On the Shanghai metro at 8:40 in the morning, the woman across from me holds her phone straight up, not sideways. She is not scrolling. She is watching a woman in Ming dynasty dress get thrown out of her own wedding, in two minute bursts, tapping to unlock the next episode every time the screen fades to black. Nobody around her looks up. This is 短剧, duanju, the vertical micro drama format that has quietly become one of the most watched forms of entertainment in China, and increasingly, one that production houses are willing to spend real money on.
The numbers say how fast this has moved. By Sina Finance’s account from January 2026, the format is racing toward tens of billions of RMB in scale this year, even as the industry itself is, in the article’s own words, “aging” at speed: the wild, anyone-can-shoot-one phase of two years ago has given way to professional managers running standardized production lines. An ordinary series now costs 400,000 to 700,000 RMB to make. Trade coverage counted more than 1,600 productions in 2025 budgeted above 1 million RMB, with the top “S-tier” series running 1.5 to 3 million. Crews that were 10 to 20 people two years ago now run 40 to 60. It is cheap compared to a real television series, and it reaches an audience that mostly stopped watching television altogether.
Here is what the headline growth number does not tell you. A Qianzhan Industry Research Institute breakdown of where a typical micro drama budget goes puts platform distribution and service fees at over 80 percent of total spend, production at under 8, cast at 5. The content is not the expensive part anymore. Buying the traffic that puts it in front of an audience is. That fact should reset how any brand thinks about a short drama line item. It is not a production budget with a small media add on, the way a branded video usually works. It is a media buying operation that happens to need a script.
I have watched this shape before. When KOL marketing arrived in China around 2015, brands treated it like a cheaper magazine ad: pay a face, get a post, done. It took years to understand that the real cost sat in the paid amplification behind the post, not the post itself. Livestream commerce went through the same correction later. Short dramas are on the same curve, compressed into months. Production has professionalized fast. Placement pricing has not, and I want to say that plainly: I looked for a published figure for what it costs to place a brand inside a hit series, and found none. What exists is case evidence, not pricing. A series called 太奶4, past 2 billion views, carried placements from the snack brand 良品铺子, Oreo, the beauty brand Chando, and Guerlain, built as episode easter eggs and brand livestreams, not a paid slot with a posted rate. Trade press says some of that budget is shifting out of long form drama sponsorship. Nobody is publishing what it costs to buy in.
Then there is the AI question, and here I want to be careful, because the hype around AI generated short dramas is real and the substance underneath it is uneven. What I can confirm: a regular AI produced series, roughly sixty two minute episodes, can be made for 120,000 to 300,000 RMB, at least half the cost of an equivalent real actor shoot. For visual effects heavy shots specifically, the gap is far wider. Traditional effects work runs around 3,000 RMB per second of finished footage, and an AI generated equivalent can come in near 3 RMB, according to a March 2026 cost breakdown. China’s first paid AI generated series, 兴安岭诡事, built 90 percent of its scenes with a five person team over three months, and still pulled more than 4 million views in nine days. That is the collapse in production cost. It is not, yet, a collapse in quality on anything that has to carry a premium brand’s visual codes. I would not put a maison’s silk or a hard case’s hardware in front of an AI camera today. The uncanny valley is still a real place, and luxury clients notice texture before plot.
So who should actually pay attention. If your brand tells its story through how something looks worn, held, unwrapped, a short drama is a genuinely new stage: fashion, fine jewelry, beauty, accessories, the categories where a close up on a bag clasp or a lipstick cap does real work. If your brand sells to a procurement department, or explains itself best in a spec sheet, this format has almost nothing to offer you. Duanju audiences want swept away, wronged, revenge, redemption. They do not want a webinar.
The window is open because the rules are not fixed yet. Placement pricing has no floor and no ceiling anyone will admit to in writing. Platforms are still experimenting with revenue splits, iQiyi paying up to 130 percent of guaranteed revenue on exclusive premieres, Youku giving non exclusive dramas the full ad split. That instability makes an early conversation cheaper than it will be once agencies start publishing rate cards the way KOL platforms eventually did.
A jewelry client of ours in Shanghai tested this last year at a far smaller scale: not a co produced series, but a single scene placement negotiated directly with a mid size production house, a bracelet worn by a supporting character across three episodes of a modest budget romance. The series never became a breakout hit. It still pulled several hundred thousand views, at a cost well under one day of livestream KOL fees. Nobody mistook it for a campaign. Everybody involved learned how the negotiation actually works, and that alone was worth the spend.
The mechanism is simpler than the industry’s jargon makes it sound. There are three doors in. The cheapest is paid placement inside a series already in production, your product worn or used for an agreed number of seconds, negotiated directly with the production house or through an MCN that represents several of them. The middle door is co production, your brand funding part of a series for a storyline built around your product category, closer to what 太奶4’s beauty partners did. The most expensive, most controlled door is a fully custom branded series: your story, your casting approval, your platform buy.
Budget realistically for the placement route. Production side costs alone will echo the 400,000 to 700,000 RMB range if you are helping fund the series, plus a separate and usually larger sum for the traffic buy that actually gets the episode seen, the same 80 plus percent of spend the Qianzhan data points to. That traffic buy is not optional. A beautifully placed product nobody watches is worse than no placement at all. As for AI production, honestly, not yet, not for anything that has to look premium. It fits the fast, cheap testing tier, a branded vertical piece to gauge whether your category even works in this format, before you commit real money to a live action shoot. Treat the AI version as a sketch, not the finished painting.
This is a negotiation problem before it is a creative one, which is the gap our video production team in China spends its time closing: sourcing the right production house, negotiating placement terms nobody publishes, briefing a script team so your product belongs in the story rather than bolted onto it. The casting pool overlaps with the world we already work in through our influencer and KOL network, since duanju’s breakout actors often carry a fanbase straight into the placement.
Some buyers I talk to think short drama placement will get regulated the way livestream commerce eventually was: disclosure rules, verified pricing, a paper trail. Others think the format is too fragmented for any regulator to get a grip on it the same way. I lean toward regulation catching up within two years. But I would rather be wrong out loud than right in silence. Where do you land?