Friday, November 20, 2026

Understanding China. Seeing what’s next.

Understanding China. Seeing what’s next.

China is using 55 percent of its AI compute on micro-dramas. Why that might actually save its tech industry

短剧烧算力被批浪费?揭秘娱乐产业与科技创新的隐秘联系

There is a deeply ingrained hierarchy in how people evaluate artificial intelligence. Fundamental scientific research sits at the apex, software coding is celebrated as a productivity engine, commercial marketing is tolerated as secondary, and entertainment—particularly short videos—is frequently dismissed as a scandalous waste of precious computing power.

By Chen Bai

China’s booming AI-generated micro-drama industry has recently found itself at the center of a fierce public storm, accused of burning through domestic computing resources at an alarming rate.

The controversy stems from a disclosed data set. According to figures released by Chinese All Digital Publishing Group (Chinese All), a prominent digital content provider in China, the distribution of domestic AI token consumption breaks down as follows: micro-drama and video generation account for 55%, e-commerce and marketing take up 24%, while software development languishes at just 15%. Immediately, a provocative narrative swept across Chinese social media: more than half of China’s artificial intelligence computing capacity is being squandered on serialized trashy melodramas featuring overbearing CEOs, revenge plots, and miraculous reincarnations.

Public indignation flared. Critics lamented that scarce, high-end computing power meant for cutting-edge technology and academic breakthroughs was being hijacked to produce digital junk food. Some commentators even elevated the debate into a geopolitical framework, darkly implying that while the United States uses AI to supercharge industrial productivity, China is using it to manufacture spiritual escapism.

It is a narrative dripping with theatrical tension, tailor-made to feed tech anxieties.

Yet I would like to pose a question that may not easily win popular resonance: Leaving aside the statistical flaws of the data itself, suppose that half of AI computing power is indeed going into micro-dramas—why, exactly, is that such a profound problem?

Must a technology only be deemed valuable if it is confined to a laboratory, an academic journal, or a heavy manufacturing plant? Is helping exhausted ordinary workers find a fleeting moment of relaxation and joy during their leisure hours truly an unworthy consumer of computing power?

At its core, this ongoing debate reflects a deeply rooted conceptual bias that has accumulated over decades.

01

First, it must be clarified that the assertion that “over half of China’s computing power is directed toward micro-dramas” is itself an amplified misunderstanding.

The figure originated from an investor communication document published by Chinese All.

The original statement noted that micro-dramas and video generation accounted for 55% of token consumption, while software development accounted for 15%. However, token consumption volume does not equate one-to-one with physical compute utilization, nor can software development comprehensively represent all scientific research and innovation activities. Furthermore, the report lacked detailed statistical breakdowns and sample descriptions, leaving outsiders unable to independently verify the claims.

A more profound question remains: Even if we accept this description of compute allocation, what valid conclusions can actually be drawn?

Today, many observers instinctively divide AI applications into rigid tiers. Applied to fundamental science, it belongs to a high-value category. Used for code generation, it is a labor-efficiency tool. Deployed in advertising, it feels trivial. But when harnessed for micro-dramas and entertainment, it is treated as a reckless dissipation of technical resources.

Underpinning this evaluative framework is a remarkably antiquated industrial mindset: production trumps consumption, scientific research outranks commerce, and serious topics supersede entertainment.

Micro-dramas certainly suffer from crude production and homogenization, but these are issues of artistic quality and market sorting. We cannot dismiss the legitimacy of the entire entertainment industry receiving technological support simply because some micro-dramas fall short.

In fact, if AI-generated micro-dramas have genuinely generated massive computing demands, it proves precisely the opposite: China’s AI commercialization demand has achieved considerable scale. It signifies that creative professionals are willing to pay for generative video, content producers are willing to acquire AI tools, and various platforms are willing to deploy the necessary technical infrastructure.

These tangible economic activities embody the crucial process of technology transitioning from the ivory tower into broad industrial application.

02

Throughout the history of scientific and technological evolution, the entertainment sector has never been excluded from the engine of innovation. Many technologies that ultimately rewrote the course of history derived their initial developmental momentum precisely from the demands of gaming, film, and consumer markets.

The most paradigmatic example is the GPU, the fundamental infrastructure upon which today’s AI era runs. Nvidia‘s rise is intimately intertwined with the explosive growth of the video game industry. To make virtual characters in games come alive with fluid, realistic visuals, chip designers continuously expanded the parallel processing capabilities of graphics processors.

Who could have foreseen that technological advancements initially accumulated to enhance a gamer’s virtual experience would later form the foundational bedrock of the artificial intelligence revolution?

Had critics argued decades ago that video games were mere time-wasting pastimes and that chipmakers should not invest vast capital into optimizing gaming graphics, the developmental trajectory of today’s AI industry would be unrecognizable.

The cinematic technology sector tells a similar story. From computer-graphics imagery and digital modeling to motion capture and 3D rendering, the filmmaking industry has long served as a testing ground for cutting-edge algorithms. The evolutionary path of Pixar is particularly instructive. George Lucas established a division dedicated to computer film technology research as early as 1979, laying the technological foundation for what Pixar would eventually become.

To this day, digital entertainment studios continue pushing the boundaries of machine learning, ray tracing, geometric algorithms, and physics engines. A deeply symbiotic ecosystem of technological progress has long existed among gaming, film, and artificial intelligence.

Of course, this does not imply that every single current AI micro-drama will yield a major technological breakthrough. Yet the trajectory of innovation is notoriously tortuous. It is rarely possible to accurately predict in the nascent stages of a technology exactly where it will trigger cascading effects down the road.

AI micro-dramas can be viewed precisely as a valuable testing ground for video-generation algorithms. As tens of thousands of creators utilize models to generate content, encounter roadblocks, and spawn new demands, model developers receive invaluable feedback straight from the operational front lines. This cyclical feedback loop between technology and commerce drives application-layer innovation forward.

And application innovation happens to be a prominent comparative advantage of China’s AI industry. China boasts a massive internet user base, a relatively complete digital content industry chain, and an exceptionally rich, diversified array of commercial application scenarios. These endowments facilitate the rapid landing of new technologies, continuously driving down deployment costs and optimizing user experiences through fierce market competition.

Observers frequently analyze how China’s AI sector is playing catch-up in foundational models, core algorithms, and chip manufacturing, yet they routinely overlook another form of competitiveness: the capacity to rapidly transform cutting-edge technology into commercial products and achieve large-scale deployment in real-world environments.

03

I once stumbled across a video series on a Chinese platform called “Mao Ziqi.” The videos feature two orange-and-white tabby cats living in the countryside, going about their days with human-like mannerisms—cooking, farming, and enjoying a leisurely pastoral life.

The videos are, naturally, entirely generated via AI modeling, and the storylines are hardly sophisticated. Yet whenever I watch those cats earnestly going about their busy routines, a profound sense of comfort washes over me.

These clips do not teach me new knowledge, nor do they boost my work efficiency, let alone push forward the boundaries of human knowledge. Yet they genuinely bring me joy. And truthfully, that is quite enough.

For many people, I imagine this is the fundamental value of entertainment. Daytime labor, grueling commutes, and the endless friction of work and family life already consume an immense share of human emotional bandwidth. Not everyone wants to immerse themselves in deep literature, parse frontier physics, or ponder the ultimate destiny of human civilization after a grueling day on the clock.

Often enough, people simply want to watch warm, adorable imagery, listen to a simple narrative, and snatch a fleeting moment of peace amid exhaustion. What possible grounds are there to condemn such a desire?

For quite some time, our discussions surrounding the entertainment industry have been tinged with a sort of implicit moral smugness. Video games were once maligned as time sinks; web novels dismissed as lacking artistic merit; short videos condemned as fragmented information junk; and now AI micro-dramas are indicted as the chief culprits of compute waste.

Admittedly, entertainment products can induce addiction, algorithmic feeds can devour human attention span, and low-quality content can crowd out exceptional works. These genuine issues warrant deep exploration.

However, we must not use these concerns to negate the inherent value of entertainment itself, nor should we demote the ordinary person’s demand for leisure into a low-grade technological application.

Particularly in the age of artificial intelligence, technological progress ought to lead humans toward richer, more colorful lives.

Naturally, we hope artificial intelligence will conquer cutting-edge scientific hurdles, accelerating drug discovery, energy reform, and productivity leaps. But we must equally permit it to be deployed to help individuals create family archives, record moments of daily life, express personal thoughts, or simply make an exhausted office worker crack a smile upon scrolling past a video of an endearing cat late at night.

As for whether compute allocation is appropriate, evaluation is certainly warranted. But that evaluation should be grounded in market efficiency, energy consumption levels, ecological costs, intellectual property protection, and technological yields—rather than a simplistic moral judgment over whether the content is sufficiently solemn.

If there is idle capacity, vicious price wars, or excessive cash burning, we should criticize the specific commercial logic and resource allocation decisions. For technological advancement, trial, error, and iteration are inherent, necessary phases; no one can accurately predict in advance which corner the next breakthrough will sprout from.

In other words, what should be the ultimate service object of a technology tasked with transforming humanity? In my view, the answer should not be restricted to grand industrial objectives.

Technological progress certainly requires grand ambitions, but ultimately, it must return to the routines of everyday life. Academic breakthroughs, economic development, and productivity gains are meaningful precisely because they point toward a singular, overarching care: making human existence broader, more fulfilling, and happier.

We cannot simultaneously call for AI to benefit human life while looking down upon people’s genuine daily desires.

So, the next time someone laments that China’s AI computing power is being co-opted to produce micro-dramas, I would pose a counter-question: Why shouldn’t a short video that neither taxes the mind nor fails to bring laughter deserve to consume a bit of computing power?

Leave a Reply

Your email address will not be published. Required fields are marked *