Huawei Reportedly Partnering with Shenzhen Manufacturer to Build 12-Inch DRAM Wafer Fab
Taylor Wilson
Huawei is reportedly working with Shenzhen's SwaySure and the Chinese government to build a state-backed 12-inch DRAM fab targeting 140,000 wafers per month at 28nm. This means Huawei is attempting to replicate its logic-chip self-sufficiency playbook in memory — a direct response to tightening U.S. export controls.
What is this fab supposed to do?
The project is a 12-inch DRAM wafer fab. DRAM — the chip that serves as short-term memory in every phone, PC, and AI server — is one of the most supply-constrained components in tech.
It will start at 28nm process technology with a target of 140,000 wafers per month, a mid-scale facility globally but a ground-zero moment for China's domestic DRAM capacity.
This means → Huawei is not running a lab experiment; it is building a mass-production line meant to deliver real supply.
Who is running the project?
Three parties: Huawei brings demand and system-integration expertise, SwaySure (盛威旭科技) brings memory manufacturing experience, and the Chinese government backs funding and approvals.
The project has recruited a former TSMC fab director as CEO and a former Elpida executive as chief strategy officer — one knows how to build fabs, the other knows DRAM.
In plain terms = Huawei has no memory-chip manufacturing experience of its own, so it found a partner with a manufacturing base and hired industry veterans to lead.
Why does Huawei need its own DRAM now?
Samsung, SK Hynix, and Micron together control roughly 95% of the global DRAM market. Chinese tech firms rely almost entirely on these three for memory supply.
Huawei's AI infrastructure — Ascend chips, cloud services — and its smartphone business are both scaling fast, driving surging DRAM demand.
This reflects a core tension: the bigger Huawei's compute ambitions grow, the more dangerous its dependence on foreign memory supply becomes — especially as U.S. export controls continue to tighten.
Can this change the global DRAM landscape?
Analysts believe the fab is unlikely to challenge Samsung, SK Hynix, or Micron's global dominance in the near term — 28nm trails the industry frontier by a wide margin.
But the project's real goal is not global market share. It is to give Huawei a supply buffer: a fallback source if external procurement channels are cut off entirely.
This means → the measure of success for this fab is not market share — it is whether Huawei's product lines can keep running under a worst-case supply scenario.
Content is for reference only, not financial advice.