Baidu: Can the Kingdom of Search rely on AI to defend the city?

In the first quarter, the core performance of Baidu.US slightly exceeded expectations, with the highlight being on the profit side. However, market expectations have already been adjusted due to company guidance, leading to a relatively average quality of the earnings report exceeding expectations.

Specifically:

1. Profit exceeded expectations due to personnel optimization: Profit beat came from the compression of research and development expenses. Despite an increase in AI expenses, the company was more aggressive in optimizing personnel (salaries) in the research and development team. Looking at the stock option incentives for research and development expenses, Q1 directly decreased by 24% year-on-year, indicating the extent of layoffs. As a result, Baidu's core research and development expenses directly decreased by 1% year-on-year.

2. Advertising pressure is significant: Core advertising growth was only 2.7%. Before AI truly brings a significant net increase to advertising, Baidu's traditional advertising business is mainly influenced by macro factors. Especially with the end of the dividend period of competition between giants and e-commerce, it is difficult to maintain an independent advantage.

As search advertising market share has been eroded by short videos, e-commerce, and other types of advertising, the current low single-digit share of AI advertising is far from enough. Meanwhile, Meta and Google across the border have contributed around 10% of revenue share from AI, while competition in traditional businesses has eased.

It is worth paying attention to whether the management will update their expectations on AI during the conference call. Since the beginning of the year, the traffic of Wenxin Yiyuan has continued to increase (from 100 million user registrations in December to 200 million in April). With the subsequent updates and iterations of AI assistants, as well as the increase in cognitive penetration rate, it is expected to reach and convert more users.

3. Recovery in cloud business: The AI cloud business continued to maintain a recovery trend from the previous quarter. Compared to the pull from advertising, the demand for enterprise AI applications may contribute more significantly to the revenue growth of Smart Cloud (with daily API calls exceeding 200 million in April, more than double the 50 million disclosed in February).

4. Poor performance in other smart businesses: Excluding revenue from Smart Cloud, the remaining businesses such as Xiaodu, Smart Transportation, and autonomous driving performed relatively poorly, showing a year-on-year negative growth.

5. Share buybacks and cash: Free cash flow in Q1 was 3.3 billion, an 8% year-on-year decrease, mainly due to increased capital expenditures from AI servers, chip investments, etc.

In the first quarter, Baidu's share buyback scale remained stable, consuming 230 million US dollars, corresponding to an annual dividend yield of approximately 2% to 3%. The company still has investments + cash of 26.6 billion US dollars on its balance sheet. Excluding short-term interest-bearing debts (including loans, notes, and convertible bonds), the corresponding net cash is 23.5 billion US dollars, which is still relatively abundant. Looking at the proportion of share buyback expenditure, the company still tends to retain cash for future operating investments and short-term deposits/financial products

6. Detailed Financial Data Overview

Dolphin's Viewpoint

After going through the period of dismantling walls for giants and offline recovery, Baidu's traditional advertising business seems to have returned to a "matured" stage, with growth basically following macro fluctuations. As traffic duration continues to be eroded, growth faces additional pressure. This also confirms that Baidu's future increasingly relies on AI.

So far, although Wenxin Yiyu has gained enough popularity, AI is still only filling the gap squeezed by traditional advertising by peers, and additional traction has not been clearly seen yet. Perhaps because ToB business clients make decisions slowly, such as advertisers have not yet realized/perceived the help of AI in improving conversion efficiency, so the budget allocation to Baidu still depends more on macro environment and competition for traffic duration among platforms. In this situation, it is more difficult for Baidu's advertising revenue to gain independent advantages.

As the saying goes, the lack of attractiveness in old businesses is the main reason why the market has been unwilling to give Baidu a reasonable valuation (compared to Baidu's advertising business, the market generally only gives 8-10x PE). Before AI brings meaningful contributions, a decent valuation repair is also expected to be slow to start.

On the other hand, from the perspective of overall shareholder returns, if Baidu is still in the growth stage of AI, fully utilizing the ample cash lying on the balance sheet to increase the intensity of repurchases or dividends may help raise the bottom valuation level in each round of fluctuations.

However, the window of waiting for AI to grow is also narrowing. Although Baidu's technology is currently leading the industry, with the release of open-source models by global giants, other platforms also have a great opportunity to catch up. If Baidu's leading gap does not further expand to a level that users can clearly perceive, other giants are likely to make up for the technological gap through ecological advantages.

The following is a detailed interpretation of the financial report

Baidu is a rare internet company that details its performance as follows:

Baidu Core: covering traditional advertising business (search/information flow advertising) and innovative businesses (smart cloud/DuerOS Xiaodu speakers/Apollo, etc.);

iQIYI Business: membership, advertising, and copyright transfer authorization, among others.

The separation of the two businesses is clear, and with iQIYI as an independently listed company with detailed data, Dolphin Research will also break down the two businesses in detail. Due to the offsetting items of approximately 1% (between 2-4 billion), there may be slight discrepancies between the detailed breakdown of Baidu Core data by Dolphin and the actual reported numbers, but it does not affect trend analysis.

I. Advertising Pressure is Not Small

In the first quarter, Baidu Core advertising grew by 2.7% year-on-year, basically meeting the market expectations guided by the company (compared to the guidance given after the previous quarter's financial report of +5%, which was lowered by 3 percentage points to 2%). Among them, the proportion of high-margin sponsored search ads continued to decline to 50%, down 1 percentage point from the previous quarter.

The macroeconomic situation in the first quarter was weak, with poor business expectations, resulting in overall industry growth pressure. The year-on-year growth of internet advertising was only 5%, slower than the previous quarter.

Comparing with the industry, Baidu's performance in the fourth quarter was significantly weaker than its peers. In a period of growth headwinds, the competition gap for platforms will also appear larger.

Dolphin Jun presented a one-year trend of advertising growth for non-e-commerce platforms (with expectations for Kuaishou, Weibo, and Bilibili). It can be seen that: (1) Tencent's performance is outstanding, with scale and growth; (2) Bilibili, due to its relatively small scale and recent acceleration in commercialization, is also running fast; (3) Kuaishou, focusing on non-e-commerce advertising, has experienced a period of pressure and is currently in recovery; (4) Although not as bad as Weibo, Baidu's trend of quarterly slowdown is not good, feeling the pressure of rapid growth in the first quarter of this year, similar to Focus Media.

The core reason lies in Baidu's search traffic ecosystem itself lacking advantages, with only high scale but weak user stickiness and duration. In the first quarter, the monthly user base of the Baidu App reached 676 million, an increase of 9 million compared to the previous month, showing a seasonal rise. Despite the continuous increase in user base, search market share has been declining year by year under the catalysis of short videos and live e-commerce in recent years.

Looking ahead to the second quarter, despite the popularity of cultural and tourism activities, due to the higher base last year and the strong traffic advantage of new platforms like Xiaohongshu in offline activities such as tourism, Baidu may face significant growth pressure. It remains to be seen how much incremental growth AI can bring, and attention can be paid to the progress of AI advertising in conference callsLikewise, in the second half of the year, there is hope for a recovery in growth as the base decreases and more businesses use AI advertising.

II. Smart Cloud Warms Up, Other Areas Lag Behind

In Baidu's core other businesses (non-advertising businesses), nearly 80% of revenue comes from Smart Cloud, while the remaining 20% is mainly from autonomous driving technology solutions, smart speakers, and other sources.

In the first quarter, revenue from other businesses reached 6.7 billion, a year-on-year increase of 5.7%. This growth was mainly driven by Smart Cloud achieving double-digit growth under the influence of a low base and AI. However, performance in other areas outside of Smart Cloud remains weak, with implied revenue continuing to decline year-on-year, with Xiaodu and Intelligent Transportation being the main drag.

In the first quarter, Baidu expanded its Wenxin series models, launched multiple lightweight large language models, and improved its MaaS platform tools for enterprise customers, aiding in the development of AI native applications. Looking specifically at the AI cloud business, it continues to maintain a 12% warming trend.

In other businesses, although smart driving is steadily advancing, there has been a slowdown in growth. In the first quarter, Luobo Kuaipao provided a total of 826,000 ride-hailing services, a 25% year-on-year increase (significantly slower than the 50% growth in the previous quarter) and a 2% decrease from the previous quarter. As of April 19, Luobo Kuaipao has accumulated over 6 million ride orders.

III. Workforce Optimization Offsets AI Incremental Spending

Concerns in the market about the impact of AI investment on profit margins have been partially offset by optimization efforts. Although there have been some confirmations in the first quarter, the overall impact on gross profit margin remains relatively manageable. At the AI conference in April, the company mentioned that the current inference cost is only 1% of the same period last year. However, in terms of capital expenditure, Baidu's investment in AI servers and chips is not low. While capital expenditure growth slowed down on a quarter-on-quarter basis in the first quarter, it still increased by 56% year-on-year, and will gradually be reflected in costs and R&D expenses in the future.

However, compared to operating costs, as AI business develops, expenses such as unavoidable server depreciation will increase. The key driver for the company is adjusting personnel salary costs, as mentioned in our financial report review from the previous quarter. Surprisingly, in the first quarter, R&D expenses saw a year-on-year decline, mainly due to layoffs/salary reductions as indicated by the 24% decrease in SBC for R&D personnelThe final operating profit of Baidu's core business in the last quarter was 4.5 billion, with a profit margin of 19%, a year-on-year improvement of 1 percentage point, and a seasonal increase compared to the previous quarter.

However, the one-time efficiency improvement brought about by layoffs is not what the market really wants to see, especially under revenue pressure. Just like Meta in 2022, if continuous investment does not translate into performance improvement over a long period of time, the market may even impose punitive valuations. Therefore, how to use AI to maintain its own traffic base and truly achieve a technical improvement in advertising conversion efficiency is where the real benefits of AI lie.

Dolphin Research on "Baidu" Historical Articles:

Financial Reports (showing the past year)

February 28, 2024 Conference Call "Baidu: AI Boosts eCPM and User Engagement (Baidu 4Q23 Conference Call Summary)"

February 28, 2024 Financial Report Review "Baidu: AI is the Key to Turning the Tables"

November 22, 2023 Conference Call "AI Improves Advertising Conversion, Reshaping the Cloud Market (Baidu 3Q23 Conference Call Summary)"

November 22, 2023 Financial Report Review "Can AI Support Baidu's 'New Story'?"

August 23, 2023 Conference Call "AI Investment Amortized Over Several Years, Minimal Short-Term Financial Impact (Baidu 2Q23 Conference Call Summary)"Financial Report Review on August 23, 2023: "Baidu: Cloud Rests, Advertising Supports the Facade" (https://longportapp.cn/zh-CN/topics/9286518?invite-code=032064)

May 16, 2023 Conference Call: "AI has Played a Role in Advertising Business (Baidu 1Q23 Performance Conference Call Summary)" (https://longportapp.cn/zh-CN/topics/6116889?invite-code=032064)

Financial Report Review on May 16, 2023: "Baidu: Not Just an AI Story, Advertising Also Supports the Plate" (https://longportapp.cn/zh-CN/topics/6112067?invite-code=032064)

February 23, 2023 Conference Call: "Advertising Relies on Macro, Cloud Emphasizes Quality over Quantity, Unhesitatingly Investing in AI (Baidu 4Q22 Conference Call Summary)" (https://longportapp.com/en/topics/4298055?invite-code=)

February 22, 2023 Financial Report Review: "Surviving the Epidemic, Welcoming ChatGPT, Is Baidu Ready for a Second Spring?" (https://longportapp.com/en/topics/4294048?invite-code=)

November 22, 2022 Conference Call: "Baidu: Advertising Resumes with Unsealing, AI Will Continue to Have High Investment (Summary)" (https://longbridgeapp.com/en/topics/3679023)

Financial Report Review on November 22, 2022: "Baidu: Valuation is All 'Cash', What is the Market Still Afraid of?" (https://longbridgeapp.com/en/topics/3676440)

Financial Report Review on August 30, 2022: "Advertising Relies on the Weather for Food, Reversal Still Depends on the Cloud" (https://longbridgeapp.com/topics/3390595?channel=t3390595&invite-code=276530)

August 31, 2022: "Reducing Costs and Increasing Efficiency, Paving the Way for AI Cloud and Autonomous Driving" (https://longbridgeapp.com/topics/3391254?channel=t3391254&invite-code=276530)

May 26, 2022 Conference Call: "Baidu's Intelligent Cloud and Autonomous Driving Business Return to Market Focus (1Q22 Conference Call Summary)" (https://longbridgeapp.com/topics/2671980) May 26, 2022 Financial Report Review "Bright Future Ahead, Is Baidu's Turnaround Near?" at Longbridge App

March 1, 2022 Conference Call "Relying on Advertising Now, Baidu Envisions a New Business Symphony" at Longbridge App

March 1, 2022 Financial Report Review "Advertising Still in Purgatory, But Baidu's 'Change of Heart' May Have Potential" at Longbridge App

In-depth

December 21, 2022 "Is Consumer Spending Warming Up or Cooling Down? The Unstoppable Spring of Advertising" at Longport App

March 17, 2021 "Digging into Baidu's Finances: How Much Room for Revaluation Does 'Hong Kong Version' Baidu Have Left?" at Longbridge App

Risk Disclosure and Disclaimer for this Article: Dolphin Research Disclaimer and General Disclosure