread through gaming, online literature, entertainment, apps and AI-enabled services, China gains another channel for globalisation that is less dependent on factories and harder to address with conventional tariffs.
服务业不会很快取代中国的制造业引擎。相反，制造业的竞争能力正开始转移到服务业。
Services will not replace China’s manufacturing engine any time soon. Rather, the competitive capabilities built in manufacturing are beginning to migrate into services.
中国数万亿美元的商品盈余在港口可见。其下一波出口浪潮可能会越来越多地通过智能手机抵达。
China’s trillion-dollar goods surplus is visible at the port. Its next export wave may increasingly arrive through the smartphone.