On August 21, KUAISHOU-W (01024) announced its second quarter 2025 financial results. Driven by AI technology and ecosystem development, the company achieved record highs in both total traffic and profitability this quarter. Revenue increased 13.1% year-over-year to RMB 35 billion, while adjusted net profit surged 20.1% year-over-year to RMB 5.6 billion. Both gross margin and adjusted net profit margin reached historical peaks at 55.7% and 16.0% respectively.

This quarter, leveraging high-quality user growth strategies, optimized traffic allocation mechanisms, strong community interaction atmosphere, and differentiated content community ecosystem, KUAISHOU-W achieved record total traffic levels. Average daily active users reached 409 million, up 3.4% year-over-year, while average monthly active users reached 715 million, up 3.3% year-over-year. Total user engagement time increased 7.5% year-over-year.

In terms of revenue structure, online marketing services, live streaming and other services (including e-commerce and Kling) served as core revenue sources, contributing 56.4%, 28.7%, and 14.9% of total revenue respectively. Additionally, overseas business performed steadily with Q2 revenue reaching RMB 1.3 billion, up 20.5% year-over-year.

KUAISHOU-W founder and CEO Cheng Yixiao stated: "In Q2 2025, our AI large model capabilities were applied across numerous platform ecosystem scenarios, not only improving operational efficiency for merchants and creators but also providing more immersive creative experiences for content creators. Furthermore, Kling AI continues to iterate and maintain its industry-leading position, offering differentiated video generation solutions for creators and driving rapid growth in Kling AI's commercialization scale. Looking ahead, we will continue exploring how AI technological transformation can empower KUAISHOU-W's business scenarios to create high-quality growth and generate long-term value for our users, partners, and shareholders."

**Model and Application Capabilities Continue Evolving: From Smart AI to Practical AI**

As a technology-driven company, KUAISHOU-W maintains deep investment in AI technology. This quarter saw breakthroughs in both large model technology and application scenarios, with comprehensive upgrades from underlying algorithms to upper-layer applications, achieving evolution from "smart AI" to "practical AI."

Regarding recommendation models, KUAISHOU-W launched the industry's first OneRec end-to-end generative recommendation large model. This innovative technical paradigm breaks traditional cascaded architectures, using one model to predict video content matching user preferences through generation from the entire video library. Thanks to architectural advantages, it can effectively introduce reward mechanism-based preference alignment methods, utilizing reinforcement learning to significantly enhance recommendation model effectiveness. Meanwhile, OneRec dramatically improves training and inference computing efficiency for recommendation large models, now widely applied in KUAISHOU-W's content recommendation scenarios, effectively enhancing user engagement time and retention rates.

Kling AI maintains rapid iteration pace. On April 15, Kling AI announced another base model upgrade, officially releasing Kling 2.0 video generation model and Ketu 2.0 image generation model globally. The new models significantly improved generation quality in semantic response, dynamic quality, and aesthetic appeal, while comprehensively upgrading controllable video and image generation and editing capabilities. On May 29, Kling 2.1 series models went online, generating 5-second videos in high-quality mode in under one minute, leading industry standards.

At the application level, Kling AI continuously iterates, successively launching "video sound effects," "multi-image reference," "dynamic canvas," and "new beginning and ending frames" functions. The video sound effects function generates high-quality, spatially rich stereo sound effects while creating videos. Multi-image reference effectively maintains consistency in characters, subjects, and scenes. Dynamic canvas functions like a creator's one-stop creative workstation, integrating unlimited visual space, intelligent creative assistance, and multi-person real-time collaboration for seamless, efficient creative experiences. The new beginning and ending frames feature achieves more fluid "cinematic-level" camera control, smooth natural transitions, and precise complex semantic understanding.

Through deep exploration in application scenarios, Kling AI has become the preferred tool for professional creators. In film and television content creation, Kling AI partnered with KUAISHOU-W Starmang Short Drama to co-produce the world's first AI anthology series "New World Loading," produced by Outliers team with 7 episodes covering science fiction, fantasy, historical and other diverse themes, achieving nearly 200 million global views and becoming an important milestone in AIGC film content creation. Additionally, Kling AI launched the "NEXTGEN New Image Investment Plan," providing support for AI creators in series production, promotion, and IP development.

From individual creators to enterprise clients, Kling AI demonstrates strong commercialization potential across multiple fields. Data shows Q2 2025 Kling AI operating revenue exceeded RMB 250 million, with accelerating monetization.

**Business Engine Upgrade: AI's Endogenous Power Releases E-commerce and Commercialization Growth Momentum**

Deep AI technology application not only enhances KUAISHOU-W's technical capabilities but more importantly injects new growth momentum into e-commerce and commercialization businesses.

In e-commerce, Q2 2025 KUAISHOU-W e-commerce GMV increased 17.6% year-over-year to RMB 358.9 billion. KUAISHOU-W continues deepening integration and efficient connection between quality content and quality products, further meeting diverse user consumption needs and driving e-commerce breakthroughs across multiple dimensions.

Q2 average monthly e-commerce buyers reached 134 million, continuing year-over-year growth, while active e-commerce user repurchase frequency also increased year-over-year, demonstrating dual enhancement in platform user stickiness and consumption activity.

In advancing comprehensive operation strategy, the pan-shelf field became a new growth engine. Data shows Q2 2025 pan-shelf e-commerce GMV exceeded 32% of total e-commerce GMV. During 618 shopping festival, pan-shelf product card GMV increased over 50% year-over-year, while search GMV surged over 140% year-over-year.

For new merchants in pan-shelf fields, KUAISHOU-W deepened marketing management tool applications, efficiently reusing content materials and product information to help new merchants achieve one-stop basic marketing operations. For larger-scale existing merchants, the platform provides richer operational resources and strategies like "Big Brand Big Subsidy" and "Super Connection." Q2 mall field daily active merchants increased 30% year-over-year.

Short video e-commerce also performed impressively, with Q2 GMV increasing over 30% year-over-year, mainly benefiting from increased supply of cart-attached short videos while improving supply quality through live streaming hit product return videos, effectively strengthening matching efficiency between short video e-commerce target audiences and core product portfolios.

Furthermore, improved new merchant support systems injected continuous vitality into the platform ecosystem. Benefiting from expanded new merchant scale and optimized field construction, active small and medium merchant numbers continued growing year-over-year. Through gold fighting plans, service provider empowerment, field traffic support and other methods, the platform helps merchants quickly adapt to platform environments and achieve individual scale growth, driving Q2 new merchant registration scale up 50% year-over-year.

AI technology also plays key roles in e-commerce full-chain operations, with large models efficiently extracting product information and user interests to restructure person-product matching efficiency. AIGC content production, intelligent broadcasting, intelligent customer service and other AI tools significantly reduce merchant costs for entering content e-commerce.

In online marketing business, KUAISHOU-W utilizes large models to mine correlations between user comprehensive behavior and marketing service conversions, directly predicting and generating marketing materials users find interesting end-to-end, improving marketing material click-through rates and bringing low single-digit growth in online marketing service revenue.

This quarter, KUAISHOU-W online marketing service revenue reached RMB 19.8 billion, up 12.8% year-over-year, with growth rates improving compared to Q1.

In external circulation marketing services, content consumption, local life services, and automotive industries became important growth drivers. In content consumption, short drama industry advertising consumption maintained high double-digit year-over-year growth. KUAISHOU-W improved short drama user payment willingness and viewing time through IAP intelligent pricing and IAA marketing node dynamic adjustments, expanding marketing clients' overall returns.

Local life and automotive industry clients mainly focus on lead generation objectives, with KUAISHOU-W's native private messaging products helping clients efficiently reach users and improve conversion rates.

Intelligent advertising product penetration continues rising, providing strong support for marketing efficiency improvements. UAX fully automated advertising products maintain rising penetration, with UAX consumption accounting for approximately 65% of total external circulation consumption.

**Online-Offline Integration: Serving People's Livelihood and Empowering Real Economy**

Today's KUAISHOU-W has become an everyday window for ordinary people to embrace digital life. Through technological empowerment and scenario innovation, online content ecosystems deeply integrate with offline real services, meeting diverse needs of over 700 million users.

Take live streaming big stage as an example - this model expands traditional live streaming rooms' indoor fixed, single-host scenarios to hundreds of square meters of indoor and outdoor venues. Through "offline staging + online live streaming," it not only rapidly attracts offline popularity but transforms high traffic into consumption power, activating urban night economy and driving coordinated growth across dining, retail, cultural tourism, entertainment and other business formats while enhancing city brand influence.

In July, KUAISHOU-W launched the live streaming big stage hundred cities plan, aiming to create "10,000-person live streaming markets" through collaboration among cultural tourism departments, live streaming platforms and hosts, building sustainable "live streaming night economy" ecosystem. Currently, this project has landed in 57 cities across Hebei, Guangdong, Anhui, Shandong, Jiangsu and other provinces.

In local life business, Q2 KUAISHOU-W optimized user experience through intelligent subsidies and refined city operations, providing users with cost-effective products. Meanwhile, it continuously optimized local lead-type advertising products to meet diverse client advertising needs. Data shows Q2 2025 local life business GMV grew steadily year-over-year, with revenue increasing nearly 120% year-over-year.

Looking ahead, KUAISHOU-W will continue adhering to "users first," deepening AI applications across full business chains to continuously create long-term value.