Kuaishou: After a Near-Death Plunge, Can AI Power a Revival? ---

$KUAISHOU-W(01024.HK) released its Q4 print after the Hong Kong close on Mar 25 Beijing time. The quarter slightly beat, with the key highlight being stable top-line growth coupled with internal efficiency gains lifting profit. However, within segments, Kling fell short of the Street’s early-year optimism, and rising industry competition since Feb likely contributed to the share-price pullback.

That said, management’s call suggested Kling’s performance was not bad and its outlook remains constructive. By contrast, legacy businesses face clear pressure from the policy environment, intensifying competition and the recognition of AI-related spend.

While we estimate Kling’s implied valuation is still not fully priced in, especially amid the current LLM-fueled rally, investors may stay cautious due to the drag from legacy. A further de-rating cannot be ruled out in the near term. Whether the stock can continue to re-rate likely hinges on Kling delivering more upside surprises.

Details below. Key points follow.

1. Can operating efficiency still improve this year? AI investment is being recognized in P&L, with equipment depreciation up 10% YoY and R&D up 34% YoY. But the company controlled basic opex, with total employee costs up 4% and marketing roughly flat YoY. As a result, core OPM reached 13.5%, up 350bps YoY and +40bps QoQ.

We expect AI spending to weigh more on margins as more AI revenue is recognized (management guided margin erosion this year). Whether the market accepts this margin dilution without cutting multiples depends on growth delivered. Specifically, can legacy break the natural slowdown and re-accelerate, or can Kling withstand competition and industry shifts to sustain strong sequential growth?

2. Kling missed near-term expectations, but guidance is upbeat: Kling v2.6 launched in Dec with improved motion control and synchronous audio-visual output, drawing solid user feedback. Monthly grossing hit a record US$20 mn, prompting Street Q4 revenue expectations to be raised from the RMB 300 mn guide to RMB 350–400 mn, with 2026 full-year estimates nearing RMB 2 bn, implying almost a doubling.

Actual Q4 revenue was RMB 340 mn, near the low end of guidance and Street. Companies typically preview with the sell side, and after Dec’s spike, iOS data showed a quick pullback in Jan; while Kling 3.0 rolled out in Feb, the viral surge of Seedance 2.0 also curbed Kling’s rebound in grossing. As such, institutions likely marked estimates down accordingly, weighing on the share price since Feb.

On the call, management disclosed Jan ARR above US$300 mn, +25% vs. Dec, indicating Kling’s grossing did not weaken as much as external trackers suggest. Dolphin Research believes third-party data only captures iOS app performance, while the platform is fundamentally productivity-oriented with a stronger B-side that skews to PC; B-side mix has been rising (40% in Q3). This is exactly what third-party data misses.

3. E-comm growth steady: Q4 e-comm GMV grew 13% YoY, slowing from Q3, with full-year GMV near RMB 1.6 tn. Under creator distribution, full-service and the 'super link' offering, take-rate improved by 8bps YoY (down QoQ due to seasonal commission rebates), driving e-comm revenue +22% YoY. This slightly beat top-tier sell-side expectations (+18% YoY).

4. Ads slightly beat; watch the 'traffic tax' impact: Q4 marketing ad revenue rose 15% YoY, a modest sequential recovery and a small beat. The market has been cautious on ads given macro consumption, competition and the 'traffic tax'. Whether the 'traffic tax' impact expands further into 2026 warrants attention on management commentary.

Growth benefited from last year’s low base, stable e-comm, and sustained strength in short dramas (incl. AI comics) and mini-games. It was also helped by the OneRec recommendation model, UAX fully automated buying (outer-loop ad consumption penetration at 80%, +10ppt QoQ), and incremental AIGC ad tools. These factors collectively supported ad momentum.

5. Live streaming under expected pressure: Q4 live-streaming payments fell 2% YoY, indicating softness. This reflects industry contraction and the platform’s own ecosystem governance, and we expect the pressure to intensify.

6. User ecosystem stable: With the platform now mature, user metrics are steady, with 740 mn MAU and 410 mn DAU, both up low single digits YoY. Stickiness is 55%, and user time spent is flat YoY.

7. Shareholder returns may rise: The company stepped up buybacks during the Q4 share-price correction. It repurchased HKD 3.2 bn for the year and declared HKD 3.0 bn in dividends, implying a 2.7% shareholder return. With a solid cash position and renewed share-price pressure, we think sizable repurchases may resume post-blackout; watch management’s color on the call.

8. Detailed financials at a glance

Dolphin Research View

Q4 was decent, especially at ~9x P/E, implying expectations have been reset and may even be conservative. That said, the market’s concerns are not unfounded: user growth is near stall speed, offering little 'extra' support for legacy businesses amid the policy backdrop ('traffic tax' and weak goods consumption) and competition. Put differently, Kuaishou’s growth optionality is largely in AI, including ads, AI comics for ToC, AI digital marketing and AIGC solutions for ToB, and the key growth engine Kling.

AI on the ad side mainly slows the deceleration, while Kling is a true incremental business that captures investor imagination. Since Kling’s grossing inflected in Q2 last year, investors quickly re-rated Kuaishou, with many applying a high-growth framework and ~35x P/S to Kling. This premium sets a high bar for consistency.

Kling’s lofty expectations also amplify the sensitivity of overall valuation to near-term run-rate swings. For example, ARR of US$150 mn vs. US$300 mn drives ~HKD 35 bn of value swing, or about 15% of current market cap. Actual long-term revenue optimism last year was even higher, implying greater volatility.

Street expectations for 2026 remain above management’s small-meeting guidance, and while management is habitually 'conservative', the gap is material and likely needs to be revised down. Two areas stand out for adjustment, in our view.

1) Legacy revenue growth is expected at just 5% for 2026, a clear slowdown vs. 11.7% this year. This reflects continued pressure on live-streaming tipping, and ads/e-comm being affected by the 'traffic tax' and competition. We think these headwinds are realistic.

2) Benefiting from v2.6 in Dec and v3.0 in Feb, management disclosed Kling’s Jan ARR above US$300 mn and guided at least a doubling in 2026 (RMB 2.1 bn+), above the market’s +85% YoY growth view. Street iOS grossing trackers miss the growing ToB PC-side contribution, which is harder to track but increasingly meaningful.

On profits, we make further adjustments for AI’s margin impact. We assume a 200bps hit to GPM from AI (capex of 260 amortized over five years implies a 3–4ppt drag, vs. an original 1–2ppt trend improvement), and +1ppt to R&D ratio from AI. On a US$30 bn market cap, that implies the legacy business trades at ~11x adj. P/E, within the current valuation range for China e-comm and social platforms.

In other words, the market cap has not fully priced Kling, but legacy drag keeps sentiment cautious near term. Under a base case, valuing legacy at 8x P/E implies ~US$20 bn (profit RMB 17.5 bn, -15% YoY; 8x is not cheap in a difficult period). Assuming Kling delivers US$350 mn revenue this year per management’s doubling guide, applying 30–40x P/S in line with peers yields US$10.5–14.0 bn. That leaves ~3–15% upside vs. the current level.

On a more cautious take, given the discount typically applied to single-model companies and a strong rival in Seedance, a 20–30x P/S for Kling implies US$7.0–10.5 bn. Summing up to US$27.0–30.5 bn vs. the current US$29.5 bn market cap, there remains risk of sentiment-driven de-rating.

Detailed charts below