Nomura Maintains Buy on Baidu: AI Revenue Exceeds Half for First Time; Kunlunxin IPO Imminent

Overview

This report summarizes key takeaways from Nomura's discussions with Baidu management (including CFO Henry He and IR Director Juan Lin) at the Singapore NIFA conference in early June 2026. The report notes that Q1 2026 was a critical inflection point in Baidu's AI transformation, with AI-driven business revenue accounting for more than half of general business revenue for the first time, signaling a shift in Baidu's growth engine from traditional search advertising to AI. Nomura maintains a 'Buy' rating on Baidu with a target price of USD 190, optimistic about the commercial validation of its AI Cloud Infrastructure, potential value unlocking from Kunlunxin, and disciplined capital allocation strategy.

Core views

AI business has become the primary growth engine, marking a fundamental structural shift. In Q1 2026, Baidu Core AI-driven business revenue exceeded RMB 13.6 billion, up 49% YoY, accounting for 52% of Baidu General Business (BGB) revenue. This is the first time AI business contribution has exceeded half, indicating that Baidu's growth profile is no longer primarily determined by traditional search advertising. Within the AI business, AI Cloud Infrastructure is the largest and fastest-growing component, reaching RMB 8.8 billion in revenue, up 79% YoY; subscription-based GPU cloud revenue grew 184% YoY, highlighting strong demand for AI computing power and inference infrastructure. AI-native marketing services grew 36% YoY to RMB 2.3 billion, while AI application revenue was RMB 2.5 billion, roughly flat YoY. AI Cloud Infrastructure is transitioning from commoditization to scarce computing power. Management believes the Chinese cloud market is undergoing a structural shift from traditional CPU cloud and IT resource procurement toward AI infrastructure, GPU cloud, and workflow-driven AI cloud services. Unlike traditional cloud, which is viewed as a cost center with weak gross margins, AI cloud can reshape workflows and create new commercial value. Growth drivers include: tight supply of advanced AI servers (liquid cooling, dedicated IDCs, etc.) with high barriers to entry on the supply side; demand for embedding AI Agents into real enterprise processes in banking, healthcare, and automotive sectors; improved competitive landscape with higher entry barriers for AI infrastructure compared to traditional cloud, reducing price war risks; and Baidu's neutrality advantage, enabling better service to enterprise clients of internet peers with e-commerce or social businesses. Kunlunxin (KLX) is a core value-unlocking catalyst. Positioned as the foundational layer of Baidu's full-stack AI strategy, Kunlunxin's strategic value lies not only in external sales or an independent IPO but also in reducing Baidu's own AI infrastructure costs, optimizing token economics, and improving AI cloud and application margins. Kunlunxin focuses more on inference rather than competing head-on with previous-generation cutting-edge global chips in pre-training, aligning with trends of large-scale, repetitive, and cost-sensitive inference demands driven by Agents and AI applications. Management stated that the Kunlunxin IPO process is on track, and as a domestic entity rather than a VIE or red-chip structure, the IPO process may be more straightforward. ERNIE Bot accelerates catch-up efforts while balancing financial discipline in investment. Management acknowledged that foundational model progress lagged over the past 1-2 years but noted increased budgets in 2026 for infrastructure, talent, and data training, aiming to return to China's top tier of foundational models. Given mature training knowledge, clear scaling laws, and accelerated monetization via AI Agent proliferation, the catch-up path is more feasible than two or three years ago. The CFO emphasized that AI investment will adhere to ROI discipline and cash flow management, avoiding sensational capex targets for signaling purposes, instead investing in areas with clear revenue paths, margin improvement potential, and long-term strategic control. Regarding shareholder returns, Baidu will continue buybacks and implement its newly introduced dividend policy while exploring mechanisms to allow shareholders to benefit from Kunlunxin value unlocking to reduce holding company discounts.

Analysis framework

Nomura values Baidu using a Sum-of-the-Parts (SOTP) methodology, dividing businesses into traditional advertising core business and emerging AI businesses (cloud, chips, SaaS, Robotaxi, etc.). The analytical logic follows the main thread of 'technology commercialization validation -> competitive moat construction -> capital allocation efficiency': first confirming the AI transformation inflection point via changes in revenue mix; then dissecting supply-demand dynamics and neutral competitive advantages in AI Cloud Infrastructure to demonstrate gross margin improvement potential; next assessing Kunlunxin's value unlocking path as an independent asset and its synergistic effects on overall costs; finally combining management's restrained attitude toward capex and shareholder return policies to judge sustainability of valuation recovery. This analytical approach emphasizes shifting perspective from purely revenue growth to earnings quality and cash flow returns.

Methodology notes

• Valuation MethodologySOTP Valuation

  Sum-of-the-Parts Valuation

  The research report splits Baidu's business into undervalued traditional advertising business (assigned lower PE multiples) and high-growth emerging AI businesses (assigned higher valuations or evaluated separately), valuing each part separately and summing them to derive the target price. This method is commonly used for diversified tech companies with vastly different growth logics across segments to more accurately reflect hidden asset values.

• Industry/Sector Analysis FrameworkSupply-demand framework

  AI Computing Power Supply-Demand Analysis

  By analyzing hard thresholds on the AI server supply side (liquid cooling, dedicated facilities) and workflow embedding characteristics on the demand side, the report demonstrates the scarcity and pricing power of AI Cloud Infrastructure compared to traditional cloud, serving as core logic for judging industry prosperity and corporate competitiveness.

• Competition & Strategy FrameworkMoat / competitive advantage

  Neutrality as Differentiated Advantage

  The report points out that due to lacking large-scale e-commerce or social businesses, Baidu has fewer conflicts with clients, thus possessing a 'neutrality' moat when serving enterprise customers. This is an important perspective for identifying asymmetric advantages in competition among platform-based internet companies.

• Company Fundamentals & Financial FrameworkFree cash flow analysis

  Balancing Capex and Shareholder Returns

  The report focuses on management's balance between heavy AI investment and shareholder returns (buybacks, dividends), emphasizing 'disciplined investment' and ROI orientation, key metrics for assessing financial health and valuation floors during heavy-asset transformation periods.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Baidu (BIDU.US / 09888.HK)

  Direct beneficiary, core vehicle for AI transformation

  Strengths

  Leading commercialization of AI Cloud Infrastructure, autonomous and controllable Kunlunxin with IPO potential, neutrality fostering enterprise client trust, robust cash flow supporting shareholder returns

  Weaknesses

  Foundational model briefly fell behind requiring sustained high investment to catch up; slowing growth in traditional advertising business

  Comparison

  Compared to pure large-model companies, Baidu possesses a closed-loop cloud+chip+application ecosystem; compared to other cloud giants, Baidu has differentiated advantages in AI-specific infrastructure and neutrality

  Risks

  Further deterioration in user growth, intensified competition from long-video platforms, regulatory risks related to online video

Key data

• AI-Driven Business Revenue Share52%Exceeded half of Baidu General Business revenue for the first time in Q1 2026, up 49% YoY
• AI Cloud Infrastructure RevenueRMB 8.8 BillionQ1 2026 data, up 79% YoY, the largest and fastest-growing segment within AI business
• GPU Cloud Subscription Revenue Growth184%YoY growth rate in Q1 2026, reflecting strong demand for AI computing power
• Target PriceUSD 190Based on SOTP valuation, implying FY26 PE of 24x
• Current Share PriceUSD 121.66Closing price as of June 5, 2026
• Traditional Advertising Business ValuationUSD 4.5 BillionCalculated based on FY26F PE of 4x
• New Business ValuationUSD 32.5 BillionIncludes cloud, chips, SaaS, Robotaxi, etc.

Impact & implications

For Baidu, AI revenue exceeding half implies its valuation logic should be reconstructed from a traditional internet advertising company to an AI infrastructure and service provider. If expectations for Kunlunxin's independent IPO materialize, it will significantly eliminate holding company discounts and unlock hidden value. For the industry, Baidu's emphasis on 'neutrality' and 'workflow-driven' services may become a competitive watershed in enterprise-grade AI cloud services, with cloud vendors having fewer scenario conflicts gaining more orders from traditional industries. Investors should watch how Baidu maintains stable shareholder returns while sustaining high capital expenditure, which will be key to supporting medium-to-long-term stock price appreciation.

Risks

• Further deterioration in user growth and time spent
• Intensified competition from long-video platforms in users and content
• Further regulatory risks related to online video platforms or video content
• AI investment returns falling short of expectations leading to excessive capex pressure
• Kunlunxin IPO process hindered or valuation failing to meet expectations

What to watch

• Specific timeline and issuance details for Kunlunxin (KLX) independent IPO
• Ranking changes of Baidu's foundational model (ERNIE Bot) in latest benchmarks
• Gross margin improvement in AI Cloud Infrastructure
• Execution of quarterly buyback amounts and dividend payouts
• Implementation cases and revenue contributions of AI Agents in vertical industries (e.g., banking, healthcare)