Morgan Stanley Maintains Cambricon Overweight and Raises Target Price to RMB 2,000
AI summary card
Morgan Stanley Maintains Cambricon Overweight and Raises Target Price to RMB 2,000
The report finds that Cambricon’s strong Q1 2026 performance, improved SMIC-related supply chain, and clearer product roadmap for MLU580/MLU690 will drive the realization of China’s AI chip localization dividend.
- Q1 2026 performance drove the stock price to hit the daily limit on April 30, 2026, up 20%, significantly outperforming the CSI 300 index’s 0.8% gain.
- MLU580 is expected to become the main shipment product in H2 2026; the SMIC-manufactured version has completed tape-out and entered production, with mass shipments expected starting from Q3 2026.
- China’s AI computing chip market TAM is projected to reach US$67 billion by 2030, with a compound growth rate of 23% starting from 2024; the domestic AI chip localization rate is expected to rise from 41% in 2025 to 86% by 2030.
- Morgan Stanley raised its 2026 revenue forecast by 3%, and its 2027 and 2028 forecasts by 16% each; corresponding EPS forecasts were raised by 8%, 15%, and 11% respectively.
- The target price was raised from RMB 1,588 to RMB 2,000; bull/bear scenario valuations were adjusted from RMB 3,000/RMB 800 to RMB 3,778/RMB 1,008 respectively.
Report interpretation
Overview
This report is Morgan Stanley’s company research and earnings review on Cambricon Technology Corporation (688256.SS). The core view is that China’s AI infrastructure development, growing inference demand, and export controls are jointly driving the domestic AI accelerator market into a phase of increasing market share; as the second-largest domestic AI chip supplier after Huawei’s Ascend, Cambricon benefits from strong demand, co-building customer software ecosystems, and improved supply chain visibility.
Core views
The report maintains an Overweight rating and raises the target price, mainly based on: Q1 2026 results validating the resilience of domestic AI demand; customer prepayments increased by 155% quarter-on-quarter to RMB 1.9 billion, indicating stronger order visibility; MLU580, MLU590, and MLU690 forming a clearer product roadmap; SMIC supply chain normalization faster than expected; and the rising domestic AI chip localization rate creating a large market window for leading manufacturers.
Analysis framework
The analysis framework combines top-down market space for China’s AI accelerators, localization rate trends, and CSP capital expenditure needs, as well as bottom-up analysis of Cambricon’s product roadmap, supply chain progress, customer orders, earnings forecasts, and residual income valuation models. The report also assesses stock price potential under different demand, market share, gross margin, and valuation multiple assumptions through bull, base, and bear-case scenarios.
Methodology notes
Deriving target price from shareholders’ equity, net profit, ROAE, and cost of equity
The target price of RMB 2,000 comes from the residual income model, reflecting upward revisions in 2026-2028 EPS, stronger year-over-year earnings elasticity, improved supply visibility, and more sustained sales ramp-up.
Research model forecasting framework
Unless otherwise specified, financial metrics in the report are based on the Morgan Stanley ModelWare framework and use Morgan Stanley Research estimates.
China’s AI chip demand and domestic substitution path
The report projects that China’s AI computing chip market will reach US$67 billion by 2030, with the domestic AI chip localization rate rising from 41% in 2025 to 86% by 2030.
Bull, Base, and Bear Case Scenarios
The bull case assumes a revenue CAGR over 130% from 2025 to 2028, continued market share gains, and gross margins above 55%; the base case assumes a revenue CAGR of 99%; the bear case assumes a revenue CAGR below 60%, market share loss, and gross margins below 40%.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- Cambricon Technology Corporation (688256.SS)Core Coverage Asset
- Strengths
- Benefits from China’s AI chip localization, CSP demand, co-building software ecosystems, and product roadmap advancement; the report considers it the second-largest domestic AI chip supplier after Huawei’s Ascend.
- Weaknesses
- Relatively high absolute valuation levels; the shift to domestic manufacturing still faces yield challenges; customer concentration and price competition may affect profitability quality.
- Comparison
- The report notes that although domestic manufacturers lag behind global leaders in process node technology, they can narrow the actual deployment gap through packaging, system architecture, and software-hardware optimization; in the Chinese market, customers place greater emphasis on token costs and deployability.
- Risks
- Slower-than-expected demand, market share loss, pricing pressure, rising wafer costs, underutilized capacity, and slower-than-expected technological iterations.
- SMICKey Supply Chain Partner
- Strengths
- The SMIC-manufactured MLU580 has completed tape-out and entered production, and the supply chain normalization is faster than expected.
- Weaknesses
- Yield remains a challenge and needs gradual improvement in 2026.
- Comparison
- Replacing overseas production with domestic supply chains helps improve supply visibility.
- Risks
- If capacity, yield, or manufacturing pace fall short of expectations, it could impact Cambricon’s shipments and revenue realization.
- NVIDIA processors for ChinaCompetition and Supply-Demand Reference
- Strengths
- Global AI chip ecosystem and performance leadership.
- Weaknesses
- China’s market supply is tightening and affected by export controls.
- Comparison
- The report believes that domestic chips offer lower TCO and comparable per-token costs in China’s AI LLM inference scenarios.
- Risks
- If overseas supply eases or price competition intensifies, the pace of domestic substitution could face pressure.
Key data
- RatingOverweightMorgan Stanley maintains a positive relative rating for 688256.SS.
- Target PriceRMB 2,000.00Raised from the previous RMB 1,588.00.
- Current PriceRMB 1,699.96Closing price on April 30, 2026.
- Implied Upside18%Based on the target price and the closing price on April 30, 2026.
- 2026 Revenue ForecastRMB 21,669 millionUpgraded by 3% from the previous forecast of RMB 20,944 million.
- 2027 Revenue ForecastRMB 38,409 millionUpgraded by 16% from the previous forecast of RMB 33,186 million.
- 2028 Revenue ForecastRMB 51,176 millionUpgraded by 16% from the previous forecast of RMB 44,256 million.
- 2026 EPS ForecastRMB 15.63Upgraded by 8% from the previous forecast of RMB 14.49.
- 2027 EPS ForecastRMB 25.92Upgraded by 15% from the previous forecast of RMB 22.50.
- 2028 EPS ForecastRMB 35.43Upgraded by 11% from the previous forecast of RMB 31.88.
- China AI Chip TAMUS$67 billion by 2030The report projects a CAGR of 23% starting from 2024.
- Domestic AI Chip Localization Rate86% by 2030The report projects a rise from 41% in 2025 to 86% by 2030.
- Valuation MultiplesApproximately 109x 2026e P/E, approximately 37x 2026e P/SThe report attributes the high multiples to strong growth, improved supply, and the company’s position as a domestic AI chip leader.
Impact & implications
The report’s investment implications for Cambricon are generally positive: if China’s AI inference demand, CSP capital spending, domestic substitution, and SMIC supply improvements continue to materialize, the company’s revenue and profit elasticity could keep unfolding, supporting its high valuation. However, the current valuation is already relatively high, and future stock performance will depend more on the pace of MLU580 mass production, the launch of MLU690, continuity of customer orders, and gross margin resilience.
Risks
- Domestic AI capital spending or LLM commercialization progress slower than expected.
- Cambricon’s market share in China’s AI chips declines, or it fails to sustain large-scale orders from top-tier customers.
- Rising wafer manufacturing costs, intense price competition, or underutilized capacity leading to lower-than-expected gross margins.
- SMIC-related capacity and yield improvements falling short of expectations, impacting MLU580 volume ramp-up.
- Slower-than-expected product iterations like MLU690, weakening technological leadership and valuation support.
- High customer concentration could amplify order volatility risk.
- Currently high P/E and P/S valuations—should growth fall short of expectations, there’s significant risk of valuation correction.
What to watch
- Actual progress of MLU580 mass shipments starting from Q3 2026.
- Whether MLU690 can start shipping in Q4 2026, and whether its performance improvement meets the report’s expectations.
- SMIC supply chain normalization, capacity, and yield improvement status.
- CSP customer AI capital spending, inference demand, and changes in order prepayments.
- Speed of domestic AI chip localization rate increase, and ongoing impact of export controls on domestic demand.
- GPU rental prices, mainstream AI LLM token prices, and signs of price competition.
- New product launches and customer ecosystem progress at industry events like WAIC 2026.