Kuaishou (Minutes): Kling accelerates commercialization, additional investment in AI

The following are the Minutes of the FY25Q2 earnings call for$KUAISHOU-W(01024.HK) organized by Dolphin Research. For earnings interpretation, please refer to《Can Kling Save the Day? E-commerce Kuaishou "Transforms" AI Kuaishou?》

I. Review of Core Information from the Earnings Report

1. Important Announcement: Ceasing Disclosure of GMV

a. Reason: The e-commerce business has matured (annual GMV exceeds 1.5 trillion), and a single GMV metric cannot fully reflect its health and long-term value (such as AI empowerment, user experience, etc.). Aligning with industry peers.

b. GMV will still be disclosed for Q3, Q4, and the annual report of 2025, providing the market with an adaptation period. Starting from Q1 2026, quarterly and annual GMV data will no longer be disclosed separately.

2. Cash Flow: Total cash and financial products amount to 101.9 billion yuan, with Q2 operating net cash flow at 8.5 billion yuan.

3. Shareholder Returns: Approximately HKD 1.9 billion worth of shares have been repurchased within 2025.

II. Detailed Content of the Earnings Call

2.1 Executive Statements of Core Information

1. Overall Performance:

a. Total revenue reached 35 billion yuan, a year-on-year increase of 13.1%. Adjusted net profit reached 5.6 billion yuan, with a profit margin of 16%, setting a new quarterly profit record.

b. Kuaishou App's average daily active users (DAU) reached a historical high of 409 million (year-on-year growth of 3.4%), and monthly active users (MAU) reached 715 million (year-on-year growth of 3.3%). Users' daily usage time was 126.8 minutes.

c. First Dividend Distribution: The board announced a special dividend distribution totaling approximately HKD 2 billion to reward shareholders.

2. Core Strategy: Comprehensive AI Empowerment:

a. Video Generation Large Model "Kling AI": Released version 2.1 in May, enhancing motion performance and physical simulation realism. Launched an independent app in July, providing a one-stop creation experience. Applied in advertising, marketing, short dramas, etc. Collaborated with Kuaishou Star Short Drama to launch the world's first AI-generated series, with nearly 200 million views. Q2 revenue from Kling AI exceeded 250 million yuan.

b. AI Empowerment of Core Business:

- Content Recommendation: Launched "OneRec" generative recommendation large model, enhancing user duration and retention rate.

- Advertising Marketing: Utilized large models to predict user behavior, improving click and conversion rates, contributing low single-digit growth to online marketing revenue.

- E-commerce: Optimized product information extraction and "person-goods matching" efficiency using multimodal large models. Future goal is to upgrade from meeting demand to predicting demand.

3. Online Marketing Services: Revenue reached 19.8 billion yuan, a year-on-year increase of 12.8%, with growth rate improving sequentially.

a. Growth Drivers:

- External Advertising: Strong demand from short dramas, local services, and the automotive industry.

- Closed-loop Advertising (E-commerce): Rapid growth in advertising revenue from general shelf scenarios, with increased penetration of full-site promotion solutions.

b. AIGC marketing material solutions improved conversion rates; UAX (one-stop comprehensive advertising placement) solutions accounted for approximately 65% of external advertising expenditure.

4. E-commerce Business: GMV reached 358.9 billion yuan, a year-on-year increase of 17.6%. Monthly average paid users reached 134 million.

a. General shelf e-commerce is the core growth engine, with GMV accounting for over 32%, growing far beyond the overall rate. During 618, product card GMV grew by over 50%, and search e-commerce GMV grew by over 50%.

b. Short video e-commerce GMV grew by over 30% year-on-year.

c. The number of new merchants increased by 50% year-on-year, supported by plans such as "Billion Golden Traffic" to assist small and medium-sized merchants.

d. Optimized influencer distribution system, with the number of available distribution products growing by 30% year-on-year, actively incubating small and medium-sized influencers.

5. Live Streaming Business: Revenue reached 10 billion yuan, a year-on-year increase of 8%.

a. Refined operation of core categories (such as group chat, multi-host live streaming), deepening the combination of "short video + live streaming".

b. Ecosystem Construction: The number of cooperative guilds grew by over 20%, and the number of guild hosts grew by over 30%.

c. "Live Streaming+" Empowering Traditional Industries:

- Quick Recruitment: Daily resume submission user count grew by over 40%.

- Ideal Home: Daily real estate lead volume grew by over 60%.

6. Overseas and Local Services:

a. Overseas Business: Revenue reached 1.3 billion yuan, a year-on-year increase of 20%. Core markets (such as Brazil) DAU remained stable, with continuous growth in user duration, e-commerce GMV, and order volume.

b. Local Services: GMV steadily grew, with operating losses narrowing year-on-year and sequentially. The strategy is to support high-ROI category growth by controlling dining subsidies.

- Daily saleable product volume grew by over 55% year-on-year.

- Advertising revenue grew by nearly 120% year-on-year.

7. Future Outlook

a. In the second half of the year, we will continue to invest firmly in AI, exploring its value in enhancing creator and customer efficiency and discovering new business opportunities.

b. Continuously promote the coordinated development of content and business ecosystems.

c. Based on user needs, improve operational efficiency, and achieve long-term sustainable growth with partners.

2.2 Q&A Session

Q: Kling AI's revenue accelerated growth in the second quarter. What are its main user profiles and core usage scenarios currently? In which application scenarios is there potential for breakthroughs or deepening in the future? How is the company planning Kling AI's long-term growth strategy and commercialization path?

A: Currently, Kling AI's users range from general enthusiasts to professional creators, with professional creators (such as content creators, designers, e-commerce advertising practitioners, film studios) being the main paying users. They use Kling AI to efficiently and cost-effectively generate content, advertising materials, or conduct preliminary concept design to enhance operational and creative efficiency.

Looking ahead, our strategic core is to push Kling AI into industrial-grade application scenarios such as game production and professional film production. We have already verified its potential through cooperation with the mobile game "Nishuihan" and the production of the world's first AI short drama "New Century Loading". Our future goal is to deeply participate in the core stages of game development and film production, breaking traditional production limitations and creating higher commercial and artistic value.

To achieve this goal, our strategy is:

Technically, continuously iterate models to enhance the controllability of generated results;

Product-wise, we will operate in layers, on one hand, providing industry solutions to tackle industrial-grade users in gaming and film, setting benchmark cases;

On the other hand, by launching innovative plays like "Kling Canvas" to attract general users, achieving user growth beyond circles. We are confident in the commercialization prospects of video generation large models and are committed to maintaining Kling AI's market leadership.

Q: Besides the star product Kling AI, how does AI empower Kuaishou's core business in other aspects?

A: Content Ecosystem and Recommendation: We launched the end-to-end generative recommendation large model "OneRec", which has been applied in some core content scenarios and local life services, successfully bringing improvements in user duration and retention.

Online Marketing (Advertising): AI empowerment is comprehensive. We launched AI solutions more tailored to the industry, such as the "Digital Human Dual Live Room" for the clothing industry, helping a men's clothing client's marketing consumption grow nearly 2 times. By optimizing bidding strategies and user behavior prediction through large models, we achieved more precise ad recommendations, contributing low single-digit growth to online marketing service revenue.

E-commerce Business: We are using multimodal large models to empower search, recommendation, and content generation. For example, by intelligently optimizing product image and text materials through AI, the transaction conversion efficiency of related products has improved by over 10%.

In summary, we have already seen the empowerment value of AI for Kuaishou's content and business ecosystems gradually being released, and we believe AI can bring more profound changes to the e-commerce business in the future.

Q: Looking forward to the second half of the year, which advertiser industries are expected to maintain strong growth? What strategies will the company adopt to seize opportunities in these high-growth industries?

A: In the second half of the year, we are confident in achieving growth in online marketing revenue beyond the market by increasing ad load rates and eCPM, and seizing opportunities in high-growth industries. We predict three main industries with strong growth in the second half:

(1) Regional Lead Industry (Local Life, Automotive): We will expand customer supply by deeply cultivating sub-industries such as healthcare and home furnishings, and promoting the entry of automotive OEMs and dealers, while optimizing conversion products such as private messages and forms.

(2) Content Consumption Industry (Short Dramas, Mini Games, Novels): We will replicate the high-growth model (native content operation + AI dynamic payment node adjustment) verified in the short drama industry over the past year to the mini games and novel fields, bringing new marketing increments.

(3) Internal Circulation E-commerce: This is our focus. We will strengthen the synergy between marketing and e-commerce by supporting new merchants, promoting intelligent marketing products like AIGC to lower placement thresholds, and optimizing traffic distribution strategies to enhance merchant operational willingness and efficiency. Especially in the general shelf domain, although its internal circulation ad revenue has achieved high growth in the first half of the year, its monetization rate is still far lower than the content domain, presenting a huge improvement space, which is the core growth point for future monetization.

Q: What strategies will be adopted to continue maintaining GMV growth above the industry level? Specifically, what is the current progress of shelf e-commerce? What is its future commercialization (monetization) pace?

A: Our second-quarter e-commerce GMV exceeded expectations due to our efforts on both supply and demand sides, accelerating the commercialization of shelf domains.

On the supply side, we activated merchant vitality through differentiated empowerment. For example, we helped content merchants reuse their live streaming materials in general shelf domains, and currently, the "operation hosting" tool covers over 20% of merchants; we also integrated goods through the "super link" model and dispatched influencer distribution, driving the daily active merchant count in the mall to grow by 30% year-on-year.

On the demand side, we focus on increasing user scale and repurchase rate. Through precise customer acquisition, e-commerce monthly active buyers have reached 134 million, while tools like wish lists and group buying enhance conversion, steadily increasing user repurchase frequency.

Shelf E-commerce Progress and Commercialization: This is the key to growth. We are enhancing its commercialization level in two ways:

On one hand, through distribution under the "super link" model and providing merchants with AI intelligent material optimization and other value-added tools, directly bringing new growth to commission income.

On the other hand, we introduce merchants' quality content into general shelf domains and optimize algorithmic placement, significantly improving the monetization rate of general shelf advertising.

Looking ahead, we will continue to empower merchants in the four major domains of live streaming, short video, distribution, and shelves, and explore certain growth spaces through AI to maintain growth above the industry rate.

Q: What is the company's plan for AI-related capital expenditure (CapEx) and expenses (OpEx)? Are there any new updates or changes? What is the company's target for adjusted net profit margin for the full year of 2025? Does the negative impact of AI investment on the annual profit margin still remain within the previously given "1 to 2 percentage points" guidance?

A: Due to the rapid commercialization progress of Kling, we have doubled its full-year revenue target for 2025 compared to the beginning of the year. Based on this, we decided to increase investment, and the related capital expenditure (CapEx) budget has also doubled, mainly for inference computing power. In terms of OpEx, it is currently expected to remain largely unchanged compared to the initial budget.

Despite significantly increasing capital expenditure, the impact of AI investment on profits remains controllable. This is mainly due to Kling AI achieving positive gross margin at the inference level, and the gross margin remains stable. Therefore, we maintain our previous judgment: the overall impact of AI investment on the adjusted net profit margin for the full year of 2025 remains around 1 to 2 percentage points. This also means we are confident in achieving the target of maintaining a stable adjusted net profit margin for the full year of 2025 year-on-year while increasing AI investment.

In the medium to long term, we believe that with technological optimization, the unit cost of AI will continue to decline, and the incremental value it brings to the business will create sustainable performance growth for the company.

<End Here>

Risk Disclosure and Statement of this Article:Dolphin Research Disclaimer and General Disclosure