Transcription
Huawei may be building something no other Chinese technology company has ever controlled at this scale. The processor, the memory, the storage, the software, and perhaps even the factories that produce the most important chips inside its devices. And that matters because Apple, despite designing some of the world's most advanced consumer processors, still depends on outside companies for the memory and flash storage inside every iPhone, iPad, and Mac. Huawei appears determined to reduce that kind of dependence as far as possible.
The latest reports focus on DRAM, the fast-working memory used by smartphones, computers, servers, vehicles, and nearly every modern electronic system. Huawei has not confirmed that it owns or operates a DRAM manufacturing network. In fact, it has denied claims that it controls a hidden group of semiconductor factories. But analysts, industry publications, corporate connections, and earlier satellite investigations have continued pointing toward a much deeper relationship between Huawei and several Chinese chip companies than official ownership records suggest.
One name now receiving major attention is SwaySure, a Shenzhen memory company that has been linked to a planned 12-in DRAM facility. Reports describe a production target of roughly 140,000 wafers every month, although that figure remains unconfirmed. If the facility reaches anything close to that scale, it would not be a laboratory experiment. It would be industrial capacity designed to supply large volumes of memory for Chinese phones, computers, servers, network working equipment, and other products.
That is where this story becomes much bigger than another new semiconductor factory. Huawei already develops its own Kirin processors to HiSilicon. It sells enterprise solid-state drives and storage systems, and it has spent years developing proprietary SSD controllers, storage software, and flash management technology. Huawei also builds phones, computers, networking equipment, cloud infrastructure, vehicle systems, and data center hardware. Adding a reliable domestic source of DRAM could connect those separate pieces into one unusually integrated technology platform.
Apple is often described as the master of vertical integration. It designs the iPhone, the operating system, the processor architecture, many supporting chips, and the services surrounding the device. That control creates excellent performance because Apple can optimize hardware and software together. But Apple does not manufacture its own leading-edge processors. TSMC produces them. Apple does not fabricate its own DRAM. Suppliers such as Samsung, SK Hynix, and Micron provide memory. Apple also relies on outside companies for NAND flash and many other essential components.
Huawei's ambition appears different. It is not merely trying to design the best phone processor or create a polished operating system. It is trying to build a supply chain that can continue functioning even when foreign governments restrict access to crucial technologies. In that environment, owning more of the stack is not only about performance or profit. It becomes a form of strategic protection.
The United States placed Huawei on the entity list in 2019, sharply limiting its access to American technology. Later export rules also restricted foreign-made chips produced with certain American software or equipment. Those measures disrupted Huawei's smartphone business, weakened its access to advanced processors, and forced the company to search for alternatives across nearly every level of semiconductor development.
For several years, many observers believed the restrictions had permanently stopped Huawei's momentum. Its global phone shipments collapsed. Google services disappeared from new international devices, and advanced chip supplies became difficult to secure. Then Huawei returned with new premium smartphones powered by domestically produced Kirin chips. Those processors did not immediately match the most advanced Apple or Qualcomm products, but their existence demonstrated that China's semiconductor network had progressed further than many outside analysts expected. Now, the next stage may be memory.
This timing is important because memory has become a geopolitical product, not merely a commodity. Washington is debating how Chinese DRAM should be treated, while Beijing is encouraging domestic substitution across strategic supply chains. A Huawei-linked memory project would therefore serve two goals at once, supplying commercial products and insulating critical Chinese hardware from future restrictions imposed unilaterally abroad.
A processor receives most of the attention because it performs the calculations and carries the recognizable brand name, but a processor without memory cannot do useful work. DRAM temporarily holds the information that the processor needs immediately. Faster memory can improve responsiveness, multitasking, gaming, professional software, and the efficiency of large computing systems. When memory is scarce, expensive, or restricted, even a powerful processor can become difficult to deploy at scale.
Huawei currently has to obtain memory through suppliers and partners, just like most technology companies. China's strongest established DRAM producer is CXMT, which has expanded rapidly and is now at the center of political pressure in Washington. American lawmakers have discussed tighter restrictions on Chinese memory chips, especially as reports suggest major global companies have evaluated Chinese components. That pressure gives Huawei another reason to support an alternative source.
If Swei Shou can produce usable DRAM in large quantities, Huawei could gain a second domestic pathway. It could potentially shape memory specifications around its own processors, storage systems, phones, servers, and networking products. It could coordinate product development earlier, test components more closely, and reduce the risk that a foreign supplier suddenly becomes unavailable because of sanctions or licensing decisions.
However, building DRAM is extremely difficult. Memory chips may look repetitive compared with complex processors, but manufacturing them economically requires enormous precision, strong yields, advanced materials, reliable equipment, and years of process knowledge. A factory can technically produce working chips and still lose money if too many wafers contain defects. Samsung, SK Hynix, and Micron have spent decades improving production and spreading their research costs across huge global businesses.
The reported Swasier facility is associated with 28 nanometer DRAM technology. That sounds old when compared with the small node numbers used to market processors, but memory processes cannot be compared directly with logic processes. Mature equipment can still produce valuable memory for industrial devices, networking systems, automobiles, consumer electronics, and many other products. The first objective may not be defeating the latest Samsung memory chip. It may be creating dependable capacity using tools China can obtain, maintain, and eventually replace domestically.
This strategy would follow a familiar pattern. Start with products that are possible today. Use domestic demands to keep factories busy. Improve yields through repeated production. Train engineers. Localize more equipment and materials. Then move toward higher density, lower power consumption, and more advanced memory generations.
Huawei has the scale to support that process. According to its official 2025 annual report, the company generated 880.9 billion yuan in revenue and invested 192.3 billion yuan in research and development. That means more than 1/5 of annual revenue went back into research. Huawei also reported that approximately 114,000 employees worked in research and development by the end of 2025. Few companies anywhere can sustain that level of engineering investment. The money does not guarantee success, but it changes the odds.
A smaller company might abandon a difficult semiconductor project after several years of losses. Huawei can spread the cost across telecommunications, consumer devices, enterprise storage, computing, automotive technology, energy products and services. It can also become the customer for its own components, giving new factories demand before they are competitive in the open market. That internal demand is one of the biggest advantages of vertical integration.
Imagine Huawei developing a new processor for a future phone. If it also influences the DRAM, storage controller, SSD firmware, operating system, and device design, engineers can optimize the entire system instead of waiting for standardized components from unrelated suppliers. They might reduce power consumption, improve data transfer, change packaging, or prioritize features designed specifically for Huawei products.
Samsung is the clearest example of this model. Samsung designs processors, manufactures chips, produces DRAM and NAND flash, builds displays, assembles phones, sells appliances, and operates enormous factories. It does not control every component in every product, but its manufacturing depth gives it options that most competitors lack. Huawei appears to be moving toward a Chinese version of that broad industrial structure. Although it remains far behind Samsung in advanced memory production and global manufacturing experience.
The comparison with Apple is more complicated. Apple's business model does not require owning chip factories. It uses its purchasing power to secure leading production from specialized suppliers, while concentrating its resources on design, software, product integration, and services. That model has been extraordinarily successful. Apple can access TSMC's newest manufacturing without carrying the cost and risk of operating fabrication plants itself.
But Apple's model depends on an international supply chain remaining available. Apple can negotiate with suppliers, diversify orders, and reserve enormous amounts of production, but it cannot directly command a memory factory it does not own. Huawei has experienced what happens when political decisions suddenly break those relationships. Therefore, Huawei may accept higher costs and less advanced technology today in exchange for greater control tomorrow.
This does not mean Huawei is already independent. It still depends on outside companies, imported equipment, specialized materials, manufacturing partners, and technologies that China has not fully replaced. Even its domestically produced advanced processors have been associated with SMIC rather than factories openly owned by Huawei. Semiconductor manufacturing equipment remains a major weakness because the most advanced lithography, inspection, deposition, and testing systems are concentrated among a small number of foreign suppliers.
There are also serious questions about the reported network of 11 fabs. Industry reports have described factories connected to Huawei through employees, financing, customers, technology transfers, or local government support. Yet, indirect relationships are not the same as legal ownership or operational control. Huawei denies operating a secret fab network, and public evidence does not allow outsiders to verify every claim. Any responsible interpretation must separate confirmed Huawei products from alleged manufacturing connections.
What is confirmed is Huawei's direction. The company openly develops processors, computing systems, enterprise SSDs, storage controllers, networking chips, operating systems, and full hardware platforms. In May 2026, Huawei presented a new semiconductor design strategy built around advanced packaging, architecture, software, and system-level coordination rather than relying only on smaller transistors. Huawei also said future Kirin chips would use its logic folding architecture to improve performance through greater integration.
That approach reveals how Huawei intends to compete while access to the most advanced manufacturing remains limited. Instead of waiting until Chinese factories can copy every part of the global supply chain, Huawei is redesigning products around the capabilities it has. Multiple chiplets can be connected through advanced packaging. Software can schedule workloads more efficiently. Larger systems can combine many processors. Memory, storage, networking, and computation can be coordinated as one platform.
DRAM fits naturally into that strategy. Memory bandwidth and latency can become bottlenecks when processors are combined into larger systems. A company that controls the processor architecture that depends entirely on external memory road maps has less freedom. A company that can influence both sides can experiment with packaging, interfaces, capacity, and power targets designed around its own products.
The same logic applies to storage. Huawei already sells Ocean Disk Enterprise SSDs and Ocean Store all-flash systems. Its public materials describe proprietary SSD controller technology and end-to-end storage development. Yet, selling an SSD does not necessarily mean fabricating every NAND memory cell inside it. The strategic goal is not instantly owning every atom of the device. It is steadily gaining control over the most important designs, controllers, software layers, suppliers, and production routes.
For China, Huawei's expansion could also support a wider domestic ecosystem. A DRAM facility needs wafer equipment, photo resists, gases, chemicals, masks, testing tools, packaging services, and trained engineers. Even if early chips trail foreign competitors, each production run creates experience for local suppliers. That industrial learning can spill into other companies and make future restrictions less effective.
For foreign competitors, the immediate threat may not be Huawei taking global market share with the world's fastest DRAM. The bigger threat is Huawei becoming harder to pressure. Sanctions work most effectively when a target depends on something only outsiders can provide. Every domestic processor, memory chip, controller, production tool, and software platform removes one potential point of leverage.
This could reshape competition inside China first. Huawei could favor its own components, direct large orders toward affiliated suppliers, and bundle processors, memory, storage, networking, and software into complete systems. Chinese customers seeking stable domestic supply may accept lower efficiency or higher prices, especially in government, telecommunications, infrastructure, and strategically important industries.
Over time, scale could reduce those disadvantages. More orders can finance better equipment. Higher production volumes can improve yields. Larger engineering teams can solve defects faster. Suppliers can justify developing specialized materials. What begins as an expensive sanctions workaround can eventually become a competitive industrial base.
Still, the path is not guaranteed. Memory markets are brutally cyclical. Prices can collapse when too much capacity enters the market. Established producers can lower prices, accelerate investment, and use their technological lead to squeeze new competitors. Huawei and SMIC would need to prove not only that they can manufacture DRAM, but that they can do so reliably, efficiently, and repeatedly across multiple generations.
There is another risk, stronger American restrictions. If Washington concludes that Huawei is using related companies to obtain restricted tools or technology, more firms could be added to export blacklists. SMIC has already faced American scrutiny. Expanded controls could make equipment maintenance, software updates, spare parts, and future factory expansion more difficult.
But restrictions can produce two opposite effects. In the short term, they can slow production and raise costs. In the long term, they can convince a large country that dependence itself is unacceptable. Huawei's current semiconductor strategy is a direct result of being cut away from suppliers it once treated as normal business partners. The more pressure increases, the more resources China may direct toward replacing every vulnerable component.
That is why the real story is not whether one rumored factory immediately beat Samsung or Micron. The real story is whether Huawei can connect enough domestic capabilities to create a self-reinforcing technology system. Huawei processors generating demand for Chinese fabs, Chinese memory supporting Huawei devices, Huawei storage products creating demand for controllers and NAND, and domestic equipment companies learning from every new production line.
Apple's ecosystem is powerful because millions of customers, developers, devices, services, and custom chips reinforce one another. Huawei is attempting to build a different kind of ecosystem based not only on software and consumer loyalty, but on industrial survival. It wants the device, the operating system, the processor, the networking, the storage, the memory, the packaging, and eventually more of the machinery beneath them.
In my opinion, Huawei is still years away from true semiconductor independence, and some claims about its hidden factories may be exaggerated. But dismissing the strategy because its first products are behind would repeat the mistake many observers made before Huawei's smartphone comeback. The company does not need to lead immediately. It needs to keep producing, learning, improving, and removing one dependency at a time.
If Huawei eventually controls reliable supplies of CPUs, DRAM, SSD technology, and advanced packaging, America will face a much harder target than the company it sanctioned in 2019. Apple may continue building faster and more profitable devices, but Huawei could possess something increasingly valuable in a divided world, the ability to keep building when outside suppliers are ordered to stop.
So, what do you think? Is Huawei creating the next Samsung-style semiconductor empire, or will manufacturing limits prevent it from closing the gap with Apple and the world's leading chipmakers? Tell us your opinion in the comments, and for more stories about China's processors, memory, storage, manufacturing, and the global technology race, subscribe to the Eastern Engine channel.