Baidu 2025: reigniting growth through AI and the cloud
- Administrateur
- 3 days ago
- 3 min read
Founded in 2000 by Robin Li, Baidu long served as China’s main gateway to the internet. Its search engine—often dubbed “China’s Google”—underpinned a highly profitable advertising empire. Yet since the late 2010s the company’s expansion has lagged that of its Chinese peers. Fierce competition, missed technology shifts and tighter regulation now oblige Baidu to reinvent itself if it hopes to regain momentum.
Activities and ecosystem
Search still sits at the business core: ads linked to search queries and the personalised news feed in the Baidu App account for more than half of revenue. Around this hub Baidu has built a wide service portfolio—Baidu Maps, Baidu Baike (encyclopaedia), Baidu Tieba (forums), Wenku (documents) plus translation, cloud-storage and browser utilities.The group has also poured resources into artificial intelligence: the Xiaodu voice assistant, Kunlun AI chips and Apollo, its autonomous-driving platform. In entertainment it controls iQIYI, one of China’s three dominant video-streaming services.
Geographic footprint and key segments
Baidu still generates the overwhelming majority of income at home, holding more than half of Chinese web searches. Internationally it runs only a handful of R&D centres and pilot robotaxi projects. Strategically the firm now rests on four pillars:
Online advertising – the historic core, recently in decline.
Cloud and AI services – the fastest-growing segment, propelled by Chinese demand for sovereign tech stacks.
Autonomous mobility – via Apollo and the robotaxi service Apollo Go, already commercial in several major cities.
Entertainment – iQIYI, still seeking the right balance between premium content and profitability.
Why growth slowed
Macro headwinds: weak consumption and property woes since 2022 have trimmed SME ad budgets, Baidu’s primary clientele.
New mobile giants: WeChat, Douyin and Toutiao now dominate user attention and ad spend, relegating search to more occasional use.
Late mobile transition: Baidu never built a super-app; its audience eroded as users shifted to more engaging, closed ecosystems.
Scandals and regulation: the 2016 Wei Zexi incident tarnished trust, while Beijing’s post-2021 rules on data and algorithms limit monetisation.
Heavy investment: over 20 % of annual revenue goes to R&D for AI, chips and robotaxis—bets that have yet to generate equivalent cash flow.
Strategic missteps and missed openings
Baidu underestimated social media, leaving Sina Weibo, WeChat and Xiaohongshu to dominate conversation. Its e-commerce forays (Youa, Nuomi) faltered, ceding a massive market to Alibaba, JD and Pinduoduo. In short-form video the group failed to counter Douyin or Kuaishou. Aggressive monetisation of barely differentiated sponsored links eroded user confidence and drew regulator scrutiny.
Assets for a rebound
Search moat: with Google absent, Baidu remains the default portal in Mandarin, yielding a unique trove of behavioural data.
AI excellence: its large language model Ernie now rivals top Western systems; the open-source PaddlePaddle framework anchors a vast developer ecosystem.
Cloud momentum: Baidu AI Cloud ranks among China’s top three providers and benefits from state-backed “tech sovereignty” programmes.
Lead in autonomy: Apollo has logged tens of millions of test kilometres; Baidu holds the data, permits and tech to scale Apollo Go rapidly.
Implicit state backing: Beijing views Baidu as an AI “national champion”, easing access to public contracts and essential licences.
iQIYI: turnaround or trouble?
After years of losses, the streamer broke even in 2023 by raising prices and trimming content costs. Subscriber growth, however, slipped in 2024 as Tencent Video and Youku battled on price. iQIYI is now testing shorter formats and modest expansion in Southeast Asia. Its fate hinges on producing hits without blowing the budget. For Baidu the asset is strategic—yet potentially for sale should management refocus capital on AI and cloud.
Outlook 2025-2030
Moderate ad rebound: growth could resume by late-2025 if China’s economy steadies, though pure search will stay a mature market.
Monetising generative AI: Baidu can charge for Ernie APIs, premium search answers or AI productivity suites.
Cloud as main engine: enterprise cloud + AI services may overtake search in profit contribution.
Robotaxi scaling: Apollo Go could reach 20 Chinese cities by 2027; overseas roll-outs depend on local rules.
Balancing returns and R&D: investors want proof that 20 % R&D spending converts to cash. Asset sales—iQIYI or niche fintech/education units—could fund cloud or autonomy growth.
Conclusion
Twenty-five years on, Baidu is still a keystone of China’s internet, yet its relative weight has shrunk versus Tencent, Alibaba and ByteDance. The firm nonetheless owns two unique levers: top-tier AI expertise and a lion’s share of Mandarin search traffic. If it can turn Ernie, its cloud arm and Apollo into recurring revenue, while revitalising its ad offering, Baidu could reclaim an upward trajectory and a richer market valuation. If not, it risks remaining a second-tier player—strong in search, burdened by costly tech bets. Execution over the next three to five years will decide whether Baidu secures a genuine second wind or lingers in the shadow of newer Chinese digital champions.
Commentaires