J.P. Morgan: Use consumer sector funds to focus on AI, robotics, and core Tier-1 city real estate

Overview

This report, based on client discussions and panel feedback from J.P. Morgan’s Global China Summit, reaffirms its thematic allocation views on Chinese equities. The core thesis is to use broad consumer sector exposure as a funding source to strategically position in artificial intelligence (AI), robotics, and developers in central business districts (CBDs) of Tier-1 cities. The report notes that although global investors currently gain AI exposure mainly through U.S., Korean, Taiwanese, and Japanese names, capital could flow back into Chinese AI tech stocks as domestic large models improve profitability. Meanwhile, spillover effects from AI capex are broadening tailwinds for Chinese industrial orders, while the property market shows pronounced structural divergence.

Core views

AI and Capital Expenditure: China’s annual AI capex is roughly 20% of the U.S. level, with the gap primarily due to supply-side constraints (e.g., limited access to advanced chips) and immature demand-side fundamentals. However, domestic substitution is progressing—for example, DeepSeek V4-Pro leads in coding capabilities and costs only one-fifth of competitors’ pricing. The firm expects that once a commercialization inflection point is reached, valuations of China’s hyperscale cloud providers will be increasingly driven by AI. Top picks include Baidu, Zhipu AI, TianShu ZhiXin, NAURA, and Weichai Power. Industrial Automation and Spillover Effects: Global AI capex is generating broad “second-order” spillovers—not only benefiting semiconductors but also boosting orders in thermal management, high-precision machine tools, injection molding, and power infrastructure. Data shows that in Q1 2026, non-financial, non-property A-share capex grew 4% YoY, while inventories rose 8%, signaling early-stage signs of a manufacturing capex cycle. This provides structural support for the factory automation segment. Property Market Divergence: Shanghai’s housing market exhibits a classic K-shaped recovery. Homes priced below RMB 3 million are selling quickly, with an inventory absorption period of only about 5 months, whereas mid-to-high-end units priced between RMB 12–15 million face slow sales and a 36-month absorption period. In the luxury segment, traditional high-end residences face over 42 months of inventory pressure, but newly launched “Gen 3.5” luxury homes are selling strongly. The firm believes equity wealth effects may support high-net-worth demand for premium properties, though a broad-based sector recovery remains unlikely. Top picks include China Overseas Land & Investment, China Resources Land, China Jinmao, and China Resources Mixc Living. Consumption and Market Sentiment: Addressing concerns about AI sector crowding, regulatory cooling, and surging U.S. Treasury yields, the firm views recent volatility as largely short-term. Regulators show no intention to dampen the bull market, and China Securities Finance Corporation (CSFC) ETF redemptions actually help raise cash buffers against downside risks. Moreover, the macro team expects the first rate hike only in Q3 2027. On the consumption front, lower-tier cities and rural areas are the main engines of incremental spending, with sustained demand for health-oriented products (e.g., sugar-free beverages, premium tea).

Analysis framework

The firm employed a top-down strategy combined with bottom-up survey feedback. First, it gathered insights from institutional investors and industry experts at the Global China Summit to identify market consensus and divergences (e.g., geographic preferences for AI exposure, regulatory concerns). Second, it applied a supply-demand framework to analyze the AI industry, comparing China-U.S. capex gaps and assessing technical progress and commercialization inflection points in domestic substitution. Third, it used supply chain transmission logic to examine how AI capex spills over into industrial automation and related manufacturing sectors, validating the cycle phase with macro data (capex, inventory growth). Finally, in real estate and consumption, it adopted a granular, structural approach—eschewing aggregate judgments and instead drilling into price bands and product types (e.g., absorption cycles for different-priced properties in Shanghai)—to capture structural opportunities.

Methodology notes

• Industry/Sector Analysis FrameworkSupply-demand framework

  Assessing the AI industry’s development stage and investment timing by analyzing supply-side constraints (chip limitations) and demand-side maturity (commercialization inflection point).

  Rather than focusing solely on overall AI hype, the report dissects why China lags the U.S. in AI investment (supply constraints, immature demand) and identifies when improvements in these factors will trigger valuation resets.

• Industry/Sector Analysis FrameworkUpstream-Midstream-Downstream Supply Chain Transmission

  Second-order spillovers—AI investment benefits not only direct beneficiaries (chips) but also drives demand for upstream equipment and related materials.

  Helps investors understand that AI exposure should extend beyond software or chips to include industrial companies supplying 'picks and shovels' like cooling systems and precision machine tools for AI data centers.

• Cycle and Momentum FrameworkInventory cycle (Kitchin)

  Using year-over-year changes in capital expenditure (Capex) and inventory to determine whether manufacturing is in a restocking or capacity expansion phase.

  The report cites Q1 data showing simultaneous growth in capex and inventory (excluding finance and real estate), suggesting manufacturing may be entering a new upswing in the business cycle.

• Event Arbitrage and Behavioral FinanceExpectation Gap / Expectation Management

  Analyzing discrepancies between market fears (regulation, IPOs, U.S. yields) and actual policy intentions (CSFC support, delayed hikes) to identify oversold opportunities.

  Highlights that markets are overreacting to short-term headwinds while overlooking supportive liquidity policies, thus recommending buying on pullbacks.

Asset mapping & comparison

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

• Baidu (9888.HK/BIDU)

  Core AI large model and cloud services play; benefits from domestic substitution and commercialization inflection

  Strengths

  Leading large model tech stack and cloud infrastructure

• Zhipu AI (2513.HK)

  Pure-play large model company; direct beneficiary of improved AI profitability

• TianShu ZhiXin-H (9903.HK)

  AI chip infrastructure; benefits from domestic compute demand

• NAURA (002371.SZ)

  Semiconductor equipment leader; benefits from AI chip manufacturing expansion

• Weichai Power-H (2338.HK)

  Industrial power and automation exposure; benefits from recovering manufacturing capex

• China Overseas Land & Investment (0688.HK)

  Tier-1 city core-location developer; seen as a lagging catch-up play

  Strengths

  Strong balance sheet, focused on core cities

  Comparison

  More defensive than other property stocks

• China Resources Land (1109.HK)

  Premium commercial property operator; benefits from core asset value preservation

• China Jinmao (0817.HK)

  High-end residential developer; benefits from quality-focused improvement demand

• China Resources Mixc Living (1209.HK)

  Property management and commercial operations; light-asset model favored

Key data

• China’s AI Capex as % of U.S.Approx. 20%Significant gap relative to U.S. scale
• DeepSeek V4-Pro Price AdvantageApprox. 1/5 of competitorsCompared to blended pricing of GLM-5.1/Kimi K2.6
• Q1 2026 A-Share Capex Growth4%YoY growth excluding financials and real estate
• Shanghai <RMB 3M Property Inventory Absorption PeriodApprox. 5 monthsFast absorption, strong demand
• Shanghai RMB 12–15M Property Inventory Absorption PeriodApprox. 36 monthsSlow absorption, weak demand
• Fed First Rate Hike Expected TimingQ3 2027Per J.P. Morgan Economics Research Team forecast

Impact & implications

For investors, this implies a need to adjust traditional sector allocation logic. Rather than relying solely on broad macro consumption data, profits should be sought through structural opportunities. In tech, the focus should shift from pure hardware import substitution to domestically developed large models with commercialization capabilities and their supply chains. In industrials, AI-driven automation upgrades represent a long-term trend, with related order growth gradually reflected in earnings. In real estate, extreme selectivity is required—only developers with prime Tier-1 city locations and strong product offerings can withstand sector-wide headwinds. Strategically, stable cash flows or valuation recoveries in consumer stocks should fund exposure to high-growth AI and robotics segments.

Risks

• Unexpected domestic regulatory tightening to cool the market
• Massive IPOs by CXMT and YMTC draining market liquidity
• Unexpected surge in U.S. Treasury yields
• Domestic macro consumption data underperforming expectations
• Persistent household balance sheet stress in real estate hindering broad recovery

What to watch

• Arrival of commercial profitability inflection point for Chinese large models
• Sustainability of A-share manufacturing capex and inventory data
• Changes in inventory absorption periods across price segments in Tier-1 cities like Shanghai
• Consumer spending trends in lower-tier markets and health-oriented categories
• Interest rate policy paths of the Fed and other major global central banks