Baidu Inc Earnings - Q4 2025 Analysis & Highlights
Baidu's Q4 2025 earnings call focused on the company's transformation into an AI-native business, with significant progress in AI infrastructure, applications, and autonomous driving, alongside announcements of shareholder returns and the proposed Kunlunxin chip spin-off.
Key Financial Results
Total revenue in Q4 2025 was RMB 32.7 billion, increasing 5% quarter-over-quarter.
Full year 2025 total revenue was RMB 129.1 billion, decreasing 3% year-over-year, primarily due to a decrease in legacy business, partially offset by an increase in Baidu Core AI-powered Business.
Baidu General Business total revenue in Q4 was RMB 26.1 billion.
Core AI-powered Business revenue exceeded RMB 11 billion in Q4, accounting for 43% of Baidu General Business revenue.
Cost of revenues was RMB 18.3 billion in Q4, remaining flat quarter-over-quarter.
Cost of revenues was RMB 72.4 billion in 2025, increasing 10% year-over-year, primarily due to an increase in costs related to Baidu Core AI-powered Business.
Operating income was RMB 1.5 billion in Q4 with an operating margin of 5%.
Operating loss was RMB 5.8 billion in 2025 with an operating loss margin of 5%, but excluding impairment of long-lived assets, operating income was RMB 10.4 billion.
Non-GAAP operating income was RMB 3.0 billion in Q4 with a non-GAAP operating margin of 9%.
Non-GAAP operating income was RMB 15.0 billion in 2025 with a non-GAAP operating margin of 12%.
Net income attributable to Baidu was RMB 1.8 billion in Q4 with a net margin of 5%, and diluted earnings per ADS was RMB 3.71.
Non-GAAP net income attributable to Baidu was RMB 3.9 billion in Q4 with a non-GAAP net margin of 12%, and non-GAAP diluted earnings per ADS was RMB 10.62.
Net income attributable to Baidu was RMB 5.6 billion in 2025 with a net margin of 4%, and diluted earnings per ADS was RMB 11.78.
Total cash and investments as of December 31, 2025 were RMB 294.1 billion.
Operating cash flow was RMB 2.6 billion in Q4.
Operating cash flow for 2025 was negative RMB 3.0 billion, but remained positive for the past two consecutive quarters.
Baidu General Business had approximately 29,000 employees as of December 31, 2025.
Business Segment Results
AI Cloud Infra revenue reached approximately RMB 20 billion in 2025, up 34% year-over-year, outpacing industry growth.
Subscription-based revenue from AI accelerator infrastructure grew 143% year-over-year in Q4, accelerating further from 128% in Q3.
AI Cloud Infra and AI Applications combined reached RMB 30 billion in revenue for the full year 2025.
Revenue from AI applications exceeded RMB 10 billion for the full year 2025.
Apollo Go delivered 3.4 million fully driverless operational rides in Q4, with weekly rides peaking at over 300,000.
Total rides increased by over 200% year-over-year.
Apollo Go delivered over 10 million fully driverless operational rides in 2025 alone, with a total of over 20 million rides provided to the public cumulatively.
Apollo Go's global footprint reached 26 cities worldwide as of February 2026.
AI-native Marketing Services, including digital humans and agents, sustained strong growth with revenue up 110% year-over-year.
The number of digital humans livestreaming on Baidu's platform increased nearly 200% year-over-year in December 2025.
Production costs for hyper-realistic digital humans declined to roughly one-third of previous quarter levels in Q4.
Miaoda's international version MeDo saw users globally create over 1 million AI applications as of early February.
ERNIE Assistant's MAU exceeded 200 million in December.
AI search API adoption accelerated in Q4 with call volume up over 110% quarter-over-quarter.
Capital Allocation
Baidu announced a new $5 billion share repurchase program to be executed on a regular basis in a disciplined and transparent manner.
Baidu adopted a dividend policy for the first time.
Baidu has invested over RMB 100 billion in AI since launching ERNIE in March 2023 and will continue to maintain this level of investment density going forward.
The company is exploring alternative ways of supporting financial needs, including operational and financing leasing, as well as access to low-cost interest banking borrowing with rates as low as below 2%.
Industry Trends and Dynamics
Enterprise AI adoption is accelerating, with customers integrating AI into core operations.
Demand growth is occurring in both training and inference workloads, with AI computing demand expected to keep expanding.
Robotaxi has reached a tipping point globally, with more countries and regions creating supportive environments for robotaxi operations.
The industry is expected to accelerate in 2026 as robotaxi operations expand.
Embodied AI showed notable momentum, with revenue from this vertical doubling quarter-over-quarter in Q4.
Many major cities are short of human drivers, and robotaxi services can provide additional supply while stimulating ride hailing demand.
Competitive Landscape
Baidu's full-stack, end-to-end AI architecture provides a key differentiator in capturing enterprise AI opportunities.
Baidu's AI infrastructure achieves an excellent balance across performance, efficiency, and cost, powered by a diverse mix of chips with deep expertise in heterogeneous computing and unified scheduling.
Kunlunxin AI chips deliver stable, high-performance AI computing at scale with broad compatibility across different models and frameworks, enabling customers to deploy faster with lower integration costs.
Kunlunxin chips have a proven track record of large-scale, real-world deployments with leading enterprises across diverse industries including financial services, telecommunications, energy, and Internet sectors.
Customers choose Kunlunxin chips for reliable performance, stable supply at scale, exceptional software compatibility, and strong efficiency, especially in inference workloads.
Apollo Go is a clear global leader in autonomous ride hailing with over 20 million cumulative rides and an outstanding safety record.
Apollo Go's fully driverless vehicles experience an airbag deployment accident only once every over 12 million kilometers, making it the safest globally.
RT6 is the world's first purpose-built production vehicle designed from the ground up for Level 4 autonomous driving, with a cost under $30,000 per vehicle, offering the industry's best cost structure.
Apollo Go was the first to achieve unit economics breakeven in Wuhan in late 2024.
Very few players have entered right-hand drive robotaxi markets, while Baidu has already established a presence and is making rapid progress.
Baidu's ERNIE Assistant maintains low error rates with minimal hallucinations, delivering highly trustworthy content to users.
The AI-to-C product market is highly competitive, with competitors using aggressive marketing strategies to rapidly scale user bases.
Macroeconomic Environment
No specific macroeconomic factors were discussed in the earnings call.
Growth Opportunities and Strategies
Baidu is pursuing one of the most comprehensive and ambitious AI search transformations globally, continuously improving the quality of AI search results while expanding what users can accomplish directly within search.
AI-generated infographics have been introduced into search results to visualize text-based information and make key insights immediately clear and digestible.
MCP capabilities have been integrated across key scenarios including e-commerce, health care, and local services, enabling actions such as shopping, booking, and health care consultation to be completed seamlessly within the search experience.
OpenClaw, a recently popular open-source agent framework, has been integrated directly into Baidu App with one-click access, enabling users to immediately benefit from cutting-edge agentic AI capabilities.
ERNIE Assistant has been enhanced with broader multimodal capabilities, improving the user experience.
Baidu's AI search API is opening up broader possibilities for the international market with industry-leading authority, comprehensiveness, and newly added multilingual capabilities.
Digital humans combine visual presence, voice, and real-time interaction to create more engaging and effective experiences, with leading companies including Jingdong, Zuoyebang, and TikTok partnering with Baidu.
Miaoda, Baidu's vibe coding platform, enables users without coding experience to build applications through natural language, including WeChat mini programs, websites, and mini games.
Yijian, Baidu's advanced visual intelligence platform, enables enterprises to automate operational compliance and safety checks through intelligent visual analysis.
FM Agent, Baidu's self-evolving agent, is designed to solve complex operational challenges by autonomously reasoning across data, rules, and real world constraints.
Baidu established the Personal Super Intelligence Business Group (PSIG), unifying Baidu Wenku and Baidu Drive to accelerate the rollout of new applications and foster robust growth driven by application layer innovation.
Apollo Go is expanding fully driverless operations at pace while advancing international expansion across 26 cities globally.
In the UK, Baidu advanced partnerships with Uber and Lyft, with plans to pilot autonomous vehicles in London with testing expected to begin in the first half of 2026.
In Switzerland, Baidu initiated testing in St. Gallen following market entry last quarter.
In Abu Dhabi, Baidu launched fully autonomous ride hailing services on Yas Island in January with AutoGo.
In Dubai, Baidu secured the city's first two fully driverless testing permits from the Roads and Transport Authority and announced the next phase of its global partnership with Uber to bring fully autonomous ride hailing services to Dubai via the Uber platform.
In Asia, Baidu entered South Korea, starting with the Seoul metropolitan area, further expanding its presence across the Asian region.
In Hong Kong, Baidu expanded open road testing into Tsuen Wan and initiated cross-district testing between Airport Island and Tung Chung, bringing it closer to commercial readiness.
Baidu's focus for Apollo Go is on three areas: aggressively scaling up safe and comfortable operations by deploying more vehicles, continuously improving unit economics with the goal of achieving unit economics breakeven in more cities, and expanding with flexible business models both domestically and internationally.
Baidu is leveraging diverse strategic partnerships to accelerate global expansion, collaborating with Uber and Lyft in London and Dubai.
Management takes an application-driven approach with ERNIE, where model improvements are guided by the most valuable and promising use cases.
Baidu restructured its model development organization into two dedicated teams: one advancing ERNIE's state-of-the-art foundation model capabilities, and another tailoring models for specific business needs to reduce costs and improve efficiency.
Baidu is open to using other models where they are better suited, with the goal of achieving the best application outcomes.
Financial Guidance and Outlook
Management believes the Core AI-powered Business will become the majority of Baidu General Business in the foreseeable future.
Management remains highly confident in sustaining strong growth momentum in AI Cloud in 2026.
Management expects AI Cloud Infra to maintain strong momentum with AI accelerator infrastructure continuing to serve as a core driver, propelling the overall cloud business toward a more sustainable and high-quality growth mode.
As enterprise AI deployments deepen further, management is confident that the cloud business will continue to grow faster than the industry.
Management expects more cities to achieve positive unit economics over time for Apollo Go.
Management expects the autonomous ride hailing sector as a whole to see valuations better reflect the transformative potential of this technology over time, creating meaningful upside opportunity for Apollo Go.
Management will continue to maintain AI investment density while balancing investor focus on profitability and return timelines.
Management believes that even with significant AI investment, operating cash flow will remain positive going forward.
AI Chip Strategy and Kunlunxin Spin-off
Baidu announced the proposed spin-off and separate listing of Kunlunxin, its proprietary AI chip subsidiary, after more than a decade of steadfast investment in self-developed AI chips.
Kunlunxin's AI chips are built on a proprietary architecture developed in-house from day one.
The proposed Kunlunxin spin-off and separate listing represents a milestone that validates Baidu's long-term strategic vision and unlocks new opportunities for value creation.
Kunlunxin serves as a key component of Baidu's own cloud platform's computing power, playing an important role in the overall AI infrastructure.
Kunlunxin is a result of over a decade of investment and represents a critical infrastructure component of Baidu's full-stack AI capabilities.
Management believes the Kunlunxin spin-off and separate listing will receive strong market recognition and unlock significant value for Baidu as a group.
Foundation Model Development
Baidu remains fully committed to advancing its proprietary foundation model, ERNIE.
Following the unveil of ERNIE 5.0 last quarter, Baidu launched an updated version in January.
ERNIE has clear strengths in several key areas such as creative writing, omni-model understanding, and instruction following.
Management is confident ERNIE will keep improving its performance across key application scenarios.
High-value application scenarios continuously provide ERNIE with real data and feedback, driving model iteration and making ERNIE better.
