Key points of investment
Performance outlook: We expect the company's revenue in 2024Q1 to be 31 billion yuan, the same year on year; of these, Baidu's core revenue is 23.4 billion yuan, up close to 2% year on year; iQiyi's revenue is 7.8 billion yuan, down 6% year on year. The company's non-GAAP operating profit is estimated at 6 billion yuan, of which Baidu's core non-GAAP operating profit is 5.2 billion yuan, and Baidu's core non-GAAP operating profit margin is 22.2%. The company's non-GAAP net profit is estimated at 5.1 billion yuan; Baidu's core non-GAAP net profit is 4.8 billion yuan.
We expect Baidu's core revenue to fall short of previous expectations, mainly due to pressure on the advertising business. We expect Baidu's core revenue to be 23.4 billion yuan in 2024Q1, an increase of nearly 2% over the previous year. Among them, advertising revenue is affected by macroeconomics and is expected to increase slightly year-on-year. Mainly in the automotive, real estate and other industries, customer advertising performance is weak, and travel/medical/e-commerce industries are relatively good. The cloud computing business is expected to grow by a high number of units over the same period last year, and the share of AI continues to increase. The non-cloud and non-advertising business is exploring development directions in the short term, and we expect revenue to be under slight pressure. We expect the trend in the first half of the year to be similar to Q1. Overall advertising revenue will be under pressure, and the cloud computing business will contribute a major increase.
The AI business continues to advance, and I am optimistic that it will drive the company's cloud business to continue to grow. 1) The company's AI products and services are progressing steadily. According to Baidu's disclosure at the 2024 AI Developers Conference in April, Wenxin's average number of users and daily API calls have exceeded 200 million, the number of service customers has reached 85,000, and the number of native AI applications developed using the Qianfan platform has exceeded 190,000. Compared with a year ago, the algorithm training efficiency of Wenxin's big model was increased 5.1 times, the average weekly training efficiency reached 98.8%, inference performance increased 105 times, and inference costs were reduced to 1%. 2) The model matrix is gradually being improved, which is expected to serve more types of customers/application scenarios. Baidu's big model has 5 new versions, focusing on lightweighting, cost performance, and actual scenarios. Among them, the version with the smallest number of parameters, Tiny emphasizes extreme low cost, low latency, and high concurrency, and is suitable for scenarios such as end-side inference. The company also launched ERNIE Characters and ERNIE Functions for vertical scenes, which can be used for role-playing application scenarios and tool call scenarios, and has reached cooperation with smart hardware vendors and travel apps respectively.
3) We are optimistic that the company's AI model capability improvement and scenario expansion will drive cloud computing revenue growth. As of March '24, Baidu Smart Cloud has served 80,000 enterprise users, helping users refine 13,000 large models and develop 160,000 large model applications. In the future, as the model matrix is improved to explore more customer needs, the cloud business is expected to continue to contribute incrementally.
Profit forecast and investment rating: We are optimistic about the long-term development of the company's AI business, but considering that the overall advertising business in the first quarter is still under pressure, we lowered the company's 2024-2026 revenue forecast from 1432/1524/1607 billion yuan to 1396/1500/158.5 billion yuan, and lowered non-GAAP net profit from 290/315/35.2 billion yuan to 288/311/34.5 billion yuan, corresponding to non-GAAPPE in 2024, 9 times, maintaining the “buy” rating.
Risk warning: policy risk, risk of technology development falling short of expectations, AI ethical risk, autonomous driving, technical risk, market competition risk.