JPMorgan Upgrades Baidu to Overweight on AI Cloud Growth

Baidu’s AI Cloud revenue is projected to surge 61 percent in 2026 as proprietary Kunlun chips drive enterprise adoption despite advertising declines.

Key Takeaways

Key Takeaways

• JPMorgan upgraded Baidu to Overweight from Neutral with a $188 price target (up from $110), citing AI and cloud computing as primary growth drivers.
• Baidu’s cloud revenue is projected to surge 61% in 2026, following an estimated 23% growth in 2025, powered by proprietary Kunlun AI chips and strong domestic demand. Non-online marketing revenue (including AI Cloud) reached over RMB 10 billion quarterly for the first time in Q2 2025, with AI Cloud specifically hitting RMB 6.5 billion, up 27% year-over-year.
• Core advertising revenue declined 15% year-over-year in Q2 2025, with non-advertising revenue now comprising approximately 38-40% of Baidu Core operations. In Q3 2025, advertising fell further by 18%, highlighting ongoing monetization challenges in AI-driven search.

Introduction

The market has long struggled to value Baidu, caught between the company’s eroding search advertising franchise and its ambitious repositioning as China’s premier AI infrastructure provider. JPMorgan’s recent upgrade to Overweight, accompanied by a 71% price target increase to $188, suggests that calculus may be shifting. The firm now values Baidu’s cloud business at $34 billion, more than half the company’s enterprise value, a remarkable reappraisal for a division that generated just RMB 6.5 billion in quarterly revenue as recently as Q2 2025.

This reassessment arrives at an inflection point. Baidu’s core advertising business continues its precipitous decline, falling 18% year-over-year in Q3 2025 and marking the company’s largest quarterly revenue contraction on record. Yet beneath this surface deterioration, the company is executing a high-stakes transformation anchored in proprietary silicon and enterprise cloud services. The question for investors is whether Baidu can scale its AI infrastructure business quickly enough to offset the structural decay in search monetization.

The Silicon Advantage

At the heart of Baidu’s strategy sits the Kunlun chip family, a bet on vertical integration that has gained urgency amid sustained U.S. export restrictions on advanced semiconductors. The company now operates a 30,000-chip cluster for training large language models, with updated M100 and M300 variants scheduled for commercial launch through 2027. Early traction among Chinese hyperscalers, telecommunications operators, and state-owned enterprises suggests genuine product-market fit in a domestic market increasingly wary of foreign dependencies.

Macquarie estimates Kunlun will generate RMB 5 billion in revenue during 2025, with projections calling for that figure to double in 2026. These numbers, while modest in absolute terms, represent a meaningful foothold in China’s broader push toward semiconductor self-sufficiency. More importantly, they create differentiation in an increasingly commoditized cloud infrastructure market. Alibaba may dominate with 35.8% of China’s AI cloud services, but Baidu’s integrated approach to chips, models, and deployment platforms offers a distinct value proposition to enterprises seeking alternatives to foreign technology stacks.

The Kunlun advantage extends beyond hardware economics. By controlling the full stack from silicon to inference, Baidu can optimize performance characteristics that matter most to Chinese enterprises: latency, throughput, and data sovereignty. This vertical integration mirrors strategies deployed successfully by hyperscalers globally, though Baidu’s execution occurs within the unique constraints and opportunities of China’s technology ecosystem.

Revenue Mix in Transition

The financial evidence of transformation is unmistakable. Non-advertising revenue now comprises roughly 38% to 40% of Baidu Core operations, up from negligible levels just years ago. AI Cloud specifically grew 27% year-over-year in Q2 2025, moderating slightly to 21% in Q3 but maintaining momentum as advertising revenue collapsed. Broader AI-powered businesses, including cloud infrastructure and model-as-a-service offerings, surged more than 50% year-over-year in Q3, reaching approximately RMB 10 billion.

These growth rates, while impressive, occur from a relatively small base. Baidu holds just 6.1% of China’s AI cloud market in H1 2025, trailing not only Alibaba but also ByteDance, Huawei, and Tencent. The company ranks fifth in a rapidly expanding market projected to reach 51.8 billion yuan for the full year, more than doubling from 20.83 billion yuan in 2024. JPMorgan’s projection of 61% cloud revenue growth in 2026 assumes Baidu can capture meaningful share as enterprises accelerate AI adoption.

The advertising decline, meanwhile, shows no signs of abating. Core advertising revenue fell to RMB 15.3 billion in Q3 2025, an 18% year-over-year contraction that reflects both macroeconomic headwinds and structural challenges in monetizing AI-powered search. AI search now accounts for 64% of ad impressions, but conversion rates remain problematic. Baidu has yet to demonstrate that generative AI search can generate comparable revenue per query to traditional link-based advertising, a challenge confronting every major search provider globally.

Strategic Positioning Amid Geopolitical Fracture

Baidu’s transformation unfolds against China’s ambition to build a $150 billion AI industry by 2030, equivalent to roughly 1 trillion yuan. This national priority creates both opportunity and obligation. The company’s two-tier governance approach, filing base models for approval while avoiding app-level restrictions, enables rapid iteration. Recent model releases including ERNIE 4.5 and ERNIE 5.0 demonstrate technical competitiveness, while the Qianfan platform provides enterprises with rapid deployment capabilities.

China’s cloud infrastructure market operates as a closed ecosystem, largely impenetrable to AWS, Microsoft Azure, and Google Cloud due to regulatory requirements and data sovereignty concerns. This protection has enabled local champions to flourish, with Alibaba, Huawei, Tencent, and Baidu collectively controlling the domestic landscape. For Baidu, sixth in overall cloud market share at 9%, the question is whether specialized AI capabilities can drive disproportionate growth as enterprises move beyond general-purpose infrastructure.

The company’s Apollo Go autonomous driving division adds another dimension, delivering over 2.2 million driverless rides in Q2 2025, up 148% year-over-year. Weekly fully driverless rides exceeded 250,000 by October, with profitability projected before year-end. While near-term financial contribution remains limited, Apollo Go reinforces Baidu’s positioning as an integrated AI company rather than a cloud infrastructure provider alone.

The Margin Pressure Reality

Transformation carries costs. Baidu Core’s adjusted EBITDA margin compressed to approximately 24% in Q2 2025, down from healthier levels as R&D spending and infrastructure investment increased. Management acknowledges near-term margin pressure from advertising weakness and development expenses, with efficiency measures aimed at stabilization rather than expansion. This stands in contrast to mature cloud providers like AWS, which operates at meaningfully higher margins despite ongoing AI infrastructure investment.

The legacy iQIYI streaming business continues its slow fade, with revenue declining 11% in Q2 and 8% in Q3 2025. This further constrains consolidated financial performance even as AI Cloud accelerates. Total revenue fell 7% year-over-year in Q3, the steepest quarterly decline on record, underscoring the turbulence inherent in Baidu’s pivot.

Investment Framework

JPMorgan’s upgrade reflects confidence that Baidu’s AI infrastructure business justifies a premium valuation despite near-term financial pressure. The $34 billion cloud valuation implies aggressive assumptions about revenue trajectory, margin expansion, and market share gains in a competitive landscape. Yet Baidu’s stock, trading at $114.11 as of November 2025, has appreciated roughly 28% year to date, suggesting investors are beginning to assign value to the company’s transformation narrative.

The risk-reward framework is asymmetric. On the downside, advertising continues to deteriorate faster than cloud revenue scales, margin compression persists, and Baidu fails to differentiate meaningfully in AI infrastructure. On the upside, Kunlun achieves widespread adoption, enterprise cloud demand accelerates as projected, and margin efficiencies emerge at scale. The 71% price target increase embedded in JPMorgan’s upgrade clearly assumes the latter scenario materializes.

For institutional investors, Baidu represents exposure to China’s AI ambitions through a company with genuine technical capabilities and strategic alignment with national priorities. Whether that exposure justifies current valuations depends on conviction that cloud revenue can grow 61% in 2026 while advertising stabilizes and margins recover. The evidence remains incomplete, but the thesis is increasingly coherent.