The new inflection point driven by AI in Kuaishou's financial report
On August 21st, Kuaishou Technology officially released its financial report for the second quarter ending June 30, 2025. Strikingly, in this quarter, Kuaishou achieved record - high levels in both total traffic and profitability.
According to the financial report, during the reporting period, the revenue increased by 13.1% year - on - year to 35 billion yuan, and the adjusted net profit increased by 20.1% year - on - year to 5.6 billion yuan. Both the gross profit margin and the adjusted net profit margin reached historical peaks, hitting 55.7% and 16.0% respectively. Meanwhile, its average daily active users reached 409 million, a year - on - year increase of 3.4%, and the average monthly active users reached 715 million, a year - on - year increase of 3.3%. The total user usage time increased by 7.5% year - on - year, indicating a new breakthrough in overall traffic.
In the official press release of Kuaishou, such achievements are attributed to "the dual - wheel drive of AI and the ecosystem."
This is in line with the shift of the market's focus. In fact, the perspectives of the industry and capital have shifted from past user growth and e - commerce GMV to the driving effect of AI. In the past few years, Kuaishou established its industry position through the traffic dividend and the dividends of the short - video and live - streaming e - commerce sectors. However, against the backdrop of reaching the traffic ceiling and intensifying competition, a new growth logic is being reconstructed.
Its progress in AI is regarded as the core indicator of whether Kuaishou can open up growth space in the next stage.
This shift is not an isolated event. Almost all global Internet companies are facing the same question: how to translate AI into practical business applications and turn technological concepts into real revenues and profits. At Kuaishou, the answer to this question is becoming clearer: AI can either be a new independent business or a fundamental reconstruction of existing businesses.
I. Leapfrog Development Driven by AI
In the eyes of external investors, the value of AI to Kuaishou presents a "dual - curve" pattern.
On the one hand, Keling AI's rapid growth as an independent business is becoming a new anchor point for Kuaishou's valuation. On the other hand, AI is also transforming Kuaishou's core business in a more intrinsic and in - depth way.
As an independent business, Keling AI was born in 2023, a time of intelligent emergence, and its entry was not late. After two years of expansion, it has maintained a high - frequency iteration. Its functions are not limited to video generation but have also formed a relatively complete ecological closed - loop in areas such as multi - modal video editing, smart animation canvas, and sound effect generation. It covers apps, web platforms, and in - app mini - programs, becoming a cross - platform productivity tool.
This productivity is not limited to individual users but is also penetrating deeper into the industry. In June this year, Keling AI co - produced the world's first AI - themed anthology series, "New World Loading," with Kuaishou Xingmang Short Dramas. Produced by the Outliers team, the series consists of seven unit stories covering diverse themes such as science fiction, fantasy, and history. It is currently the AIGC drama work with the widest range of themes and the largest narrative volume in the industry. The global cumulative playback volume has approached 200 million, becoming an important milestone in AIGC film and television content creation.
The latest data shows that its global users have exceeded 45 million, with a cumulative generation of over 200 million videos and 400 million pictures, and it serves more than 20,000 enterprise customers. The rapid improvement of functions and the high - frequency use by users have become a rapidly rotating flywheel. According to Kuaishou, in the second quarter of this year, Keling AI achieved revenues of 250 million yuan.
The market has always maintained a positive attitude. In the first quarter, when Keling AI's revenues reached 150 million yuan, institutions such as UBS and CICC gave it an independent valuation, with the range roughly between 3 billion and 6 billion US dollars. This growth rate means that Keling has evolved from an "ancillary product" within the Kuaishou system to a business unit capable of independently supporting market narratives.
What really determines Kuaishou's future direction is the penetration of AI into its existing ecological system. From the perspective of underlying technology, Kuaishou has obviously completed its internal infrastructure. The OneRec end - to - end generative recommendation large - model publicly disclosed by Kuaishou's technology team before is a prime example.
Content has always been the foundation of Kuaishou, and the recommendation algorithm is the nerve center supporting content distribution and reach. Different from the traditional cascaded architecture, OneRec uses a single - model approach to directly predict content that matches user preferences from the entire video library through a generative mechanism. This not only significantly simplifies the recommendation process but also endows the recommendation system with new capabilities. It can predict user preferences using AI, which not only improves the accuracy of recommendations but also significantly reduces computing power costs.
The data publicly released by Kuaishou shows that the computing power utilization rate has been significantly increased to 23.7% and 28.8%, the operating cost has dropped to 10.6% of the traditional solution, and the user retention time and user lifecycle indicators have been significantly improved. In other words, AI not only enhances the quality of recommendations but also makes Kuaishou's core business more efficient and resilient.
II. Deep Integration of AI and Business Reconstruction
Kuaishou's AI practice has shifted from technology demonstrations to full - chain business integration. The infrastructure and ecosystem are achieving self - consistency internally, becoming the key force driving the platform's commercialization and the reconstruction of the e - commerce model. This force has also spilled over to become a powerful driving force for reshaping merchants' operations.
In terms of commercialization, Kuaishou's AI practice has achieved a full - process reconstruction of the advertising chain. Overall, it has systematically transformed the commercial marketing chain in aspects such as AIGC content production, advertising placement agents, large - scale marketing recommendation models, and large - scale bid - reasoning models, and achieved phased results.
A pain point in the advertising placement industry is that traditional performance - based advertising is becoming increasingly difficult to break through in the "era of limited growth." Kuaishou's solution is to promote intelligence in the deeper links of the marketing business. Specifically, it continuously optimizes the generative bidding logic, relying on the quality of the data set and the concept of reinforcement learning, and explores strategies guided by long - term value modeling to improve the conversion rate. On the recommendation side, Kuaishou uses the content understanding and reasoning capabilities of large - language models for generative retrieval, mining the potential relationship between users' overall behaviors and advertising conversions, thereby increasing the click - through rate of advertising materials and driving a single - digit growth in online marketing revenues. For advertisers, this means that they are no longer simply "buying traffic" but directly entrusting the optimization of business results to the platform's AI system.
More importantly, Kuaishou has not stopped at the middle and back - end links but has penetrated the entire chain through a concrete AI product matrix. For example, for video content used as advertising materials, Kuaishou has integrated its AIGC capabilities to create the Cili Chuangxiang platform, which can achieve one - stop video creative production, from inspiration recommendation, script generation to video editing and digital human presentation, forming a complete chain. While significantly lowering the threshold for content material production, Kuaishou's internal calculations show that this AIGC capability can reduce the production cost of marketing materials by 60% to 70%, especially helping small and medium - sized merchants lacking content capabilities to accelerate their transformation towards digital and intelligent operations.
Take a certain merchant as an example. After using the platform, through the "hit replication + intelligent remix" model, its weekly video production capacity increased from 1,000 to 3,000. The production of these materials only accounted for 7.4% of its total human resources, becoming the core growth engine for its advertising placement on Kuaishou. In other words, AI is injecting new growth elasticity into the advertising placement end with almost exponential production capacity and efficiency.
Beyond short - videos, live - streaming is an inevitable business form. The implementation of Nüwa Digital Human Live - streaming has further lowered the business threshold for merchants. Through digital human technology, merchants can quickly set up accounts, launch multiple live - streams in batches, and flexibly test different scripts, significantly shortening the cycle of advertising placement and conversion. More importantly, AI anchors can be online 24/7 and have strong learning abilities, rapidly iterating in terms of interaction rhythm and scenario dialogue. This not only improves efficiency but also shifts live - streaming from being "human - driven" to "technology - driven," injecting new vitality into Kuaishou's live - streaming ecosystem.
At the earnings conference call after the release of the financial report, Kuaishou also revealed the latest applications of its AIGC capabilities in different industries. Taking the clothing industry as an example, Kuaishou launched the function of a two - person live - streaming room with a "host + model." The model can change clothes in real - time according to the host's explanations, solving the pain point of merchants lacking the ability to shoot with real people. After a clothing customer tried this function, the ROI increased by 15% compared to ordinary digital human live - streaming rooms, and the overall consumption increased by 35%, directly demonstrating the boosting effect of AI on marketing results.
Beyond strong live - streaming categories like clothing, digital human live - streaming has opened up new possibilities for other categories. A certain industrial belt merchant mainly sold on other platforms in the past. In July this year, it "entered the market at zero cost" on Kuaishou through digital human live - streaming. It tried different scene setups such as home and factory in the live - streaming room and had real - time interactions with fans through the comment section and audio. The conversion effect of digital human live - streaming was almost comparable to that of real people. In just 10 days, the ROI stabilized at around 1.8.
In the e - commerce segment, the value of AI is reflected in the matching of people and goods and transaction efficiency. Multi - modal large - models can automatically extract product features and, combined with scenario reasoning capabilities, transform complex product information into recommendation language that is easier for users to understand. In the second quarter, including the 618 promotion, Kuaishou's e - commerce achieved significant growth in multiple dimensions: on the one hand, the e - commerce GMV reached 358.9 billion, maintaining a double - digit year - on - year growth rate of 17.6%. On the other hand, the general shelf field made great progress, with the GMV of general shelf e - commerce accounting for more than 32% of the total e - commerce GMV during the same period.
Behind these data is the systematic role of AI in precise recommendation, intelligent customer service, and full - chain analysis. Small and medium - sized merchants have particularly benefited. In the past, they often faced problems such as low efficiency in product selection, advertising placement, and conversion. The intervention of AI products such as the digital employee π has helped them achieve higher efficiency at a lower cost.
Concrete products have become the anchor points for the deep integration of AI into the Kuaishou system and the engines for its ecological entities to share the AI dividends and evolve synchronously. According to Kuaishou's latest data, as of August, the daily average consumption of products such as Cili Chuangxiang, Nüwa Digital Human, and Digital Employee π has reached 34 million times, indicating that AI has become an everyday infrastructure in Kuaishou's business ecosystem.
These cases together outline the real - world picture of Kuaishou's AI ecosystem. More broadly, Kuaishou is transforming into an "intelligent business platform" by integrating AI into the platform's logic and becoming the "underlying productivity." Advertisers are no longer just buying exposure but deeply entrusting their business goals to the platform. Merchants no longer rely solely on manual labor and experience but rely on AI tools to complete the entire process from product selection to conversion. The role of AI here is more like a new means of production, becoming the most important underlying element in the platform economy. When the three pillars of content, commercialization, and e - commerce are reshaped by AI, Kuaishou's business model is also undergoing a structural reconstruction.
III. New Inflection Point Beyond the Financial Report
Kuaishou is at a critical turning point. From an initial content platform relying on traffic, to the full - scale expansion of commercialization and e - commerce, and now to the stage of AI penetration, its form has undergone profound changes.
Looking at Kuaishou's actions, whether it is seeking a business model after stabilizing the content ecosystem, finding a new model for matching people and goods after the expansion of the advertising business, or currently trying to break through the constraints of efficiency boundaries driven by AI... We can see that each of its changes is not a one - time strategic leap but is accompanied by years of "self - reconstruction."
The outside world often regards AI as a new narrative dividend, but for Kuaishou, it is more like a structural update. The AI path chosen by Kuaishou is not "AI for the sake of AI" but rather embedding AI as a production tool into the entire chain of users, merchants, and advertisers. From video production by creators, to the optimization of advertising placement in the e - commerce chain, and then to the improvement of efficiency in all aspects such as customer service, private domains, and recommendation algorithms, AI is redefining the platform's organizational and business efficiency. This means that Kuaishou is not just "using AI" but rewriting its relationships with users, merchants, and the capital market with AI.
The future Kuaishou will undoubtedly have a more mature ecosystem. Short - videos and live - streaming will still be the most intuitive entry points, but at the underlying level, it hopes to achieve a seamless connection "from content to transaction" through AI. Generative AI can lower the production threshold, enabling more ordinary people to create and operate. The data capabilities driven by large - models can make advertising placement more precise and e - commerce matching more efficient. And the platform - level infrastructure can stably support this content and transactions. This integrated closed - loop is gradually transforming Kuaishou from a single - traffic platform to an AI - driven digital business system.
This is precisely where the capital market is sensitive. If the core of Internet competition in the past decade was "how to distribute traffic," then in the next decade, a more important question may be "how to define efficiency." In this regard, Kuaishou is providing a new answer: while using AI to improve efficiency, it maintains a strong connection with real - user relationships. This is also its difference from other platforms. Short - videos defined the Internet landscape of the past decade, and Kuaishou hopes to become a full - chain ecosystem capable of generating content, supporting transactions, and providing business services in the next decade.
Perhaps this kind of evolution is the real reason for the capital market to bet on Kuaishou.