On November 21st, the "Science and Technology Innovation Board Daily" (Reporters Xu Cihao, Researcher Liang Youyun) reported that Kuaishou released its financial report for the third quarter of 2024 after the closing of the market on November 20th.

The financial report shows that Kuaishou's third-quarter revenue increased by 11.4% year-on-year to 31.131 billion yuan, with adjusted net profit increasing by 24.4% to 3.9 billion yuan. In the first nine months, Kuaishou's cumulative revenue was 91.514 billion yuan, a year-on-year increase of 13.1%. During this period, the profit was 11.37 billion yuan, a year-on-year increase of 308.0%, and the adjusted net profit was 13.015 billion yuan, a year-on-year increase of 120.3%.

After the opening of the Hong Kong stock market today, $KUAISHOU-W (01024.HK)$ a sharp decline, once dropped by nearly 13%, now the decline has narrowed to within 10%, trading at HK$47.6, with a turnover of over 3 billion Hong Kong dollars.

It is worth mentioning that AI has been a key focus for Kuaishou this year. During the financial report conference call, Kuaishou's founder and CEO, Cheng Yixiao, stated that the monthly revenue from intelligent AI has exceeded tens of millions of RMB, and they plan to explore more diversified monetization models, expecting rapid growth in revenue scale next year.

In addition to the video large model generation, Cheng Yixiao mentioned at the earnings conference that in the third quarter, Kuaishou continued to optimize the performance of the pedestal large model, deepening the applications of large models in content understanding, recommendation, production, and user interaction. The upcoming MoE model significantly reduces the cost of model training inference. Currently, Kuaishou's large model technology has been implemented in scenarios such as short videos, live streaming, comments, commercialization, and e-commerce.

Average daily active users exceeded 0.4 billion.

As an important sector of Kuaishou, the e-commerce business showed a 15.1% year-on-year growth in the third quarter, reaching 334.163 billion yuan in GMV. The average monthly number of e-commerce buyers increased to 0.133 billion, with the proportion of pan-shelf in the total e-commerce GMV rising to 27%. The average monthly active sellers increased by over 40%, and the number of new merchants joining increased by over 30%.

Apart from the e-commerce business, within Kuaishou's core commercial revenue, the contribution of marketing service revenue increased, reaching 17.634 billion yuan, a 20% year-on-year growth; the revenue contribution from live streaming business decreased slightly to 9.338 billion yuan, down by 3.9% year-on-year; other services contributed revenue of 4.159 billion yuan, a 17.5% year-on-year growth.

In the third quarter, the average daily active user count of Kuaishou reached 0.408 billion, an increase of 5.4% year-on-year; while the average monthly active users reached 0.714 billion, an increase of 4.3% year-on-year. The average daily usage time per user also reached 132.2 minutes, a 7.3% year-on-year growth. In addition, the daily average live streaming and short video playback times approached 110 billion.

Furthermore, since the beginning of this year, Kuaishou has been improving the product infrastructure and business model of the paid short video industry, creating a Native original link based on the service account, enabling users on the Kuaishou platform to watch dramas and make payments without the need for technical development by the streaming party, only requiring simple episode uploads, layout configurations, etc.

It is reported that Kuaishou has implemented the IAA model for short dramas. The IAA model allows advertisers to choose to unlock more episodes through incentive and feed ads to C-end users, with Kuaishou sharing the streaming fees with short drama advertisers in a certain proportion. The IAA model not only allows customers to acquire more incremental converted users but also adds secondary monetization value to their existing drama assets.

At the earnings conference held this evening, Kuaishou Technology's founder and CEO Cheng Yixiao revealed that with the vigorous development of the short video industry, the marketing budget for short videos has increased significantly. In the third quarter, the marketing expenditure of Kuaishou's short videos increased by over 300% year-on-year, with daily peak consumption in recent days exceeding 40 million.

Kuaishou's AI, KelinAI, broke through tens of millions in monthly revenue.

"AI has been a key focus of Kwai in recent years," said Kwai's CFO Jin Bing at tonight's earnings conference.

Since the introduction of the video AI model KelinAI in June this year, it has served over 5 million users, generated 51 million videos, and over 0.15 billion images.

Recently, KelinAI by Kuaishou announced the addition of the v1.5 high-quality model and video extension feature to its API (Application Programming Interface) service, supporting 1080P HD video output, with multiple extensions totaling up to 3 minutes.

Kuaishou Technology's founder and CEO Cheng Yixiao revealed tonight at the earnings conference that the user retention of Kelin AI has been increasing month by month, and achieved over 1.5 million monthly active users in September.

In addition to the large-scale video generation model, Cheng Yixiao mentioned at the earnings conference that in the third quarter, Kuaishou has been continuously optimizing the performance of the large-scale base model, deepening the application of the large model in content understanding, content recommendation, content production, and user interaction. In terms of the base model, Kuaishou is developing the Kuayi large model MoE model, with total parameters only in the scale of several billion, several indicators in the model pre-training phase surpassing the Kuayi 175 billion model, significantly reducing the cost of model training inference, and better adapting to Kuaishou's existing business scenarios.

Regarding the specific application scenarios and data performance of the large model in the business, Cheng Yixiao introduced in the earnings report that in terms of content understanding, Kuaishou continues to promote the large model technology in short video, live streaming, comments, commercialization, e-commerce, and other scenarios, achieving progress in user retention, commercial consumption, and e-commerce GMV in multiple aspects; in terms of content recommendation, the upgraded recommendation large model technology has brought significant growth in time spent; in terms of content production and interaction, this quarter, AIGC marketing material consumption continues to rise, with daily consumption exceeding 20 million, and the interactive assistant AI Xiao Kuai has also integrated into more scenes including Olympic betting, user private messages, and comment section interactions, with a peak monthly active users exceeding 18 million.

Regarding the commercialization progress of Kelin AI that investors are concerned about, Cheng Yixiao stated that Kelin AI has successively launched a member payment system for domestic and overseas users since the end of July, providing users with a differentiated model effect experience. In late September, the API (Application Programming Interface) service was officially opened, covering modules such as video generation, image generation, and virtual try-ons. Currently, Kelin AI's monthly commercial revenue exceeds tens of millions of RMB, progressing as expected. In the future, we will explore more diverse monetization models and are confident in achieving rapid growth in Kelin's revenue scale next year.

Overseas business losses narrowed by 75.9% year-on-year.

Notably, in order to further promote globalization, Kuaishou's international version Kwai recently announced the launch of its e-commerce platform Kwai Shop in the Brazilian market. Kwai platform users can add products appearing in specific areas of short videos, live streams, or stores to the shopping cart, and complete payment and order tracking without leaving the Kwai application.

"In the past two years, one particularly important thing we have been doing is trying to replicate the Chinese business in Brazil. The core strategy of Kuaishou in Brazil is localization," said Ma Hongbin, Senior Vice President of Kuaishou and head of international business.

The financial report shows that in the third quarter, Kuaishou's daily active users in Brazil increased by 9.7% year-on-year, and the average daily usage time per daily active user increased by 4.2%.

"In terms of performance, by continuously iterating marketing experiences, further optimizing traffic mechanisms and efficiency while accessing new traffic scenarios to expand incremental marketing revenue." This was stated by Kuaishou in its financial report.

According to the financial report, in the third quarter of 2024, Kuaishou's overall overseas revenue reached 1.3 billion yuan, a year-on-year increase of 104.1%; the operating loss of the overseas division was 0.153 billion yuan, a year-on-year narrowing of 75.9%.

In June of this year, Kuaishou's Chief Financial Officer Jin Bing stated at the Saudi capital markets forum that Kuaishou plans to establish an office in Riyadh, Saudi Arabia soon.

"In addition to the Middle East, North Africa, and Brazil, Southeast Asia has also become one of Kuaishou's target markets." Jin Bing mentioned at the forum mentioned above.

An internet analyst, when anonymously interviewed by the Star Daily, stated that Kuaishou has the full capability to replicate its successful small ecosystem model overseas. Furthermore, Brazil, as a market with relatively less political risk, provides Kuaishou with good development opportunities.

In his view, Kuaishou only needs to successfully launch in several countries with specific languages to establish a presence in overseas markets. Arabic, French, Portuguese, and other language regions are all markets worth exploring. Despite the trial-and-error process, the future still looks promising. 'If Kuaishou can excel in localized operations, its path to overseas development will undoubtedly broaden.'

Editor/Lambor