Samsung aims to tap into deeply diversified revenue streams and intends to double its investment in memory technology for the future. recently announced (opens in new tab) We have no plans to cut capex on new memory technologies, such as GDDR7, which will ultimately be the best graphics card of the future, or ultra-high-rise V-NAND for the best SSD. Nor do we plan to reduce production despite the slowdown in market conditions.
The company believes its spending plan is necessary Secure technical superiority over competitors, consolidating its position as a leading memory chip manufacturer. To that end, Samsung is betting on a next-generation manufacturing process that can scale memory density for DRAM, VRAM, and the latest highest density NAND that can power future data-intensive workloads.
In the DRAM field, the company announced that its latest 5th generation 10 nm class manufacturing (1b) will start mass production chips by 2023. Exploratory research is underway for sub-10 nm DRAM fabrication employing new patterning, materials and structures. Designs containing High-K gates. Samsung generally leads most of its rivals manufacturing his DRAM chips at his 14nm class node.
Samsung’s next-generation GDDR7 memory is sure to grace future graphics accelerators, and it’s already in production. It offers speeds up to 36 Gbps, double the 18 Gbps of GDDR6. At these data rates, a 384-bit memory bus can provide approximately 1.728 TB/s of bandwidth. This is a significant increase from the upcoming RTX 4090’s 1 TB/s bandwidth. This will allow GPU manufacturers to have sufficient bandwidth without having to increase bus width. This can make the PCB more expensive and have an even worse price impact.
When it comes to NAND, where Samsung has established itself as the undisputed champion, the company is now making 9th and 10th generation V- Designing and prototyping NAND. Samsung is currently shipping its 7th generation 176-layer V-NAND and plans to release V-NAND chips based on its 8th generation 230-layer design before the end of the year. The latter offers a 42% density increase over a 512 Gb chip.
But Samsung is aiming for even bigger density leaps, and is on track to achieve a 1,000-layer V-NAND design by 2030. Samsung also continues to work on his QLC (Quad-Level Cell) technology, hoping to improve performance while increasing his bits of storage. density.
“Even if the current situation is not good, we will not change the course we have already drawn,” said Han Jin-man, executive vice president and head of global memory sales and marketing at Samsung. . “So far, no one at Samsung is talking about cutting chip production.”
Several other memory makers have announced cutbacks in memory production in response to slowing demand, a situation partially due to general macroeconomic conditions. Weak demand has already created a glut in the memory market, which Micron’s CEO has described as “unprecedented” and the company plans to cut capital spending by more than 30% (around $8 billion) and reduced spending on wafer fab equipment by 50%.
In Japan, Kioxia also announced a 30% reduction in memory wafer production, which it hopes will be enough to absorb inventory. The two companies have more price cuts due by the end of the year and are looking to reverse the downward pressure on his NAND’s price that took a hit throughout 2022.
Samsung, by contrast, chooses to double down on its technology leadership even as its competitors play the safe game by reducing production. Samsung’s multiple divisions and revenue streams are somewhat shielded from cyclical market and demand conditions, so rather than simply blindly keeping course, the question becomes who can do it. Samsung’s apparent refusal to lower production levels may end up helping his DIY enthusiasts. The company also believes it can widen the technological gap with its competitors by staying on course with its investment roadmap.
