China AI Demand and Domestic Substitution Support Cambricon's High Growth; Goldman Sachs Raises Target Price to Rmb1,841

Overview

The report examines Cambricon's revenue, earnings, and valuation outlook amid rising demand for AI chips in China and the domestic substitution trend. Goldman Sachs believes that expanding multi-industry applications, ongoing product specification upgrades, adaptation to domestic large models, the recovery of the R&D team, and the stockpiling of wafer raw materials will support future growth. It therefore substantially raises its medium- to long-term earnings forecasts and maintains its Buy rating.

Core views

The report first bases its growth assessment on demand for AI chips in China and investment in semiconductor self-sufficiency and controllability. Cambricon's customers now include foundation model providers, server and AI software vendors, cloud service providers, and companies in energy, education, finance, telecommunications, healthcare, and the internet. Financial institutions use its chips to optimize business operations and service efficiency, while internet customers use them for foundation models, multimodal applications, search, and recommendations. The company's chips have been adapted to mainstream AI applications and achieved scaled training in the first half of 2026. Vertical scenarios such as smart cities, smart mines, intelligent transportation, and smart retail are also expanding, including semantic search, structured video analysis, visual inspection of goods, power equipment operations and maintenance, workplace safety monitoring, traffic incident detection, foot-traffic identification, intelligent inspection, and theft and loss prevention. Goldman Sachs believes that this broader range of applications expands the company's addressable demand base. Products and R&D constitute the second major growth driver. Cambricon is developing the microarchitecture and instruction set for its next-generation intelligent processors, aiming to optimize the training and inference of natural language processing, video and image generation, and vertical-industry foundation models, while improving programming flexibility, ease of use, performance, power consumption, and chip area. The company continues to enhance computing power, memory capacity, and bandwidth, and its products have completed adaptation to leading domestic foundation models such as GLM, DeepSeek, Qwen, Kimi, and MiniMax. The company was added to the U.S. Entity List in December 2022 and experienced talent attrition in 2023. After the number of R&D engineers fell to a low of 727 in the first half of 2024, it recovered to 1,007 in the first half of 2026. The report believes that the rebound in R&D headcount is conducive to future chip iteration and the development of the software and hardware ecosystem. In terms of near-term operations, revenue grew 8% quarter-on-quarter in the second quarter of 2026, below the 53% growth recorded in the first quarter. The report attributes the slowdown to the timing of customer procurement and wafer supply. Its analysis indicates that historically, after periods of slower inventory growth, the company's quarter-on-quarter revenue growth in the following quarter has generally also been slower: inventory declined 9% quarter-on-quarter in the first quarter of 2026, followed by revenue growth of only 8% quarter-on-quarter in the second quarter. However, inventory surged 83% quarter-on-quarter in the second quarter, with the increase mainly coming from raw materials such as wafers. The report therefore expects quarter-on-quarter revenue growth to strengthen significantly in the third quarter. Raw material inventory grew 779% year-on-year in the first half of 2026, which Goldman Sachs also views as a signal of progress in advanced processes at domestic foundries and improving domestic generative AI supply capabilities. Based on these demand and supply-chain assessments, Goldman Sachs raises its 2026–2030 EBITDA forecasts for Cambricon by 1%, 41%, 69%, 71%, and 82%, respectively, primarily due to higher revenue forecasts. The report expects ongoing chip specification upgrades to drive an increase in the overall average selling price in the future. At the same time, Goldman Sachs lowers its gross margin assumptions because Chinese cloud service provider customers have greater bargaining power as they expand, and large-volume procurement typically involves average selling price discounts. The report also raises its R&D expense forecasts to support hardware including AI chip design, GPU modules, substrates, motherboards, and servers, as well as foundational software such as AI application development platforms and toolkits. The detailed forecasts show revenue increasing from Rmb6,497.2mn in 2025 to Rmb23,980.4mn in 2026, Rmb91,261.6mn in 2027, and Rmb172,469.6mn in 2028, compared with the previous 2026–2028 forecasts of Rmb23,661.8mn, Rmb39,106.5mn, and Rmb62,158.3mn, respectively. EBITDA is expected to increase from Rmb2,185.3mn in 2025 to Rmb6,744.1mn in 2026, Rmb16,004.0mn in 2027, and Rmb31,225.8mn in 2028. The new 2026–2028 earnings-per-share forecasts are Rmb10.63, Rmb22.58, and Rmb43.40, compared with the previous forecasts of Rmb9.78, Rmb15.90, and Rmb25.45. Revenue growth over the same period is forecast at 269.1%, 280.6%, and 89.0%, with EBITDA margins of 28.1%, 17.5%, and 18.1%. The initial decline and subsequent stabilization in margins reflect the trade-off between rapid revenue expansion and customer bargaining power, bulk-purchase discounts, and increased R&D investment. On valuation, Goldman Sachs continues to use a discounted EV/EBITDA methodology, with 2030 as the valuation base year. The target multiple is derived from the relationship between trading EV/EBITDA multiples of Chinese semiconductor peers and their following-year EBITDA growth and EBITDA margins. After incorporating updated peer data and Cambricon's 29% year-on-year EBITDA growth, previously 21%, and 19% EBITDA margin, previously 31%, the target EV/EBITDA multiple is reduced from 43x to 27x. Goldman Sachs applies 27x to estimated 2030 EBITDA and discounts it back to 2027 using an unchanged 12.7% cost of equity, deriving a 12-month target price of Rmb1,841, up from Rmb1,614.77 previously. The new target price implies 82x estimated 2027 P/E, which the report says is broadly consistent with Cambricon's average P/E of 84x since July 2024, supported by estimated average year-on-year EPS growth of 111% in 2027–2028. Based on the report's reference share price of Rmb1,035, the target price implies 77.9% upside, and Goldman Sachs maintains its Buy rating.

Analysis framework

Goldman Sachs first assesses the market opportunity based on China's AI demand and domestic substitution trends, and then examines Cambricon's deployment scenarios across different industries, adaptation to mainstream large models, product iteration capabilities, and changes in R&D headcount. It subsequently uses changes in inventory and raw materials to assess the quarterly revenue trajectory and adjusts its revenue, gross margin, R&D expense, EBITDA, and earnings-per-share forecasts accordingly. Finally, the report determines the target multiple based on the relationship between Chinese semiconductor peers' EV/EBITDA multiples, growth, and margins, uses 2030 EBITDA as the base, and discounts it back to 2027 to derive the 12-month target price.

Methodology notes

• Valuation MethodEV/EBITDA valuation

  Discounted EV/EBITDA Valuation

  The report applies a target EV/EBITDA multiple of 27x to estimated 2030 EBITDA and then discounts it back to 2027 at a 12.7% cost of equity, deriving a 12-month target price of Rmb1,841.

• Industry/Sector Analysis FrameworkSupply-demand framework

  Combined Analysis of AI Demand, Domestic Substitution, and Wafer Supply

  The report assesses demand for AI chips and investment in localization in China while also examining wafer supply and raw material inventory to determine Cambricon's revenue growth potential and near-term delivery cadence.

• Industry/Sector Analysis FrameworkVolume-price decomposition

  Chip Shipments and Overall Average Selling Price Analysis

  The report decomposes growth into demand and shipment growth driven by application expansion and an increase in the overall average selling price driven by specification upgrades, while considering the impact of cloud service providers' bargaining power and bulk-purchase discounts on gross margin.

• Cycle and Business Conditions Framework

  Inventory as a Leading Indicator of Next-Quarter Revenue

  Based on the company's historical relationship between inventory and subsequent revenue, the report treats the 83% quarter-on-quarter increase in inventory in the second quarter of 2026, mainly contributed by wafer raw materials, as a leading signal of stronger quarter-on-quarter revenue growth in the third quarter.

Asset mapping & comparison

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

• Cambricon (688256.SS)

  A direct beneficiary of rising demand for AI chips in China, domestic substitution, expanding multi-industry applications, and product specification upgrades.

  Strengths

  Broad industry coverage among customers, with products adapted to several leading domestic foundation models; the number of R&D engineers has recovered, and the R&D scope spanning chips, hardware, and foundational software is comprehensive.

  Weaknesses

  Cloud service provider customers have relatively strong bargaining power, and large-volume procurement may lead to average selling price discounts; R&D investment is increasing, while medium-term margins face pressure.

  Comparison

  The target EV/EBITDA is determined based on the relationship between Chinese semiconductor peers' trading multiples, following-year EBITDA growth, and margins. The target price implies 82x 2027 P/E, close to the company's average of 84x since July 2024.

  Risks

  Constrained wafer supply, slower-than-expected cloud chip development, and stronger-than-expected competition in cloud chips.

Key data

• 12-Month Target PriceRmb1,841.00Raised from Rmb1,614.77
• Reference Share PriceRmb1,035.00The target price implies 77.9% upside
• 2026–2030 EBITDA Forecast Revisions+1% / +41% / +69% / +71% / +82%Primarily driven by higher revenue forecasts
• 2026–2028 Revenue ForecastsRmb23,980.4mn / Rmb91,261.6mn / Rmb172,469.6mnPrevious forecasts were Rmb23,661.8mn, Rmb39,106.5mn, and Rmb62,158.3mn, respectively
• 2026–2028 EBITDA ForecastsRmb6,744.1mn / Rmb16,004.0mn / Rmb31,225.8mnCorresponding to EBITDA margins of 28.1%, 17.5%, and 18.1%
• 2026–2028 Earnings-per-Share ForecastsRmb10.63 / Rmb22.58 / Rmb43.40Previous forecasts were Rmb9.78, Rmb15.90, and Rmb25.45
• Second-Quarter 2026 Quarter-on-Quarter Revenue Growth8%Below 53% in the first quarter of 2026
• Second-Quarter 2026 Quarter-on-Quarter Inventory Change+83%Mainly contributed by raw materials such as wafers, based on which the report expects stronger quarter-on-quarter revenue growth in the third quarter
• First-Half 2026 Year-on-Year Change in Raw Material Inventory+779%The report believes this also reflects the development of advanced processes at domestic foundries
• Number of R&D Engineers1,007Recovered by the first half of 2026 from a low of 727 in the first half of 2024
• Target EV/EBITDA27xPreviously 43x; applied to estimated 2030 EBITDA
• Cost of Equity12.7%Maintained unchanged and used to discount the 2030 valuation back to 2027
• Target Price-Implied 2027 P/E82xBroadly consistent with the average P/E of 84x since July 2024
• Market Capitalization and Enterprise ValueRmb601.9bn / Rmb599.1bnMarket capitalization and enterprise value presented in the report

Impact & implications

The report believes that Cambricon's growth drivers are expanding from AI demand alone to a combination of multi-industry deployment, product specification upgrades, adaptation to domestic models, recovery in R&D capabilities, and development of the domestic supply chain. Earnings forecasts are therefore raised substantially, although the bargaining power of cloud service providers, bulk-purchase discounts, and higher R&D investment may weigh on margins. Even with the target valuation multiple reduced from 43x to 27x, higher medium- to long-term EBITDA forecasts still drive the target price increase to Rmb1,841.

Risks

• Wafer supply may be constrained; the company was added to the U.S. Entity List in December 2022.
• Cloud chip development may progress more slowly than expected.
• Competition in the cloud chip market may be more intense than expected.

What to watch

• Monitor whether quarter-on-quarter revenue growth strengthens significantly in the third quarter as expected by the report following the 83% quarter-on-quarter increase in inventory in the second quarter of 2026.
• Monitor progress in next-generation intelligent processor development, adaptation to leading domestic foundation models, and chip specification upgrades.
• Monitor whether the recovery in R&D engineer headcount translates into product iteration and stronger software and hardware ecosystem capabilities.
• Monitor changes in wafer supply and raw material inventory, as well as progress in the development of domestic advanced processes.
• Monitor the impact of cloud service providers' bargaining power, bulk-purchase discounts, and increased R&D investment on gross margin and EBITDA margin.