Intel has, it seems, competed in every major memory technology and exited them all. Last week we saw the company unload its NAND flash and SSD assets to SK hynix for $9bn. The deal is structured in 2 phases with $7bn expected later in 2021 and IP related to NAND manufacturing/design (along with R&D & Dalian Fab personnel) is not slated to transfer until March 2025 at an additional $2B. That two-phased approach is almost certainly down to Intel’s NAND IP entanglements with Micron, believed to expire in 2025.
SK hynix gets a significant boost in scale, something it needed in an industry where the alternative to scaling is failing. Adding Intel’s business to its own should install SK hynix as the second-largest vendor for enterprise SSD – the combined businesses fulfil around 32% of the market, compared to Samsung’s 35%. They’ll be some loss from the disruption, but SK Hynix will certainly retain second place.
SK hynix also gains access to high-density floating-gate QLC technology, which is what all the cool enterprises are looking for in their SSD solutions, and it can ramp up Chinese fabrication with the fab in Dalian.
However, in SSDs these acquisitions rarely result in 1+1=2. SK hynix must be decisive on the 2 NAND technology roadmaps it will now support, and aggressively defend them against an intense market share rush from SSD competitors. That decision will need to be clearly communicated, so buyers (and partners) understand the SK Hynix roadmap.
For Intel, this deal removes a significant drag on its margin numbers. Intel keeps Optane, which makes sense as there’s a strategic value in Optane which is missing from the NAND business. By integrating the CPU with Optane Intel will be able to eliminate the memory bottleneck, creating faster processors and creating another point of differentiation. Abandoning Optane would also have annoyed a load of Intel partners who’ve invested in the technology, so Intel is obliged to continue with scaling and reducing the cost, despite the lack of profitability.
Overall, this deal benefits both companies, however, it’s not going to will make the NAND industry healthier or more rational. The #5 player purchasing #6 player and achieving ~20% NAND sales market share in 2020 is not significant enough to restructure the industry. Given the heightened competition, and inevitable impact from China, the roller-coaster ride of boom/bust cycles will continue, and they will be more consolidation to come.
It’s never dull in the world of memory.