Business judgment on AI products
There is public usage data. The useful view is what it does, its scale, and whether the change lasts.
01
Start inside the user's day · Public facts + observable behavior · 2026-08-28
可灵 replaces the most expensive live-action shots in ads and short dramas with generated footage. Monthly actives: 8.1 million.
Public materials do not yet show how users complete this job today or what they replace.
The product targets friction in this job, but public user evidence does not yet show the cost, frequency, or consequence of leaving it unsolved.
A public record shows 8.11M mau and -1.69% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
① Whether the Hong Kong IPO launches within the expected window — filing within 12 months is the key validation point of the carve-out narrative; ② Whether the full year delivers against the 3B RMB target — Q1's 650M implies accelerating quarters ahead; ③ Whether auditable public delivery cases in p…
Worth dissecting. A public record shows 8.11M mau and -1.69% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
to monetize a creation product, first figure out which cost segment you save for whom. Kling doesn't save ordinary people editing time; it saves the film/ad/short-drama industry the cost of live shooting — those budgets are large, willingness to pay is high, and cases are auditable. When building an AI tool, target "the production segment with the largest budget," not "the scenario with the most users." B-end API by usage plus C-end subscriptions is standard; the reference-worthy piece is how it penetrates professional scenarios — by using public film cases (Peaceful Year, House of David) as sales assets. For AI tools selling to businesses, the case library is part of the product.
Two engines: enterprise API revenue plus consumer paid-membership subscriptions. Kuaishou CEO Cheng Yixiao said on the earnings call that both engines drive growth and that both enterprise clients and paid members show healthy retention. Pr… ① Whether the Hong Kong IPO launches within the expected window — filing within 12 months is the key validation point of the carve-out narrative; ② Whether the full year delivers against the 3B RMB target — Q1's 650M implies accelerating quarters ahead; ③ Whether auditable public delivery cases in p…
可灵 replaces the most expensive live-action shots in ads and short dramas with generated footage. Monthly actives: 8.1 million.
A public record shows 8.11M mau and -1.69% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
An English validation note will follow from the public evidence.
The product claims to help users complete: “可灵 replaces the most expensive live-action shots in ads and short dramas with generated footage.”. User evidence has not yet verified pain intensity or the cost of doing without it.
The assessment is recorded; an English explanation is pending.
The assessment is recorded; an English explanation is pending.
The assessment is recorded; an English explanation is pending.
02
Market comparison
The Chinese–English market comparison is not complete yet. A conclusion follows only after its coverage and verifiable evidence are recorded.
03
The call and next move come first; the full read retains the evidence and counterevidence.
Kuaishou's AI video generation platform, positioned as "a new generation AI creative productivity platform" — tool, community, and professional verticals as a trinity. It is the fastest-commercializing Chinese video model: Q1 2026 revenue above 650M RMB, ARR approaching $500M, and the largest single financing round of any video model globally.
Kuaishou. Kling AI was incubated by Kuaishou and in June 2026 raised close to $3B (led by CPE Source Peak and Tencent) at a valuation around $18B — a global record for video-model financing. Kuaishou holds 68.33% through Lucky Labs and other entities, keeps consolidating it financially, and expects to start a Hong Kong IPO within 12 months. The team is Kuaishou's AI division; headcount is not disclosed.
Read: Kling is the most complete specimen of the standard path "incubated inside a giant → independent financing → carve-out IPO." Kuaishou's compute, data (short-video ecosystem), and money all fed it; the independent round answers "how to prove independent value," not "how to survive."
Making an ad spot, a VFX sequence, or an episode of a short drama used to mean a shooting crew + studio + actors + post-production editing + a VFX house. A 30-second ad started at hundreds of thousands of yuan; an episode of short drama ran thousands to tens of thousands, on a weeks-long cycle.
Kling replaces the two most expensive segments of that chain — shooting and post-production: in the historical drama "Peaceful Year," 30–40% of the production pipeline was handled by Kling, with AI-generated shots costing under 1% of the traditional approach; in Hollywood's "House of David" season 2, Kling generated 400 shots at about one-third of a traditional studio's quote. For ordinary people the case is "Paper Phone" — two non-professional creators made nearly all of its visuals with Kling in three days, and it passed 100M plays.
Read: Kling does not replace the "editing software" tool; it replaces "live shooting" as a mode of production. The weight of that sentence is that live shooting is the root of the film industry's cost structure — replacing it means replacing the profit distribution of an entire supply chain.
Two engines: enterprise API revenue plus consumer paid-membership subscriptions. Kuaishou CEO Cheng Yixiao said on the earnings call that both engines drive growth and that both enterprise clients and paid members show healthy retention. Pricing is membership subscriptions plus usage-based API billing (stable public price points could not be verified).
| Dimension | Call |
|---|---|
| Founder-product fit | High: Kuaishou's short-video ecosystem is the ideal training-data source for a video model, and the team is Kuaishou's core AI division |
| Product insight | High: tool + community + professional verticals advancing on three lines, and it moved into film-industry work (Peaceful Year, House of David) early |
| Execution quality | High: 3.0 All-in-One full modality plus native 4K, first tier in text-to-video (alongside Seedance 2.0 and HappyHorse-1.0) |
| Timing | Strong: it caught both the video-model financing window and the short-drama industrialization boom (AI live-action short-drama supply grew dozens of times in H1 2026) |
Kling carries the strongest commercialization signal in this batch — but its magnitude gap with Jimeng exposes a structural fact.
The good side first: Q1 revenue of 650M RMB with +300% YoY, ARR quadrupling to $500M in a year, 100M global users, $3B raised at an $18B valuation — this is the only domestic video model whose revenue scale can be written into an earnings report. Its commercialization route is also clear: B-end API + C-end subscriptions, with professional verticals (film/ad/short-drama/game) generating paid demand far better than entertainment-at-scale.
The question mark is the scope: traffic board's app MAU is only 8.11M and falling 1.69% MoM, two orders of magnitude away from the official "100M global users." This is not fabrication — app-only scope and total scope are different things — but it shows: Kling's revenue center is B-end API and enterprise production, not C-end app subscriptions. A slight app-MAU decline is not fatal inside that structure, but reading only the traffic board number would badly understate its commercial size.
Read: Kling and Jimeng side by side is the most informative comparison in this batch — Jimeng uses the Douyin ecosystem to reach 87M app MAU; Kling uses professional scenarios to reach $500M ARR. One validates "distribution decides scale," the other validates "scenario decides revenue." For AI product builders these are two selectable growth models: To C via distribution, To B via scenario — don't try to win both ways.
① Whether the Hong Kong IPO launches within the expected window — filing within 12 months is the key validation point of the carve-out narrative ② Whether the full year delivers against the 3B RMB target — Q1's 650M implies accelerating quarters ahead ③ Whether auditable public delivery cases in professional verticals keep appearing (film, ad, game) — claims like "AI production at one-third the cost" need more third-party verification
Product logic: to monetize a creation product, first figure out which cost segment you save for whom. Kling doesn't save ordinary people editing time; it saves the film/ad/short-drama industry the cost of live shooting — those budgets are large, willingness to pay is high, and cases are auditable. When building an AI tool, target "the production segment with the largest budget," not "the scenario with the most users."
Pricing structure: B-end API by usage plus C-end subscriptions is standard; the reference-worthy piece is how it penetrates professional scenarios — by using public film cases (Peaceful Year, House of David) as sales assets. For AI tools selling to businesses, the case library is part of the product.
Worth watching. It is the most solidly commercialized item in this batch — revenue, financing, valuation, and IPO path are all real, with auditable cases in professional scenarios. The app MAU is small and slipping, which is not fatal inside a B-end revenue structure, but it says this is essentially an enterprise production tool, not a national creation app. For the reader: it is the best current specimen of "AI tool turning into a business" and worth long-term tracking.
05
Use these searches when the official site is missing or the current link is only a lead.