Huawei faces the challenge of being a startup in a new business as well as US government pushback
Chinese tech giant Huawei is reportedly entering the DRAM manufacturing business in a bid to cash in on the insane profitability of memory sales.
Three firms – Micron Technology, SK hynix, and Samsung Electronics — account for 95% of the DRAM on the market worldwide. The rest is small players, mostly in China. One of them, CXMT, is gearing to make a run for the market and try and take a little bit of their business. But it is a small player, especially compared to Huawei.
Huawei’s strategy is complex. It is working with various entities to circumvent the U.S. trade sanctions specifically targeting it. According to SemiconductorInsider on X.com, the Chinese government and DRAM chip maker Swaysure (formed in 2022) are planning to introduce a DRAM manufacturing plant with a capacity of 140,000 wafers per month.
For perspective, Samsung manufacturers approximately 500,000 wafers per month, and Micron makes about 250,000 wafers per month.
It all fits in perfectly with Beijing’s drive for semiconductor self-reliance uh especially after all of the U.S. sanctions. And it would be US interference, not Huawei’s inexperienced with making memory that will be its greatest challenge, says one analyst.
“Their biggest problem will be tooling up, since the US has a lot of sway over the world’s semiconductor equipment makers, and it’s using that sway to prevent tools from being shipped to China,” said Jim Handy, president of Objective Analysis.
Handy said he hears that Huawei and Swaysure are working to produce DRAM starting at 28nm. With that process they ought to be able to make 8Gb chips that yield reasonably but he doubts that they will be able to make HBM with that process, though.
Since he doesn’t know how far along they are in the process, Handy doesn’t know when something could come to market. “They may be on the cusp of shipping something, or they may be a couple of years away, but I don’t think they will wait any longer than that, so let’s say 2028. If there’s still a shortage in 2028, they are likely to ramp as hard as they can to get to something like 2-5% of the market by 2030. If the shortage ends before then, they will probably take a much longer time,” he said.
