A video cost $443 to make. Three days later, it had made $500,000.
But in that same Chinese micro-drama market, 9 out of 10 new titles don't even break even.
Here's what everyone assumes
"AI content is just slop." That phrase went so mainstream that Merriam-Webster and the American Dialect Society both named "slop" their 2025 Word of the Year. It means AI-generated content that's mass-produced, low-effort, and hollow.
Justine Moore, the a16z partner who tracks AI and consumer trends, thinks that framing misses the point entirely. Here's what she said in a recent interview about micro-dramas.
"Consumers care less about how something was made."
— Justine Moore, a16z
Low-quality content existed long before AI did. The real variable was never the tool — it's who made it, why, and how much effort they put in. And the numbers backing that up are honestly staggering.
But the numbers say the opposite
Start with scale. China's micro-drama audience is 662 million people, with market revenue past 50 billion yuan (roughly $7.23B) — already surpassing the country's entire domestic box office. By late 2024, daily time spent in micro-drama apps was matching time spent in messaging apps.
The demographics are the surprising part. Despite the "this is just Gen Z content" assumption, users 50 and older make up nearly 30% of the audience, with ages 40-49 at 19.7%. This is nowhere near a youth-only format.
Korean outlet Ohmynews reported field-level numbers that make this even more concrete. After AI adoption, production timelines dropped from 15-30 days down to under 5 days, and costs fell to roughly 1/5 of what they used to be. Some productions ran costs as low as $30/minute; the most extreme case cost about $500 total and took 48 hours. The result: over 470 AI micro-dramas were released per day this January, and AI-made titles now make up 38% of the top-100 chart, up from just 7% a year earlier.
Overseas markets exploded too. From January to August 2025, overseas micro-drama revenue hit $1.525 billion — up 195% year-over-year — and ReelShort alone pulled in $400 million overseas in 2024. Even Character.ai, the character-chatbot company, has jumped in with three in-house-produced series inside its app — leaning on 20 million MAU and 75 minutes of average daily engagement, numbers that beat most streaming benchmarks.
But 90% of it is failing
The other side of the same market
All that volume created an oversupply crisis in China's domestic market. Measured by wholesale revenue, the loss rate hit 90%. Earnings per 10,000 views collapsed from 30-100 yuan down to 5-10 yuan, and more than 80% of small and mid-sized production teams stopped updating. And yet, in just the first four months of 2026, over 8,000 new comic-drama production companies launched — a 300% year-over-year increase.
That's a signal the volume game is already over. As production costs approach zero, anyone can enter — and once anyone can enter, without real differentiation, you just disappear into the pile of cheap content. Chinese apps' share of global top-30 download rankings has already dropped from 95% to 67%, as players from the US, Korea, and elsewhere carve out different approaches.
| Playing on volume | Playing on differentiation | |
|---|---|---|
| Strategy | Same clichés, faster and cheaper | Stories and characters that can't be copied |
| China domestic reality | 90% loss rate | Breakout hits like Zombie Cleaner still exist |
| Durability | No barrier to entry — saturates fast | Characters and worlds are hard to replicate |
So what actually decides who wins?
Justine Moore's real point isn't about which AI model you use. It's about what workflow and what story creators and studios assemble those models into. Two of a16z's recent investments aim directly at that layer.
Glif is a "creative super agent" that bundles multiple AI models into one tool. Instead of hopping between apps, creators can direct an entire pipeline — script, storyboard, video — from one place. Its "Skills" feature packages proven templates for short-form video and influencer content. Co-founder Fabian Stelzer is himself an early AI filmmaker who made the series "SALT."
Phota Labs solves a narrower but very real problem: characters looking subtly different from scene to scene in AI video — a dealbreaker for anything serialized like a micro-drama. Phota Labs offers an API that generates keyframes with a locked identity, so the same face holds up across episodes.
Both companies are really betting on the same thing: the workflow layer that makes AI output consistent and intentional, not just "generated." That's also the exact line separating AI slop from an actual AI micro-drama.
Could I actually try this myself?
The technical barrier is basically gone. The real gate is somewhere else.
- Start with the premise, not the model
Before picking a video tool, nail down a strong, one-line story premise and characters. Assume you can't win on volume in this market. - Use a workflow tool to unify the pipeline
Instead of juggling separate apps for script, storyboard, and video, a tool like Glif makes it easier to keep tone consistent across episodes. - Solve character consistency early
If you're planning a series, evaluate identity-locking tools like Phota Labs up front. Viewers drop off fast if the face keeps drifting episode to episode. - Understand the monetization mechanics
Micro-dramas monetize like games — in-app purchases to unlock episodes, not ads. Payout structures vary by app (ReelShort, Character.ai, etc.), so check before you distribute. - Skip the oversaturated markets
Don't launch the same clichés into an already-saturated market like China's domestic scene. Start with a differentiated story in a less-crowded market or niche genre instead.
Go deeper
AI Micro Dramas, Generative Media, and the Future of Creativity The original a16z podcast episode featuring Justine Moore a16z.com
Investing in Glif a16z's official announcement on why it backed this "creative super agent" a16z.com
About Phota Labs The identity-preserving visual generation API built to solve character consistency photalabs.com
Character.ai Enters The Microdrama Race Why the chatbot giant is producing its own in-house micro-drama series forbes.com
Chinese micro-dramas gain global momentum as industry booms 662 million users, 30% aged 50+ — the scale and demographics of China's micro-drama market en.people.cn
South Korean reporting on AI-made short dramas On-the-ground coverage of real production costs and timelines ohmynews.com
