4,000 yuan, 96 episodes, 600 million views.
On November 13, Bai Feikong, a 40-year-old from Chengdu, topped Weibo’s trending list. He used to be a 3D home-design artist. He quit his job in March and, working alone over a month and a half, completed all 96 episodes of the AI short drama “Reborn as a Mantis: Awakening Infinite Devouring and Evolution.” It cost just over 4,000 yuan, all spent on memberships and credits for AI platforms. No team, no paid promotion, no actors.
Let’s break it down: 41.7 yuan per episode.
The time, too: 96 episodes in 45 days—that’s more than two a day. He kept regular office hours: starting at 9:30, taking a lunch break, and finishing at 6. Eight hours a day.
A traditional short drama costs hundreds of thousands to over a million yuan. Using 500,000 yuan as the benchmark, that’s 125 times the cost. The big-ticket expenses—actors, locations, post-production—were wiped out by the tools.
The numerator looks good. 600 million views divided by 4,000 yuan means every yuan bought 150,000 views.
But the number that really matters is the revenue share. He said the series brought him an extra 60,000 yuan. 600 million views for 60,000 yuan works out to one yuan per 10,000 views.
Work backward from that rate, and a 4,000-yuan production cost takes 40 million views just to break even.
Hongguo’s popularity and ratings are measured differently: 60.42 million popularity points, a rating of 8.9, 4,934 ratings, and 791 reviews. Views, popularity, and revenue share are three different metrics. Mix them up, and you’ll mistake 600 million views for 600 million yuan.
Now look at the denominator. DataEye’s figures: in the first half of 2026, more than 220,000 new AI dramas were added on Douyin, but only 1,055 surpassed 100 million views—a hit rate below 0.5%. If 50 million views is the break-even point, just 1.3% make the cut, while more than 90% struggle to attract viewers.
This news story is about one of those 1,055 out of 220,000.
Bai Feikong’s first drama had 71 episodes and 890,000 views—it was among the other 98.7%. His second drama did more than 600 times better, thanks to his decade of 3D experience, visual taste, and knack for choosing scripts—not AI itself. Douyin vice president Li Liang has publicly warned people not to treat an extremely rare success as a get-rich-quick blueprint.
Let’s go one step further upstream. 220,000 dramas times 4,000 yuan comes to around 880 million yuan. That money doesn’t go into creators’ pockets; it goes to platform credits and computing power. Some users say they burn through thousands of yuan on a single episode, which would push that figure even higher.
In this business, the ones collecting the money are the tool and computing-power sellers, not the people making the dramas.
There’s another factor people overlook. Those 220,000 dramas are all competing for the same slice of the same people’s free time. The total amount of time doesn’t change. When supply grows by two orders of magnitude, the money each view can earn only goes down. This isn’t speculation—it’s division by the size of the revenue pool.
Crypto has the same pattern.
Launching a coin used to mean writing a contract and paying deployment fees. Now it takes a few clicks, and new coins pour out every day. AI has brought the cost of making an episode down to 41 yuan, and there are 220,000 dramas a year. Production costs have fallen by two orders of magnitude, but demand-side attention hasn’t changed at all. The odds of success haven’t been raised by skill; they’ve been pushed down by the proportion of supply.
That’s why I don’t focus on individual stories like “who made how much from this.” I watch just two numbers: the daily increase in new supply, and the share that survives. If those two numbers don’t change, no matter how many success stories there are, they’re still just the numerator.
Now, $SOL . Today it’s $80.46, down 2.65% over 24 hours, with $269 million in trading volume and an intraday low of $77.92. On a day when BTC broke below $70,000, it didn’t crash along with it. That suggests selling pressure isn’t on it for now.
But the supply story doesn’t give it a valuation, either.
I look at these two things separately: one is token supply, the other is the narrative. Mix them together, and you’ll get the math wrong.
If the denominator doesn’t change, the numerator is worth nothing.
The barrier to entry in this industry has never been the tools. It’s that 0.5%.
#BlackAndWhiteAccounting
On November 13, Bai Feikong, a 40-year-old from Chengdu, topped Weibo’s trending list. He used to be a 3D home-design artist. He quit his job in March and, working alone over a month and a half, completed all 96 episodes of the AI short drama “Reborn as a Mantis: Awakening Infinite Devouring and Evolution.” It cost just over 4,000 yuan, all spent on memberships and credits for AI platforms. No team, no paid promotion, no actors.
Let’s break it down: 41.7 yuan per episode.
The time, too: 96 episodes in 45 days—that’s more than two a day. He kept regular office hours: starting at 9:30, taking a lunch break, and finishing at 6. Eight hours a day.
A traditional short drama costs hundreds of thousands to over a million yuan. Using 500,000 yuan as the benchmark, that’s 125 times the cost. The big-ticket expenses—actors, locations, post-production—were wiped out by the tools.
The numerator looks good. 600 million views divided by 4,000 yuan means every yuan bought 150,000 views.
But the number that really matters is the revenue share. He said the series brought him an extra 60,000 yuan. 600 million views for 60,000 yuan works out to one yuan per 10,000 views.
Work backward from that rate, and a 4,000-yuan production cost takes 40 million views just to break even.
Hongguo’s popularity and ratings are measured differently: 60.42 million popularity points, a rating of 8.9, 4,934 ratings, and 791 reviews. Views, popularity, and revenue share are three different metrics. Mix them up, and you’ll mistake 600 million views for 600 million yuan.
Now look at the denominator. DataEye’s figures: in the first half of 2026, more than 220,000 new AI dramas were added on Douyin, but only 1,055 surpassed 100 million views—a hit rate below 0.5%. If 50 million views is the break-even point, just 1.3% make the cut, while more than 90% struggle to attract viewers.
This news story is about one of those 1,055 out of 220,000.
Bai Feikong’s first drama had 71 episodes and 890,000 views—it was among the other 98.7%. His second drama did more than 600 times better, thanks to his decade of 3D experience, visual taste, and knack for choosing scripts—not AI itself. Douyin vice president Li Liang has publicly warned people not to treat an extremely rare success as a get-rich-quick blueprint.
Let’s go one step further upstream. 220,000 dramas times 4,000 yuan comes to around 880 million yuan. That money doesn’t go into creators’ pockets; it goes to platform credits and computing power. Some users say they burn through thousands of yuan on a single episode, which would push that figure even higher.
In this business, the ones collecting the money are the tool and computing-power sellers, not the people making the dramas.
There’s another factor people overlook. Those 220,000 dramas are all competing for the same slice of the same people’s free time. The total amount of time doesn’t change. When supply grows by two orders of magnitude, the money each view can earn only goes down. This isn’t speculation—it’s division by the size of the revenue pool.
Crypto has the same pattern.
Launching a coin used to mean writing a contract and paying deployment fees. Now it takes a few clicks, and new coins pour out every day. AI has brought the cost of making an episode down to 41 yuan, and there are 220,000 dramas a year. Production costs have fallen by two orders of magnitude, but demand-side attention hasn’t changed at all. The odds of success haven’t been raised by skill; they’ve been pushed down by the proportion of supply.
That’s why I don’t focus on individual stories like “who made how much from this.” I watch just two numbers: the daily increase in new supply, and the share that survives. If those two numbers don’t change, no matter how many success stories there are, they’re still just the numerator.
Now, $SOL . Today it’s $80.46, down 2.65% over 24 hours, with $269 million in trading volume and an intraday low of $77.92. On a day when BTC broke below $70,000, it didn’t crash along with it. That suggests selling pressure isn’t on it for now.
But the supply story doesn’t give it a valuation, either.
I look at these two things separately: one is token supply, the other is the narrative. Mix them together, and you’ll get the math wrong.
If the denominator doesn’t change, the numerator is worth nothing.
The barrier to entry in this industry has never been the tools. It’s that 0.5%.
#BlackAndWhiteAccounting