
If you want to find Hong Kong-listed companies related to memory chip design, the key is not simply to look for the “semiconductor” label, but to examine whether their products are directly connected to Flash, DRAM, non-volatile memory, memory interfaces, MCUs or FPGA supply chains. More precisely, GigaDevice 3986.HK is closer to the main memory chip design track, Shanghai Fudan 01385.HK is a composite company covering non-volatile memory, FPGA and security chips, while Montage Technology 6809.HK focuses more on AI server memory interconnect chips. You need to analyze them separately across three dimensions: memory chips, embedded control and memory interconnect.

Hong Kong-listed companies related to memory chip design can be divided into three categories: the first category consists of companies directly engaged in memory chip design; the second includes companies involved in MCUs, FPGAs, memory interfaces and other chips strongly related to memory capabilities; the third includes indirectly related companies such as wafer foundries, packaging and testing companies, and semiconductor equipment suppliers. If you screen only by the “chip stock” label, you may easily mix up design companies, manufacturing companies and equipment companies, which can blur the investment logic.
The core of memory chip design is whether the chip product itself performs the function of storing data, reading data or supporting system operation. Typical products include NOR Flash, NAND Flash, EEPROM, DRAM, SLC NAND and embedded NVM. By contrast, companies such as SMIC and Hua Hong Semiconductor are more focused on wafer manufacturing, while ASMPT is more related to equipment and packaging. They are affected by the semiconductor cycle, but they are not “memory chip design companies.”
| Category | Representative Direction | Hong Kong-Listed Examples | Relevance to Memory Chip Design |
|---|---|---|---|
| Direct memory design | NOR Flash, SLC NAND, niche DRAM | GigaDevice 3986.HK | High |
| Non-volatile memory + FPGA | EEPROM, NOR Flash, SLC NAND, FPGA | Shanghai Fudan 01385.HK | Medium to high |
| Memory interconnect | DDR5, CXL, memory interface chips | Montage Technology 6809.HK | Medium to high |
| Wafer foundry | Mature process, specialty process | SMIC, Hua Hong Semiconductor | Indirect |
| Equipment and packaging | Packaging equipment, testing equipment | ASMPT and others | Indirect |
Why should not all Hong Kong-listed semiconductor companies be called memory chip stocks? The reason is that the memory supply chain is highly specialized. Design companies determine product architecture, interfaces, capacity, power consumption and customer applications; wafer foundries handle manufacturing; packaging and testing companies handle assembly and testing; equipment companies provide production tools. They may all be affected by semiconductor cycles, but their earnings drivers are different.
You can judge relevance through four dimensions. First, whether the product is a memory chip. Second, whether memory-related businesses account for a meaningful share of revenue. Third, whether performance is affected by memory pricing cycles. Fourth, whether the product enters demand scenarios such as AI servers, automotive electronics, industrial control and IoT.
Summary: Hong Kong-listed memory chip design-related companies should not be classified simply by the “chip stock” label. A more useful classification is to place each company within its supply chain position. GigaDevice is closer to the main memory chip design track, Shanghai Fudan is a composite company covering non-volatile memory, FPGA and security chips, while Montage Technology focuses on memory interconnect. SMIC, Hua Hong Semiconductor and ASMPT are related to the semiconductor cycle, but they are more focused on manufacturing, packaging or equipment support. When screening companies, you should prioritize product revenue, technology roadmap, customer scenarios and cyclical sensitivity, rather than only looking at whether the company name contains “semiconductor” or “chip.”

If you are looking for a Hong Kong-listed company most directly related to “memory chip design,” GigaDevice 3986.HK should be placed in the first tier. GigaDevice 3986.HK was listed on the Main Board of the Hong Kong Stock Exchange in January 2026. The company is an integrated circuit design company with products covering Flash, niche DRAM, MCUs, analog chips and sensors, giving it a clearer memory-related profile than many general Hong Kong-listed chip stocks.
From a product structure perspective, GigaDevice’s core memory-related products mainly include NOR Flash, SLC NAND Flash and niche DRAM. According to the GigaDevice 2025 annual report, the company’s 2025 revenue mainly came from specialty memory chips, MCUs, sensor chips and analog chips, with specialty memory chips accounting for more than 70% of revenue. This means the company is not merely associated with the “memory concept”; its revenue structure itself is highly tied to memory chips.
| Product Line | Is It a Memory Chip? | Application Scenarios | Key Analysis Points |
|---|---|---|---|
| NOR Flash | Yes | Code storage, automotive, industrial, IoT | Capacity upgrade, pricing cycle |
| SLC NAND Flash | Yes | Industrial, communications, embedded devices | Supply-demand changes, gross margin |
| Niche DRAM | Yes | Consumer, industrial, network communications | Cyclical elasticity, customer structure |
| MCU | No, but highly related | Industrial control, automotive, home appliances, IoT | Inventory cycle, price competition |
| Analog/sensors | Not core memory | Terminal electronic systems | Cross-selling capability |
Why is GigaDevice more closely related to memory design than an ordinary chip stock? The key is that its Flash Memory products support code storage, embedded storage and industrial-grade application demand. NOR Flash is usually used for program code storage, SLC NAND is more oriented toward embedded data storage, and niche DRAM is related to supply-demand cycles, product pricing and downstream demand changes.
MCU should be understood separately. MCU itself is not a memory chip in the traditional sense, but MCUs require program storage, firmware upgrades, data recording and peripheral control, and are often used in the same system as Flash, EEPROM and embedded NVM. Therefore, GigaDevice’s MCU business expands its customer coverage in automotive electronics, industrial control, home appliances and IoT, but MCU itself should not be directly equated with memory chips.
The risks also need to be separated. Memory products are significantly affected by pricing cycles, and the business climate for niche DRAM and SLC NAND may fluctuate quickly with supply-demand changes. MCU business may be affected by inventory cycles, price competition and customer qualification timelines. After dual A+H listing, valuation differences, liquidity and market sentiment may also affect share price performance.
Summary: GigaDevice is the clearest representative of memory chip design exposure among the Hong Kong-listed companies discussed here. You should not summarize it only as an “MCU company” or a “domestic chip stock.” Instead, you should separate Flash, SLC NAND, niche DRAM and MCU cycles. Flash and DRAM determine its memory chip relevance, while MCU determines its coverage in embedded systems and terminal customers. Opportunities come from domestic substitution, industrial and automotive electronics demand, and memory pricing recovery. Risks come from pricing cycles, inventory correction, intensifying competition and valuations that may already reflect optimistic expectations.

Shanghai Fudan 01385.HK is not a pure memory chip company, but it is highly related to memory chip design. The reason is that the company clearly has a non-volatile memory product line, while also operating FPGA, security and identification chips, smart meter chips and other businesses. The Shanghai Fudan 2025 annual report summary discloses that the company has established product lines including FPGA chips, security and identification chips, non-volatile memory, smart meter chips and integrated circuit testing services.
Shanghai Fudan’s most direct memory relevance comes from non-volatile memory. Non-volatile memory retains data after power is cut off. Typical products include EEPROM, NOR Flash and SLC NAND Flash. The company discloses that its memory products cover multiple interfaces, packages and capacities. Its main products include EEPROM, NOR Flash and SLC NAND Flash, used in communications, automotive electronics, industrial control, smart meters, medical instruments and other scenarios.
| Business Line | Relationship with Memory Chips | Relevance Level | Key Analysis Points |
|---|---|---|---|
| EEPROM | Typical non-volatile memory | High | Small capacity, high reliability, industrial applications |
| NOR Flash | Code storage | High | Capacity, interface, pricing |
| SLC NAND | Embedded data storage | High | Supply changes, application expansion |
| FPGA | Related to configuration storage and on-chip memory | Medium | Communications, AI, industrial, high reliability |
| Security MCU | Embedded secure storage | Medium | Finance, identification, IoT |
From a specific product perspective, Shanghai Fudan’s NVM products cover EEPROM, SPI NOR Flash and NAND Flash. EEPROM is more suitable for small-capacity, highly reliable data retention; NOR Flash is suitable for code storage; SLC NAND is suitable for embedded data storage. Its SLC NAND Flash products are aimed at embedded scenarios such as mobile phones, set-top boxes, communications equipment and networking devices.
Why should FPGA be included in a memory chip-related topic? FPGA is not a traditional memory chip, but it requires configuration storage, on-chip cache and high-speed data paths. Especially in communications, industrial control, AI edge computing, test and measurement, and high-reliability applications, FPGA is closely related to data access, caching and bandwidth. Shanghai Fudan’s FPGA products also cover PSoC, RFSoC and FPAI directions, making the company more like a composite IC design platform.
The difference between Shanghai Fudan and GigaDevice is that GigaDevice is more focused on the Flash, niche DRAM and MCU track, while Shanghai Fudan combines non-volatile memory, FPGA, security identification and smart meter chips. Shanghai Fudan’s advantage lies in its more diversified product matrix, while its risk lies in the fact that different business cycles may not move in sync. Investors need to analyze the growth quality of non-volatile memory, FPGA and security chips separately.
Summary: Shanghai Fudan can be classified as a Hong Kong-listed company related to memory chip design, but it should not be simply equated with DRAM, NAND or HBM manufacturers. Its memory relevance mainly comes from non-volatile memory products such as EEPROM, NOR Flash and SLC NAND, as well as embedded storage, identity authentication, data security and on-chip memory demand within FPGA, security MCU and smart meter chips. When analyzing Shanghai Fudan, you should treat it as a composite IC design company, not a pure memory chip manufacturer. Key areas to watch include non-volatile memory revenue, FPGA customer adoption, smart meter cycles and R&D efficiency.
Montage Technology 6809.HK is not a traditional memory chip manufacturer, because it does not produce DRAM, NAND or NOR Flash chips. However, it is a Hong Kong-listed chip design company highly related to the AI server memory chain. You can understand it as a key part of “memory system efficiency improvement,” with its core products including DDR5 memory interface chips, memory module companion chips, PCIe/CXL interconnect chips and server platform-related products.
The HKEX 6809.HK profile of Montage Technology describes its interconnect chips as mainly including memory interface chips, memory module companion chips, PCIe/CXL interconnect chips and clock chips. According to the Montage Technology 2025 annual report, the combined revenue from memory interface chips and Jintide server platform-related products accounted for the vast majority of consolidated revenue in 2025, indicating that its business is highly concentrated in server memory and platform interconnect chains.
| Product Direction | Is It a Memory Chip? | Relationship with the Memory Industry | Main Scenarios |
|---|---|---|---|
| DDR5 memory interface chips | No | Connect CPU and memory modules | Servers, AI training/inference |
| Memory module companion chips | No | Support stable memory module operation | DDR5 RDIMM/MRDIMM |
| CXL MXC | No | Memory expansion and pooling | Data centers, cloud computing |
| PCIe Retimer | No | High-speed interconnect | AI servers |
| Clock chips | No | System synchronization | Server platforms |
Why should Montage Technology not be described as a “memory chip manufacturer”? Because it is not a memory chip producer like Micron, SK hynix or Samsung, nor is it a NOR Flash or SLC NAND design company. Its core value lies in improving data transmission efficiency among processors, memory modules and server systems. Its Memory Interface products cover DDR2 to DDR5 memory interface directions and serve cloud computing and data center markets.
Why does AI server demand increase Montage Technology’s relevance? AI training and inference require higher memory bandwidth, capacity, latency performance and reliability. DDR5 upgrades, MRDIMM, CXL memory expansion and memory pooling all increase the importance of memory interconnect chips. The CXL Memory eXpander Controller is a CXL Type 3 DRAM memory controller that supports high-bandwidth, low-latency interconnect, memory pooling and sharing. In 2026, the company also announced that it had sampled its 9200 MT/s DDR5 RCD to key customers, showing that its product upgrades are closely tied to next-generation server platforms.
Summary: Montage Technology should be included among broadly memory chip design-related companies, but its position must be stated clearly: its core is not memory chips, but memory interconnect and server data paths. When analyzing Montage Technology, it is more appropriate to place it within AI data centers, DDR5 upgrades, CXL memory expansion, memory pooling and server platform upgrades, rather than directly comparing it with NOR Flash, EEPROM or MCU companies. Its opportunities come from AI servers and memory bandwidth demand, while its risks come from customer concentration, technology iteration, valuation volatility and changes in the server cycle.
Non-volatile memory, MCU and FPGA are not the same type of chip. Non-volatile memory stores data after power is cut off. MCU is a control chip that often integrates or connects to Flash. FPGA is a programmable logic chip that often requires configuration storage, on-chip cache and high-speed data paths. Therefore, they often appear together in automotive electronics, industrial control, IoT, communications equipment and edge AI, but they should not be mixed up in investment analysis.
The core function of non-volatile memory is “storage.” EEPROM is suitable for small-capacity and highly reliable writes; NOR Flash is suitable for code storage and fast reads; SLC NAND is suitable for embedded data storage. When judging whether a company is related to memory chip design, you should first check whether it has these products and whether these products contribute meaningfully to revenue and profit.
The core function of an MCU is “control.” It executes programs, reads sensors, controls peripherals and manages communication interfaces. Many MCUs include built-in Flash and may also connect to EEPROM, NOR Flash or NAND Flash. Automotive electronics, smart home appliances, industrial control, smart meters and IoT devices all require MCUs to work together with memory, so MCU companies are often analyzed within the embedded memory supply chain.
The core function of an FPGA is “reconfigurability.” It is suitable for high-performance, high-reliability, programmable data processing tasks. FPGA itself is not a memory chip, but configuration data, on-chip RAM, cache and external memory interfaces all affect system performance. Therefore, in AI edge computing, communications, data centers, test and measurement, and high-reliability applications, FPGA has a strong connection with memory capability.
| Chip Type | Core Function | Can It Retain Data After Power-Off? | Related Hong Kong-Listed Companies |
|---|---|---|---|
| EEPROM | Small-capacity reliable storage | Yes | Shanghai Fudan |
| NOR Flash | Code storage | Yes | GigaDevice, Shanghai Fudan |
| SLC NAND | Embedded data storage | Yes | GigaDevice, Shanghai Fudan |
| MCU | Control and computing | Depends on built-in storage | GigaDevice, Shanghai Fudan |
| FPGA | Programmable logic | Not equivalent to memory itself | Shanghai Fudan |
Why do users often search for MCU and memory chips together? Because terminal devices do not use chips in isolation. A smart meter, body control unit or industrial module may require an MCU to execute control, Flash to store firmware, EEPROM to store parameters and sensors to collect data. For investors, this means the customer scenarios of MCUs and non-volatile memory overlap significantly.
Summary: The relationship between non-volatile memory, MCU and FPGA can be summarized as follows: memory stores data, MCU executes control, and FPGA handles reconfigurable logic and high-speed data processing. These three types of chips often cooperate within the same terminal system, but they are not the same. GigaDevice’s memory exposure mainly comes from Flash, SLC NAND and niche DRAM, while MCU expands its application scenarios. Shanghai Fudan’s memory exposure comes from EEPROM, NOR Flash and SLC NAND, while FPGA and security chips enhance system-level value. Montage Technology goes beyond embedded terminals and directly enters AI server memory interconnect.
When assessing Hong Kong-listed companies related to memory chip design, you should not focus only on market themes. Instead, you need to analyze product cycles, revenue structure, customer validation, gross margin, valuation and trading costs. Memory chips have clear cyclical characteristics. MCUs may face inventory and price competition. FPGA depends on R&D investment and customer adoption. Memory interconnect is more closely tied to AI servers, DDR5, CXL and cloud computing capital expenditure.
From the opportunity side, AI servers raise demand for memory bandwidth and interconnect. Automotive electronics and industrial control support demand for MCUs and non-volatile memory. Domestic substitution increases attention on local IC design companies. DDR5, CXL and edge AI also bring structural opportunities. In its report on Montage Technology’s Hong Kong listing, Reuters described the company as a major global memory interconnect chip supplier, showing that market attention toward the AI memory chain has increased significantly.
From the risk side, GigaDevice should be assessed for memory price declines, MCU competition and A/H valuation differences. Shanghai Fudan should be assessed for non-volatile memory cycles, FPGA R&D investment and high-end customer adoption. Montage Technology should be assessed for server platform upgrade timing, customer concentration, technology iteration and high valuation volatility. Popular chip stocks are often driven by sentiment, but fundamentals ultimately return to revenue growth, gross margin and cash flow quality.
| Assessment Dimension | What You Should Watch | Relevant Companies |
|---|---|---|
| Memory cycle | ASP, inventory, gross margin | GigaDevice, Shanghai Fudan |
| AI servers | DDR5, CXL, memory interface demand | Montage Technology |
| Embedded market | MCU, EEPROM, NOR Flash | GigaDevice, Shanghai Fudan |
| Domestic substitution | Customer adoption, product certification | All three companies |
| Valuation risk | A/H differences, trading volume, earnings expectations | All three companies |
You should also pay attention to actual trading costs. If you track both Hong Kong-listed memory chip companies and U.S.-listed semiconductor companies, you need to understand not only their supply chain positions, but also trading fee structures. U.S. stock trading costs usually include not only commissions, but may also include platform fees, external institution fees and transaction activity fees. If the service is available in your region and you meet the applicable conditions, you can review Biya U.S. stock trading fees. Biya charges 0 USD commission for U.S. stock trading, while platform fees, external institution fees and other charges are subject to the fee center and order page display.
What signals suggest a company’s memory business is improving? You can watch for recovering revenue growth, improving gross margin, falling inventory, growing customer orders, stabilizing or rising prices, new product ramp-up and broader downstream demand. What signals suggest risk is rising? These include industry capacity expansion, price declines, customer order cuts, inventory rebuilding, rising R&D pressure and valuations that have already priced in optimistic expectations.
Summary: The key to analyzing Hong Kong-listed memory chip design-related companies is to convert “themes” into “verifiable indicators.” For GigaDevice, watch Flash, SLC NAND, niche DRAM and MCU cycles. For Shanghai Fudan, watch non-volatile memory, FPGA, security chips and smart meter business structure. For Montage Technology, watch DDR5, CXL and AI server memory interconnect demand. On the trading side, you also need to understand fees, liquidity, exchange rates and account eligibility. Public market information, trading rules and fee structures are for research purposes only and do not constitute investment advice. Actual trading should be based on platform rules, account statements and local regulatory requirements.
If you follow Hong Kong-listed memory chip design companies, you can also expand your view to the U.S. semiconductor supply chain, such as Micron, Marvell, Broadcom, AMD and NVIDIA, which occupy different positions across memory chips, networking interconnect, AI chips and server ecosystems. When using Biya to track U.S. and Hong Kong stocks, you can compare company financials, industry cycles, fee structures and order risks within the same framework. To look up U.S. semiconductor names, you can also use U.S. stock information search to build a watchlist. Service availability depends on your location, identity verification result, platform rules and applicable laws and regulations. Before trading, you should fully understand order types, fee structures, exchange rate fluctuations and market risks. You can also download the App to check available functions and actual displayed fees.
Hong Kong-listed memory chip design companies are responsible for product definition, chip architecture and sales, while wafer foundries are responsible for manufacturing. GigaDevice and Shanghai Fudan are more design-oriented, while SMIC and Hua Hong Semiconductor are more manufacturing-oriented. Both categories are affected by semiconductor cycles, but their earnings drivers, valuation logic and risk sources are different.
Shanghai Fudan is related to memory chip design, but it is not a pure memory company. It has non-volatile memory products such as EEPROM, NOR Flash and SLC NAND, while also operating FPGA, security and identification chips, smart meter chips and testing services. Its businesses should be analyzed separately.
GigaDevice’s MCU business itself does not count as traditional memory chip business. MCU is a control chip, but it is often tied to Flash, EEPROM, embedded storage, automotive electronics and industrial control scenarios, so it strengthens the company’s memory-related application coverage.
Montage Technology is considered a broadly memory-related company because it makes memory interconnect chips, not memory chips. Its DDR5, CXL and memory interface chips are highly related to AI server memory bandwidth, capacity expansion and data transmission efficiency.
Ordinary investors should assess product revenue, gross margin, customer adoption, inventory cycles and valuation levels. Do not rely only on concept labels. It is especially important to distinguish memory chips, MCUs, FPGAs, memory interfaces and wafer foundries to avoid comparing different supply chain positions under the same logic.
The main risks include memory price cycles, MCU price competition, customer validation cycles, supply chain constraints, R&D investment pressure and valuation volatility. For trading, fees and account eligibility, you should refer to platform rules, account statements and local regulatory requirements.
*This article is provided for general information purposes and does not constitute legal, tax or other professional advice from BiyaPay or its subsidiaries and its affiliates, and it is not intended as a substitute for obtaining advice from a financial advisor or any other professional.
We make no representations, warranties or warranties, express or implied, as to the accuracy, completeness or timeliness of the contents of this publication.



