Baidu Q3 financial report released, with total revenue of 34447 billion yuan, a year-on-year increase of 23% in net profit, surpassing market expectations What information is worth paying attention to?

Baidu’s Q3 Report: Revenue and Profit Exceed Expectations

On November 21st, Beijing time, Baidu released its financial report for the third quarter!

Let’s take a look at the specific situation ~~

Baidu’s revenue in the third quarter reached 34.447 billion yuan, with a net profit attributed to Baidu (non-GAAP) of 7.3 billion yuan, a year-on-year increase of 23%. Both the revenue and profit performance exceeded market expectations.

In the past three months, Wenyi Yanyan has been open to the whole society, and Baidu has released the Wenyi Model 4.0 version. Currently, the number of Wenyi Yanyan users has reached 70 million.

Baidu’s founder, chairman and CEO, Robin Li, said, “Against the backdrop of the emerging opportunities of generative AI and basic models, Baidu continues to strengthen its technology and improve its products. In particular, we have launched the Wenyi Model 4.0, which is Baidu’s most powerful basic model. We have also used the capabilities of the Wenyi Model and Wenyi Yanyan to reconstruct consumer and enterprise-level products, as well as our own operations, continuously improving efficiency while providing AI-native experiences.”

He pointed out, “At the same time, we have fully opened the Wenyi Model API to our cloud customers, helping them develop AI-native applications and solutions. Baidu adheres to the business and product strategy centered around artificial intelligence, laying a foundation for the revenue maintenance and profit expansion of the Wenyi Model and Wenyi Yanyan ecology in the coming years.”

Baidu CFO Herman Yu stated, “Baidu’s core maintained a stable profit margin in this quarter, and our continued investment in artificial intelligence has laid the foundation for technological and product innovation. In the future, we will continue to prioritize investment in artificial intelligence, especially in generative AI and basic models, and Baidu will firmly focus on efficiency and strategic resource allocation.”

According to the financial report, Baidu’s core revenue in this quarter was 26.6 billion yuan, with a net profit attributed to Baidu’s core (non-GAAP) of 7 billion yuan, a year-on-year increase of 21%. Baidu’s core online marketing revenue increased to 19.7 billion yuan. At the same time, the accumulated service volume of Luobo Kuai Pao increased to 4.1 million orders. It is reported that in the second quarter, Xiaodu’s shipment volume ranked first in China’s smart screens and smart speakers.

Wenyi Model API calls show exponential growth

Since the opening of Wenyi Yanyan to the whole society, in October, the Wenyi Model 4.0 was released, allowing users to experience the Wenyi 4.0 version through Wenyi Yanyan, while enterprises can call the Wenyi 4.0 API through Baidu’s Intelligent Cloud Qifan Big Model Platform. Currently, there are 70 million Wenyi Yanyan users, covering 4,300 scenarios.

The Wenyi Model and Wenyi Yanyan are opening up new development opportunities for Baidu. On the first day of its opening, Wenyi Yanyan answered 33.42 million user questions, resulting in rapid growth in the number of users. Recently, the “Wenyi Yanyan Professional Edition” based on the Wenyi Model 4.0 was launched, opening up paid usage for professional needs, which sets the Wenyi Model apart from other models on the market. Compared to the free basic version, the professional edition has stronger modeling and image generation capabilities.

At the same time, the API call volume of the Wenyi Model shows exponential growth. Not long ago, Robin Li stated in a speech that the call volume of the Wenyi Model “exceeded the sum of the calls of other 200 large models”. It is reported that there are more than 10,000 companies that call the capabilities of large models on the Baidu Qifan Big Model Platform every month.

Significant increase in usage of AI-native applications, paid conversion rate of the document library significantly improved

Recently, various indicators of Baidu’s multiple AI-native applications have shown significant growth. Based on the Wenyi Model and Wenyi Yanyan, Baidu has comprehensively rebuilt its products, launching several AI-native applications such as Baidu’s new search, Baidu’s new document library, Baidu’s “Yun Yiduo” cloud, Baidu Maps, and Baidu GBI.

The reconstructed AI-native applications have achieved significant growth in user numbers: on Baidu App, 640,000 creators have used AI-assisted creation tools to produce 14 million pieces of content, with a total of 30 billion plays; after the launch of Baidu’s “Yun Yiduo”, it has become the world’s first “personal cloud intelligent assistant” with a user base of 20 million; Baidu’s new document library AI functions have been used by more than 30 million users.

At the same time, user stickiness and activity of Baidu’s AI-native applications are rapidly increasing. It is reported that the cumulative usage of the Wenyi Yanyan suggestion feature has reached 20 million times; the cumulative usage of Baidu’s new document library AI functions has exceeded 200 million times; as of mid-October, the cumulative number of requests for Baidu’s “Chao Hui Xie” inpu

Baidu’s Q3 financial report shows revenue and net profit exceeded market expectations. (百度财报超市场预期)

On November 21st, Baidu (Nasdaq: BIDU, http://09888.HK) released its Q3 financial report, which showed that Baidu’s revenue reached 34.447 billion yuan, a year-on-year growth of 6%; the net profit attributable to Baidu (non-GAAP) reached 7.3 billion yuan, a year-on-year growth of 23%. Both revenue and net profit exceeded market expectations.

The financial report showed that Baidu Core (Baidu’s combination of search services and transaction services) generated a revenue of 26.6 billion yuan this quarter, and the net profit attributable to Baidu Core (non-GAAP) reached 7 billion yuan, a year-on-year growth of 21%. Baidu Core’s online marketing revenue grew by 5% to 19.7 billion yuan, which was in line with Morgan Stanley’s expectations. Previously, Morgan Stanley lowered its revenue expectation for this performance to 19.7 billion yuan. Non-online marketing business revenue was 6.9 billion yuan, a year-on-year growth of 6%. This segment of business includes innovative businesses such as Baidu Cloud, Xiaodu, and autonomous driving business.

In terms of the performance of segmented businesses, Baidu launched Wenchin 4.0 in October this year. Users can experience Wenchin 4.0 through Wenchin’s “One Word”. Enterprises can call Wenchin 4.0’s API through Baidu’s Intelligent Cloud Qianfan Model Platform. Currently, Wenchin’s “One Word” has reached 70 million users and covers 4,300 scenarios. Baidu established a technology ethics committee in October to guide the practice of technology professionals.

The financial report showed that Baidu’s R&D expenses for the third quarter were 6.1 billion yuan, an increase of 6% year-on-year. This was mainly due to the depreciation expenditure and server hosting fees for the servers supporting Wenchin’s “One Word” research investment.

According to the information from 【商商查】, Beijing Baidu Netcom Technology Co., Ltd. was established in 2001. The company has a total of 70 business segments, including Baidu Nani, Baidu Xiaoke, Haokan Video, Xiaodu Smart Speaker, Baidu Maps, and Baidu Wenku. Earlier, it had already ventured into the fields of artificial intelligence and AI.

It is reported that the “Wenchin One Word Professional Edition” based on Wenchin’s Large Model 4.0 was launched recently. It is available for paid use to meet the needs of professionals, which distinguishes Wenchin’s Large Model from other models on the market. Compared to the free basic version, the professional edition has stronger model capabilities and image generation capabilities. Not long ago, Robin Li said in a speech that the number of Wenchin’s Large Model calls “exceeds the total of the other 200 large models.”

In addition, Baidu disclosed that starting from the third quarter of 2023, the total amount of Baidu’s share repurchases reached 126 million US dollars, and the cumulative repurchase amount under the share repurchase plan has reached 351 million US dollars.

In February of this year, Baidu announced that its board of directors had authorized a new share repurchase plan, allowing the company to repurchase up to 5 billion US dollars of shares, which is valid until December 31, 2025. As of September 30, 2023, Baidu’s cash, cash equivalents, restricted funds, and short-term investments amounted to 202.7 billion yuan, and cash, cash equivalents, restricted funds, and short-term investments (excluding iQiyi) amounted to 197.4 billion yuan. Free cash flow was 6 billion yuan, and free cash flow (excluding iQiyi) was 5.2 billion yuan.

At the time of writing, both Baidu’s Hong Kong and US stocks have risen sharply.

The Impression and Potential of Baidu in the Stock Market

If I have some spare money next year, I might invest a little in Baidu stocks.

From 2020 to 2021, holding Baidu stocks for a period of time has yielded good returns, and my impression of Baidu has always been okay. In fact, among the mainstream internet companies, Tencent, Alibaba, and Baidu, these three old school BAT, have the most stable market trend. I don’t know why, maybe because the concept of BAT has left a deep imprint in my mind during the growth stage, and my thinking has been influenced by it.

Let me mention a few details:

The first term is called “beyond expectations.”

Baidu’s external communication in recent years, from Li Yanhong to the present, has been mostly honest, rarely making outrageous claims.

In this Q3 financial report, they described themselves as beyond expectations. Whether it’s because the initial expectations were relatively conservative, the feedback from the AI model market is indeed excellent. In any case, it’s good news.

In fact, based on my user experience (sometimes unable to use GPT), Wenzhenyiyi, especially the 4.0 version, is indeed a top-tier large-scale model system in China and even in the world.

Considering the negative news and stock price fluctuations on the day of its release this year, and looking at the market feedback now, it’s reasonable to use the term “beyond expectations” to describe it.

The second detail comes from Goldman Sachs.

Although there are always rumors about Goldman Sachs buying mansions by the sea, their judgments on the mainland in recent years are highly regarded.

This time, Goldman Sachs still includes Baidu in their Regional Conviction Buy list, and they have also explicitly stated the reasons for this rating:

1. Integrating generative AI with Baidu’s existing cloud business and upgrading its search advertising business, it has the opportunity to reshape market share allocation in the next 2-3 years.

2. Generative AI enhances the stable development of the cloud business and benefits from the incremental revenue of Mobility as a Service (MaaS). The compound annual growth rate of revenue is expected to be 23% from 2023 to 2025.

3. The healthy recovery of advertising revenue and further monetization of Baidu’s traffic through the synergistic effect of generative AI.

The trend in the capital market is usually reliable, and the situation where Morgan Stanley and Goldman Sachs are both optimistic is generally a positive signal.

The third detail comes from iQiyi.

No one expected that iQiyi, which had been dragging down Baidu’s financial reports for the past three years, would actually grow.

The financial report for 2021 is the most impressive to me. Baidu’s overall performance that year was average, but the data without iQiyi looked impressive.

Surprisingly, iQiyi’s revenue also reached 8 billion yuan (1.1 billion US dollars) this year, with a growth rate of 17%.

The burden has become an engine, and it seems that iQiyi has become the second largest video website in China.

Overall, after the slump of Chinese concept stocks in the US from 2022 to 2023, Baidu’s current situation looks promising both in terms of products and in the capital market.

Chasing after a trillion-dollar Baidu may seem absurd, like the situation before the WBGS finals, but returning to $190 is not impossible.

In my opinion, next year, Wenzhen/Xunfei might outperform other large models. WPS AI might also have a chance, who knows.

Whether Baidu can firmly secure the position of “China’s number one AI stock” still depends on whether the products can be refined.

As for reputation.

How was Pinduoduo’s reputation in the past few years? Take a look at Huang Zheng’s current worth.

No notable highlights in overall view.

As a whole, I didn’t see any particular highlights. Here is a summary of the third-quarter data:

• In terms of revenue, compared to the same period last year, it increased by about 2 billion, and compared to the previous quarter, it increased by about 400 million.
• In terms of costs, sales and administrative expenses increased by about 600 million compared to the same period last year, and decreased by about 500 million compared to the previous quarter.
• In terms of costs, research and development expenses increased by 400 million compared to the same period last year, and decreased by nearly 300 million compared to the previous quarter.

Note: The reason for the year-on-year increase in research and development expenses is due to the increase in server depreciation expenses and server hosting fees for supporting the research and development of Wenzin Corp. What is the reason for the decrease compared to the previous quarter? Basically, is it because we couldn’t buy NVIDIA cards in the third quarter?

• In terms of other income, the “other” category had a loss of 4.48 billion last year and made a profit of 1.1 billion this year, which is probably due to changes in the stock market leading to fair value changes in stock investments.

Baidu Q3 Financial Review: AI Integration and Potential

Baidu Q3 Financial Report Quick Review: AI AI AI and More

Everyone is focused on the annual drama in the AI industry, titled “What to Do When the Palace Intrigue is already over when you wake up?”

OpenAI’s Game of Thrones tells us that sometimes accomplishing something big is a complex system, a mix of various strange and unique factors coming together like a chemical reaction. Whether it’s the right timing, the right circumstances, or the gathering of seven dragon balls to summon Shenlong, the process is intertwined and interconnected. It’s difficult to say that it is simply the victory of one person or one idea. It may involve pure idealistic emotions, noble altruistic aspirations, or the pursuit of commercial interests driven by money. It may even involve competition and espionage. Regardless, sometimes when you act, if God opens a door for you, He will also open a window for you, and you will succeed. But speculating on motives is the lowest and can easily lead to conspiracy theories.

Stories about scientists battling with capital groups are like soap operas, just like the cheesy plotlines in articles that cheat and excite the masses. They are laughable and belong to trash literature.

In summary, for AI to succeed, unity of thought is not necessary, but unity within the team is. I still hold the view that if AI is still far away from Level 4 autonomous driving, it is hard for me to believe that AI will awaken and destroy humanity. It’s like making a slingshot and then starting to worry that it’s not enough, that I have to create an atomic bomb, which is terrifying. This is not a problem that the current generation should ponder.

In fact, it doesn’t take a brilliant scientist to know that from beginner to expert, AI is an accumulation process. It is a very difficult and long-term process. May Huizi once said, “Anything worth doing is worth doing slowly.”

However, the recent OpenAI incident may have reshaped the entire AI landscape in the capital market, albeit in a small way.

Turning to Baidu’s financial report, the keyword must be AI. Baidu was among the earliest to enter the AI field, and its sense of smell is sharp. There is a story about Ilya Sutskever, the Chief Scientist of OpenAI, who was the protagonist of this incident. His teacher is Professor Hinton, and his classmate is Alex Krizhevsky. Their company, DNN Research, is actually the predecessor of OpenAI. At that time, there were four bidders for this company: Microsoft, Google, DeepMind, and Baidu. Dr. Kai Yu (who had early contact with Professor Hinton and was willing to invest) was in charge of the AI work and made an offer of 12 million RMB. However, Microsoft and Google had deeper pockets, and in the end, Google bought the three-person company for 44 million USD, even though it had no products. After that, Google Brain was formed and later integrated into DeepMind, the one known for its exceptional chess-playing abilities. I’m sure everyone knows that.

So it’s not an exaggeration to say that Baidu missed out on Ilra, even though it was probably destiny.

Regarding Baidu in the era of generative AI, at least in China, I personally believe that Baidu is the best AI company. It has continuously expanded its portfolio of AI products (from the early Wenxin Yiyan to the current product matrix) and has a full-stack architecture (from chips and deep learning frameworks to large models and top-level applications like search). My reasons for being optimistic are mainly as follows: 1. The combination of generative AI and Baidu’s existing cloud business can upgrade its search and advertising business (similar to what Microsoft is doing) and have the opportunity to reshape the market share allocation in the next 2-3 years. 2. Generative AI enhances the stable development of Baidu’s cloud business and benefits from incremental revenue from Mobility-as-a-Service (MaaS). The market revenue expectation for this segment is expected to reach 500 million USD by 2025 (assuming a compound annual growth rate of 23% from 2023 to 2025). 3. With the synergy of generative AI, Baidu can further monetize its existing traffic through advertising.

Let’s take a look at the Q3 financial report. All of Baidu’s businesses can be seen as having AI as the underlying infrastructure, supporting MEG, IDG, and whatever G.

Q3 revenue was 34.447 billion RMB (Baidu’s core revenue was 26.6 billion RMB, with the rest coming from iQiyi), and non-GAAP net profit attributable to Baidu reached 7.3 billion RMB, a year-on-year growth of 23%, exceeding expectations in terms of revenue and bottom line. Baidu’s core revenue from advertising reached 19.7 billion RMB. The cumulative service volume of Luobo Kuai Pao reached 4.1 million orders, and Xiaodu became the number one in China for smart screens and smart speakers. Wenxin Yiyan has 70 million individual users, covering 4,300 scenarios. I will focus on financial data in my valuation later, so I won’t analyze it quantitatively here. Let’s do a non-financial qualitative analysis.

First, Baidu’s AI native applications have shown significant growth in user and invocation metrics. These are important indicators of its AI combat capabilities.

Firstly, I believe that in the future, the importance of large models as the underlying operating system cannot be underestimated, regardless of who will ultimately succeed in China. Such an underlying operating system will undoubtedly be in great scarcity, as developing large models from scratch would be a tremendous waste of basic resources. In the end, the good money will drive out the bad, and the winner takes it all.

The current gap between Baidu and other companies is in the completion of its AI-native transformation. For example, Baidu has AI completing 20 lines of code out of every 100 lines. Moreover, the importance of developing AI-native applications cannot be overemphasized, just like citing literature. Currently, Baidu’s Wenxin large model has more invocations than the total of 200 other large models.

The same goes for commercialization. Whether a foundational system can thrive depends on whether developers are willing to use it. Currently, Baidu’s large models have been called by 8 million developers using plugins for Wenxin Yiyan. Apart from the developer team, there are users. On Baidu’s Qianfan large model platform, over 10,000 enterprises (serving more than 20,000 companies and covering nearly 500 scenarios) call upon large model capabilities every month. So, if you want to start a business or have a side job, instead of spending money on team expansion, why not try Wenxin Yiyan? Many pioneers in AI have already started making money on platforms like Xiaohongshu without much effort.

Secondly, large models bring additional revenue to advertising and cloud, primarily in the B2B segment. Baidu’s Intelligent Cloud Computing, with an EV/ Sales ratio of 3x (while others in the industry generally have 5-6x), is also a conservative estimate. Ignoring the premium related to “intelligence” or “AI,” the estimated revenue for 2023 is about $2.6 billion, multiplied by 3, giving us roughly $7.8 billion, which is the market value for this segment.

Large models, due to their ability to reduce costs and increase efficiency, have a strong potential to drive enterprises to pay for services. For example, the educational giant New Oriental has used Baidu Intelligent Cloud’s services to train and develop large models, improving the effectiveness of their models. And for Mianbi, they have trained and optimized multiple large models on Baidu Intelligent Cloud and have seen good results (I have spoken with them). So I believe their willingness to further pay for these services is high. It brings the most desired cost reduction and efficiency improvement for internet companies, and who wouldn’t jump at that opportunity?

The level of revenue generated by AI, whether through AI cloud, Xiaodu, or autonomous driving, will eventually grow exponentially like other businesses. This reminds me of Baidu’s bidding ranking business starting on September 20, 2001. The revenue on the first day was 1.9 RMB, 3 RMB on the second day, over 10 RMB on the third day, and over 200 RMB on the fifth day. Everyone knows what the subsequent revenue became – abundant prosperity.

Finally, let’s do a simple SOTP valuation.

For Baidu’s core search business, the traditional Core Search + News Feed, we can roughly correspond it to Baidu’s Mobile Ecosystem Group (MEG), a business more familiar to investors than the old actress Joan Lin. We were quite conservative in our valuation for this segment, giving it an EV/EBITDA multiple of 5-6x for 2023, which is already quite low considering peers' multiples being around 10-12x. Taking into account the expected growth rate, let’s discount it by 50%. I don’t know if Baidu’s AI will give this segment a second wind, but my conservative assumption is that it won’t, so I give it a multiple of 5, emphasizing conservatism. Baidu’s 2023 EBITDA is likely to be over 5 billion USD (we already have specific numbers for three quarters), which translates to a market value of around 25 billion USD. Can it be any more conservative than this? It’s quite difficult. Of course, EV = MV + debt - cash, a rough estimate, gives us a market value of equity around 26-27 billion USD. Let’s go with 260 billion USD.

Baidu Intelligent Cloud Computing, EV/Sales of 3x (typically 5-6x for peers), is also a conservative estimate. Let’s not assign a premium for “intelligent” or “AI.” The estimated revenue for 2023 (our estimation) is around 2.6 billion USD, multiplied by 3, giving us a market value of around 7.8 billion USD.

IDG’s autonomous driving + Jidu Auto (Gill AI), 3x sales, is 10 billion USD + 2 billion USD = 12 billion USD.

Xiaodu, 3 billion USD (based on the valuation from January’s financing).

Equity investments in iQiyi + Ctrip, totaling around 4 billion USD.

Net cash of 17.5 billion USD.

Adding all of this up and applying a discount (30% conglomerate discount), we get a total of around 65 billion USD. This corresponds to a target price per ADS of 180 USD; whereas the current price (latest closing price) is 108 USD.

Using the relative valuation method, even if we give a PE ratio of only 15x, assuming that Baidu’s full-year revenue in 2023 is less than 140 billion RMB and the estimated net profit is 27 billion RMB (3.8 billion USD), then multiplying by 15 gives a valuation of 57 billion USD, corresponding to 162 USD per ADS. Baidu’s valuation is like searching for something everywhere, going back and forth, moving in different directions, protruding and curling back, no matter how we value it, the conclusion is that it is undervalued. However, valuation is something you understand. A low valuation doesn’t mean it will rise in the short term. Nevertheless, we can at least say that in an expanding growth pattern, the lower the valuation, the more logical it is to invest.

---

Disclosure: The author holds a long position in Baidu.

This article does not constitute any investment advice.

The Trend of Baidu AI Applications and the Future Development of Large Models

Are there any more AI players? It’s not too late to join now.

Baidu’s Q3 financial report can be seen as an indicator: AI native applications have already become the main track in the future business field. Its popularity and speed of adoption are estimated to be no less than the popularity and speed of mobile devices today.

When Baidu’s “Wenxin Yiyun” was released, although there were a lot of jokes and ridicule on social media platforms, it did attract everyone’s attention. On the first day of its launch, “Wenxin Yiyun” answered 33.42 million questions from users. “Wenxin Yiyun” has been continuously optimized and developed. So far, version 4.0 has achieved significant improvements in understanding, generation, logic, and memory capabilities. The training algorithm efficiency has increased by 3.6 times, with a weekly training effectiveness rate exceeding 98%. The reasoning performance has increased by 50 times, and its overall level is no less than GPT-4.

Therefore, the launch of the professional version of “Wenxin Yiyun” is indeed a natural progression. It has stronger model capabilities and image generation abilities, and it is open for a fee to meet the needs of professionals. This also gives “Wenxin Yiyun” a fundamental distinction from other models in the market.

In fact, as a technology company, the development of large models will eventually need to be commercialized, and commercialization requires the ability to provide more stable and comprehensive services. Currently, there are more than 230 basic large models available in China, but “Wenxin Yiyun” is still the mainstream choice. Its usage surpasses the sum of the other 200 models, and other large models cannot compete with its plugins and API data. In recent times, Baidu has released a series of AI native applications, including Baidu’s new search, Baidu’s new library, Baidu Cloud “Yun Yi Duo”, Baidu Maps, and BaiduGBI. These applications are all based on the reconstruction of Baidu’s large model and “Wenxin Yiyun”. They have also gained widespread recognition from users. For example, in the Baidu App, 640,000 creators use AI-assisted tools to produce 14 million pieces of content, receiving 30 billion views. And after the launch of Baidu Cloud “Yun Yi Duo”, it immediately became the world’s first “personal cloud intelligent assistant” with a user base of 20 million.

These applications based on “Wenxin Yiyun” bring extraordinary experiences and values that other large models that have not yet been implemented cannot provide.

Therefore, this financial report must be viewed in conjunction with user feedback, which can basically indicate its value. Baidu’s approach to AI research and operations is completely correct, and the resulting profits are also evident. However, how to use them in the future, I think this is the focus of Baidu’s future operations.

I estimate that in the next year, all of Baidu’s product lines will be penetrated by large models and fully adapt to different scenarios. At the same time, model performance will be optimized according to different needs, while synchronously incorporating feedback from end users, establishing a solid competitive barrier for Baidu’s core business.

And “Wenxin Yiyun” is likely to become the operating system of the AI native era, just like the iOS or Android system. With the prosperity of the ecosystem, applications based on “Wenxin Yiyun” will gradually be launched, and Baidu is likely to define the next wave of AI development and become a mainstream enterprise of the AI era.

Of course, this is just my speculation. As for how it will develop exactly, it still depends on Baidu’s future operations.

Increase Ad Concentration for Higher Revenue

10 Answers, 24,000 Heat…

The concentration of advertisements is not high enough. If the concentration of advertisements can be increased to 100%, revenue can be further enhanced.

Baidu Q3 Revenue and Profits Exceed Expectations, AI Applications Growing Significantly

Baidu’s third-quarter revenue and core operating profit exceeded expectations, with significant growth in the data indicators for various native applications in AI commercialization, indicating that Baidu will continue to focus on AI in the future.

On November 21, before the US stock market opened, Baidu released its financial report for the third quarter of 2023. According to the report, Baidu’s third-quarter revenue was RMB 34.4 billion, a year-on-year increase of 6%, surpassing the market’s estimated RMB 34.21 billion.

Baidu’s net profit attributable to the company (non-GAAP) reached RMB 7.3 billion, a year-on-year increase of 23%.

In terms of profitability, the operating profit was RMB 6.3 billion, and Baidu’s core operating profit was RMB 5.5 billion, with a core operating profit margin of 21%. Adjusted EBITDA (earnings before interest, tax, depreciation, and amortization) was RMB 9.505 billion, a year-on-year increase of 7%.

Under non-GAAP accounting principles, the operating profit was RMB 7.6 billion, significantly exceeding the market’s estimated RMB 6.88 billion, and the core operating profit was RMB 6.7 billion, far surpassing the market’s expectation of RMB 4.4 billion. The core operating profit margin under non-GAAP accounting principles was 25%.

After the financial report was announced, Baidu’s stock rose 2% in pre-market trading. Since the third quarter of 2023, Baidu has repurchased shares for a total amount of $1.26 billion, and the cumulative repurchase amount under the share repurchase plan has reached $3.51 billion.

Continued growth in core revenue, increased R&D and sales costs

Specifically, Baidu’s total revenue was RMB 34.4 billion, a year-on-year increase of 6%; and its core revenue was RMB 26.6 billion, a year-on-year increase of 5%. Among them:

• Core online marketing revenue was RMB 19.7 billion, a year-on-year increase of 5%.
• Non-network marketing revenue was RMB 6.9 billion, a year-on-year increase of 6%.
• iQiyi’s revenue was RMB 8 billion, a year-on-year increase of 7%. iQiyi had an average daily total subscription member count of 10.75 million in this quarter, and the average monthly membership service revenue contributed by each member increased by 12% to RMB 15.54.

In terms of costs, the operating costs for the third quarter were RMB 16.3 billion, which remained relatively unchanged compared to the same period last year. Sales costs were RMB 16.3 billion, which basically stayed the same compared to the same period last year.

R&D and sales costs have increased. Sales and management expenses were RMB 5.8 billion, a year-on-year increase of 11%, mainly due to increased channel and marketing expenses. R&D expenses were RMB 6.1 billion, a year-on-year increase of 6%, mainly due to increased depreciation expenditures on servers supporting Wenxin Yiyan research and server hosting fees.

Significant growth in AI native application data indicators, overall mobile ecosystem remains robust

In terms of the commercialization of AI, according to Baidu, multiple AI native applications have shown significant growth in data indicators. In the Baidu App, 640,000 creators have used AI-assisted creative tools to produce 14 million pieces of content, which have received 30 billion views. The user base of Baidu Wangpan (Baidu Cloud) “Cloud Yiduo” has reached 20 million. The new AI features in Baidu Wenku (Baidu Library) have been used by more than 30 million users.

At the same time, user stickiness and activity of Baidu’s AI native applications are rapidly increasing. According to reports, the cumulative usage of prompt words in Wenxin Yiyan has reached 20 million times. The cumulative usage of the new AI features in Baidu Wenku has exceeded 200 million times. As of mid-October, the cumulative number of requests for the Baidu Input Method’s “Super Writing” has exceeded 100 million, with a user satisfaction rate of over 80%.

In the past three months, Wenxin Yiyan has been opened to the whole society. Baidu has released Wenxin’s powerful basic model 4.0. The user base of Wenxin Yiyan has reached 70 million.

In terms of the overall mobile ecosystem, Baidu App had a monthly active user count of 663 million in September, a year-on-year increase of 5%, and the pages hosted on the platform accounted for 53% of Baidu’s core online marketing revenue in the third quarter of 2023.

In terms of intelligent driving, Baidu’s autonomous ride-hailing service, RoboTaxi, provided 821,000 ride services in the third quarter of 2023, a year-on-year increase of 73%. As of September 30, 2023, RoboTaxi has provided a total of 4.1 million rides to the public.

In other businesses, the shipment volume of Xiaodu, Baidu’s smart screen and smart speaker, maintained its position as the top seller in China in the second quarter.

Baidu’s founder, Robin Li, said:

Against the backdrop of new business opportunities arising from generative AI and basic models, I am particularly satisfied with our continuous enhancement of technology and product improvement. Notably, we have launched the most powerful basic model, Wenxin 4.0. We have also reshaped our products and self-operated business targeting consumers and enterprises using Wenxin and Wenxin Yiyan, creating native AI experiences and helping users and clients improve efficiency. At the same time, we have fully opened Wenxin API to cloud enterprise customers, empowering them to develop their own AI native applications and solutions. Our AI-centric business and product strategy will continue to promote the continuous prosperity of Wenxin and Wenxin Yiyan’s ecosystem, and lay a solid foundation for long-term sustainable business growth.

Baidu CFO Herman Yu stated:

Baidu’s core profit margin remained stable this quarter. Looking ahead, we will continue to prioritize investment in artificial intelligence, especially generative AI and basic models, but we will also focus on efficiency and strategic resource allocation.

Source: WallStreetCN

No adverts needed.

Do we have to include advertisements in every single sentence? If that’s the case, who would bother looking at advertisements when they have nothing better to do?

1 Baidu’s strong data foundation.

2 The booming economic environment of the AI industry. 3 Issues and concerns with Baidu AI. Baidu has inherent advantages in the field of AI. However, if they want me to pay, I will have to think carefully.

In fact, Baidu’s data for the second quarter is even more impressive:

On August 22, Baidu released its unaudited financial report for the second quarter ending June 30, 2023. In the second quarter, Baidu achieved a revenue of 34.1 billion yuan, a year-on-year increase of 15%; Baidu’s net profit attributable to Baidu (non-GAAP) reached 8 billion yuan, a year-on-year increase of 44%.

1. Strong data foundation of Baidu.

Baidu AI has achieved significant development in the field of artificial intelligence with its powerful data collection capabilities as a search engine. Now, the number of users of Wenxin Yiyan continues to grow. In the absence of overseas AI entry, Wenxin Yiyan is expected to occupy a significant market share in China, just like Baidu Search.

Baidu AR was also based on a large data foundation in the past and developed well for a while. However, due to market changes, this track is no longer viable.

Research work on artificial intelligence has existed for half a century, but it has experienced rapid development in the past five years. In 2017, AI research accounted for 3.06% of all scientific literature that year. By September 2022, this figure had reached 5.69% for that year, becoming the fundamental driving force behind the emergence of AI applications.

2. Rapid development in the booming environment of the AI industry.

With the rapid development of artificial intelligence, the popularity of overseas chat GPT, and the booming research on big data and mobile internet economy in China, the development of AI in China has reached the stage where it can bear fruit in the past two years.

The launch of various AI products represented by Baidu will bring about significant changes and opportunities in our lives.

Baidu’s Chief Financial Officer, Luo Rong, said, “Baidu maintained a stable profit margin in the core business this quarter. Our continued investment in AI lays the foundation for technological and product innovation. In the future, we will continue to prioritize investment in artificial intelligence, especially generative AI and basic models. Baidu will firmly focus on efficiency and strategic resource allocation.” The financial report shows that Baidu’s core business revenue this quarter was 26.6 billion yuan, with net profit attributable to Baidu’s core business (non-GAAP) reaching 7 billion yuan, a year-on-year increase of 21%; Baidu’s core online marketing revenue increased to 19.7 billion yuan. At the same time, the cumulative service volume of Radish Run has increased to 4.1 million orders. It is reported that in the second quarter, Xiaodu’s shipment volume ranked first in China for smart monitors and smart speakers.

3. Issues and concerns regarding Baidu AI.

① Although the model is powerful, the quality of generated content is not high, and the support for future development is insufficient.

When it comes to writing, it can only be said to be more convenient than Baidu search and the like, but it is far from being able to replace the level required to create a complete article on its own. The content generated often resembles excerpts from websites and lacks the organization of one’s own language. This is why I am currently not considering paying to use it!

In the future, facing various obstacles set by developed countries such as the United States, it is unlikely that Baidu’s hardware and data foundation for AI development will be as easily obtained as they are now. Baidu AI is unlikely to be self-sufficient.

② Worries about collecting too much daily privacy data through AI.

I use Baidu Input Method, and the voice input conversion to text is also very useful. But I’m afraid that all kinds of data will be collected by Baidu. How to protect personal privacy is a problem.