AI / LLMChina
← Sticker Wall
JOURNEY · DETAIL
Baidu: A Long-Distance Runner Transforming from Search Engine Hegemon to an AI Cloud and Autonomous Driving Full-Stack Player
Founded: Robin Li, Eric Xu · Baidu Online Network Technology (Beijing) Co., Ltd.
JOURNEY
Key Fields
FIELD STAMPSIndustryAI / LLM
RegionChina
ScaleGiant
ChannelOther
Origin
In 2000, Robin Li returned to China from Silicon Valley with a patent for hyperlink analysis to start his own business. Initially, it served as a backend supplier providing search technology to portals. After realizing the limitations of the technology licensing model, he decisively pivoted in 2001 to launch an independent search engine directly for users. At the time, Google stood in the way, and the Chinese search experience was far from mature. Betting that the Chinese search market would be dominated by local players, Li built a moat using localized word frequency processing and content ecosystems like Baidu Tieba and Baidu Zhidao, eventually becoming the primary traffic gateway for the Chinese internet.
Milestones
2000
Founding and Model Transformation Turning Point
In January 2000, Robin Li and Eric Xu founded Baidu at the Peking University Resource Hotel. The initial model was selling search technology licenses to portals like Sina and Sohu for about $80,000 a year, which had a clear growth ceiling. In 2001, despite opposition from the board, Li pushed for a pivot to an independent Chinese search site and launched pay-for-placement advertising. This was the first life-or-death decision, with investors threatening to withdraw funding.
2005
IPO and Traffic Hegemony PMF
On August 5, 2005, Baidu debuted on NASDAQ with an offering price of $27. It surged 354% on its first day to close at $122.54, setting a record for the first-day gain of an overseas company at the time. After Google exited mainland China in 2010, Baidu's share of Chinese search exceeded 70%. Search advertising became a cash cow, with 2010 annual revenue reaching 7.915 billion RMB, a year-on-year increase of 78%.
2013
Early Bet on 'All-in AI' Inflection Point
In 2013, Baidu established the Institute of Deep Learning (IDL) with Robin Li as dean. In 2014, Andrew Ng was hired as Chief Scientist to form the Baidu Silicon Valley AI Lab. At the time, AI commercialization seemed distant, and the heavy investment drew significant skepticism. However, this laid the decade-long foundation for Apollo and the Ernie Bot, despite internal concerns that these investments were dragging down financial reports.
2016
Wei Zexi Incident Damages Trust Failure
In April 2016, the Wei Zexi incident erupted. Baidu's medical advertising bidding model faced unprecedented public outcry. A joint investigation team ordered reforms, leading to the mass termination of medical advertising clients and a sharp slowdown in ad revenue growth. This crisis exposed the moral and regulatory risks of a single bidding-based business model, forcing Baidu to accelerate its search for a second growth curve.
2017
Apollo Open Platform Launch Turning Point
In April 2017, Baidu released the Apollo autonomous driving open platform. Under Lu Qi's leadership, the company strategy shifted from 'All-in O2O' to 'All-in AI,' shedding peripheral businesses like food delivery. However, in July 2017, Robin Li's ride in an autonomous vehicle on the Fifth Ring Road resulted in a traffic ticket, highlighting the gap between regulation and reality. After 2019, Apollo's commercialization stalled, and executives Wang Jin and Lu Qi left, leading to mockery of autonomous driving as 'PPT-based car manufacturing'.
2019
Smart Cloud and Autonomous Driving Commercialization Ramp-up Growth
In 2019, the Baidu AI Cloud brand began independent operations to target government and enterprise markets, competing directly with Alibaba Cloud and Huawei Cloud while focusing on AI differentiation. In 2021, Apollo Go launched normalized paid operations. In August 2022, it received fully driverless commercial pilot licenses in Wuhan and Chongqing, becoming one of the first in China to operate without safety drivers, validating that Baidu's decade of investment in autonomous driving was not in vain.
2023
Ernie Bot and AI Revenue Milestone PMF
In March 2023, Baidu rushed to launch Ernie Bot, the first ChatGPT-like product from a major Chinese tech firm. Robin Li admitted the launch wasn't perfect but noted the market couldn't wait. By the 2026 interim report, AI-related revenue reportedly reached about half of the total. Management stated they would continue to invest and push for a primary listing in Hong Kong by the end of the year, with the goal of returning the Ernie Bot to the top tier.
Turning Points
- 2001: Abandoned the stable business of selling technology to portals, pivoted to an independent search engine, and built its own traffic gateway.
- 2010: Google exited mainland China, allowing Baidu to capture over 70% of the Chinese search market.
- 2016: The Wei Zexi incident forced Baidu to downplay medical bidding and accelerate the search for an AI-driven second growth curve.
- 2023: Rushed the launch of Ernie Bot, betting the company's next decade on large models.
Failures & Pitfalls
- Baidu Waimai, built with heavy spending during the O2O expansion era, was eventually sold at a loss and integrated into the Ele.me system.
- Early start, late arrival in smart cars: Partnerships for mass production with companies like WM Motor fell short of expectations, viewed internally as a major failure in smart vehicle strategy.
- The departure of key figures like Lu Qi, Wang Jin, and Andrew Ng led to multiple leadership changes during the AI transition, causing strategic instability.
- The medical bidding model was completely discredited during the Wei Zexi incident, causing long-term damage to brand trust.
关键成功要素
- Over a decade of sustained investment in deep learning and autonomous driving, creating a complete technology stack and patent moat.
- Using stable cash flow from search advertising to fund AI business, surviving long periods of losses.
- Adopting a dual-track strategy of open ecosystems and Robotaxi operations to mitigate the risk of single-point failure.
- Proactively reducing the share of advertising revenue after the crisis, restructuring revenue through AI Cloud and large models.
Lessons
- Technological bets must be viewed on a ten-year scale; short-term financial pressure and long-term moat building must be addressed simultaneously for capital markets.
- The moral risks of a single business model will eventually erupt; the lessons from pay-for-placement advertising are profound.
- Being an early mover doesn't guarantee success: Without a closed loop for mass production and operational deployment, early advantages can be overtaken by latecomers.
- Founders driving the transformation personally is more effective for maintaining strategic continuity than relying on parachuted professional managers.
Core Data
- 2023 Annual Revenue：Approx. 134.6 billion RMB (based on public data, not independently verified)
- 2022 Annual Revenue：Approx. 123.7 billion RMB (based on public data, not independently verified)
- Net Profit Performance：2023 net profit attributable to Baidu was approx. 20.3 billion RMB (based on public data, not independently verified)
- R&D Investment：Annual R&D investment consistently exceeds 20 billion RMB, accounting for nearly 20% of revenue (based on public data, not independently verified)
- Number of Listings：2005 (based on public data, not independently verified)
- 2026 AI Business Revenue Share：Reportedly reached nearly half according to interim report analysis (based on public data, not independently verified)
- 2005 IPO Price and First-Day Closing Price：Issued at $27, closed at $122.54 on the first day (based on public data, not independently verified)
Competitors / Peers
In search, Baidu long suppressed 360 Search and Sogou, but is now seeing traffic diverted to ByteDance's Douyin Search and WeChat Search. In the large model race, it faces price wars from Alibaba Cloud's Tongyi Qianwen, ByteDance's Doubao, Tencent's Hunyuan, and the phantom-like DeepSeek. In the AI Cloud market, Alibaba Cloud holds the top share and Huawei Cloud is growing rapidly, while Baidu AI Cloud maintains 4th to 5th place with its full-stack AI capabilities. In autonomous driving, Apollo Go competes with Tesla FSD's deployment in China, as well as local players like Didi Autonomous Driving, Pony.ai, and WeRide, with competition shifting from technical demonstrations to operational licenses and unit economics.