Generative media · China

Kling AI (Kuaishou)

A video model born inside a listed short-video giant, with free distribution and revenue disclosed on earnings calls, is the most auditable media case on the list; retention and margin are still in the dark.

Verdict: Close to a real reading

Updated: 2026-10-06

Profile

What it does: Text-to-video and image-to-video model and product (Kling 1.0 to 3.0). Subscriptions for consumers, API for businesses. CEO: "revenue primarily comes from API call fees from B-end enterprise clients and subscription fees from C-end paying members" [2]
Founded: Parent company Kuaishou Technology (HKEX: 1024), China's second-largest short-video platform. Kling launched in June 2024 (secondary, to be verified) [4]
Founders: Founding team not disclosed; Kuaishou CEO Cheng Yixiao presents Kling himself on earnings calls [3]
Pricing model: API call fees on the B side, subscriptions on the C side; the company says "average monthly payment amounts" are rising [2]. Revenue is tied directly to credit consumption.
Funding: June 2026: spin-out round of roughly $2.8B at an $18B valuation; Alibaba, Tencent, Baidu (citing WSJ, secondary data) [4]
Hong Kong IPO targeted for 2027 (secondary data) [4]

Growth timeline

2025-03
ARR around $100M (earnings-call comparison, secondary report)
"$100 million in the same period last year"
[2]
2025-12
Monthly revenue above $20M, ARR $240M (secondary report)
[1]
2026-01
ARR above $300M; 2026 revenue expected to "more than double"
"Kling's ARR already topped USD 300 million by January"
[3]
2026-03
Q1 2026: quarterly revenue above 650M yuan, +300% year on year; ARR around $500M (secondary report)
[2]
2026-06
Spin-out: $18B valuation, 2027 IPO target (secondary report)
"People are paying for the product now"
[4]

Growth loop

1Creator generates video (credits)
2Circulation on Kuaishou and social
3New creators and subscribers
4Subscription and API revenue
5New model release funded by capex
Leak: capex far above revenue; the parent subsidises

The loop runs on output that advertises itself; the red arrow shows that the model-training bill is paid by the parent's treasury, not by revenue.

The drawing shows the shape of the loop; each node is explained in the list below. The red arrow marks value leaking out of the loop.

Mechanism

Acquisition channels: Three channels. First, the parent's short-video platform: 771.7M monthly active users, a ready creator base, channel cost close to zero [2] (opinion). Second, the global web product and API; orchestration layers such as Higgsfield packaging Kling as a model is a B2B channel (opinion). Third, the model release cadence: every major version creates a news cycle [3].
Loop: A creator generates a video with Kling, the video circulates on Kuaishou and social, a new creator tries it, subscription and API revenue comes in, capex improves the model, better output creates new circulation. The viral-output pattern; output quality, not a marketing budget, drives growth (opinion).
Activation: No public activation disclosure. The first moment of value is the first generated video (opinion).
Retention evidence: Only a qualitative statement: "Both B-end enterprise clients and C-end paying members have maintained favorable retention trends" [2]. No figure. No retention evidence.
Role of pricing: API call fees on the B side, subscriptions on the C side; the company says average monthly payment amounts are rising [2]. Revenue is tied directly to credit consumption; model upgrades can change the price of a credit (opinion).
Distribution moves: Advertising, short drama and gaming use cases; expansion into the US, Europe and Japan [4]. According to the Q4 2025 call as reported, more than 60M global users, more than 600M generated videos, more than 30,000 enterprise customers and developers [1] (secondary data).
Team: Headcount and revenue per employee not disclosed. The spin-out will make it a separate company [4]. 2026 capex of roughly 26B yuan, focused on Kling training, inference and data centres [1] (secondary data).

Real reading

The five questions of the Growth Claim Check, applied to this company's public claims.

01
What was measured, and by whom?
unclear
ARR and quarterly revenue come from a listed company's earnings call; whether Kling is reported as a separate segment, and the ARR definition, are unverified.
02
Perception or outcome?
yes
A revenue figure, not a usage claim.
03
Is there a stability or quality metric?
no
Retention is only described as "favorable"; no margin data.
04
Is there a counterfactual or cohort?
no
No control group or cohort disclosure.
05
Nominal or real?
unclear
ARR is derived as twelve times monthly revenue; model release spikes and credit consumption can inflate it when annualised.

Verdict

Close to a real reading

The only media case on the list whose figures rest on management statements in a listed company's quarterly disclosure; the figures come from secondary relays of the earnings call, and the primary IR source was not opened. Without retention and margin data it cannot be called profitable growth; capex is not paid by revenue.

Test your own claim with the same five questions: Growth Claim Check →

Shadow side

Management says on its own call that the market is "immature and crowded" and that rival multimodal models create a monetisation risk [3].
2026 capex of roughly 26B yuan; even if Kling revenue doubles it stays far below the investment, and the parent subsidises it [1] (opinion).
US chip export controls are cited as a risk [4].
Two different revenue figures circulate for the same quarter in secondary reports (Q4 2025: 314M and 340M yuan); December 2025 ARR of $150M and $240M contradict each other [1] [3] [4]. An example of how unreliable secondary reporting is.

What it teaches a founder in Türkiye

This section is opinion; a reading, not a source.

01
(opinion) A product born inside a large distribution platform starts with channel cost near zero. The Turkish equivalent is launching a product from inside a company that already has a customer base.
02
(opinion) When the output is the ad, output quality and shareability drive growth, not the marketing budget.
03
(opinion) Annualised revenue is the slipperiest metric in a credit-consuming product. In your own product, track three months of realised revenue and paying members instead of ARR.
04
(opinion) The capex of a business that trains models is not paid by revenue. Model orchestration and model ownership are two different capital games; read this next to the Higgsfield case.
05
(opinion) A listed company's management statement is a calibration. Read the other cases' claims against this scale.

A question for you: remove this company's logo and drop the loop onto your own product. Which node stays empty?

Sources

Numbers are the references in the text. An 'opened' source was read by the researcher; 'secondary' ones are verified against their primary source before publication.

[1]
Kuaishou Accelerates AI Commercialization: Kling AI's ARR Exceeds $300 Million, 2026 Capex to Reach 26 Billion Yuan
BigGo Finance (earnings-call report) · 2026-03-26 · opened
“"strong confidence" that revenue would "more than double year-over-year"”
[2]
Kuaishou's Kling AI Revenue Surges Threefold in Q1, ARR Nears $500 Million
BigGo Finance (earnings-call report) · 2026-05-29 · opened
“revenue primarily comes from API call fees from B-end enterprise clients and subscription fees from C-end paying members”
[3]
Kuaishou Earnings Call: AI Bets Drive Next Phase
The Globe and Mail (press release report) · 2026 (Q4 2025 call) · opened
“Kling's ARR already topped USD 300 million by January and expects 2026 Kling revenue to more than double”
[4]
Kuaishou's Kling AI is spinning out at a $20 billion valuation
Startup Fortune (citing WSJ) · 2026-06-17 · opened
“People are paying for the product now”
[5]
Kuaishou investor relations results announcements (primary, could not be opened)
Kuaishou Technology · 2026-03 and 2026-05 · secondary
Note: Primary release; the evidence reviewer must open it before publication.

Open questions

Is Kling revenue a separate line in Kuaishou's official release; is ARR twelve times monthly revenue or contract-based?
Was Q4 2025 Kling revenue 314M or 340M yuan; was December 2025 ARR $150M or $240M?
Kling launch date (June 2024?) and 2024 revenue.
User and paying-member counts, B versus C revenue split, retention figure.