Intel Considering a Return to the Storage Market or Exploring CPU and Storage Chip Stacking Solutions
2026-08-17
Recently, Intel CEO Chen Liwu stated that memory is no longer just a "commodified" business, but has become a strategically significant field again, and hinted that the company is evaluating new memory architectures and the possibility of stacking memory on CPUs. This conversation took place against the backdrop of his discussion on the return and innovation of the US chip industry, and also reflected that the industrial environment has changed since the rise in memory demand driven by AI data centers.

Chen Liwu pointed out that in the past he would think that storage was not worth investing in, but now the situation has changed because many new technologies are emerging. He mentioned that he is currently following a project that he is personally interested in, which is related to a new type of memory architecture, but it is not yet possible to disclose it publicly. He also stated that stacking memory and CPU may be a meaningful approach, and believes that there are many exploratory stacking methods between CPU and memory.

The outside world has also noticed that Intel recently hired former CEO of SK Hynix, Lee Seok hee. Chen Liwu mentioned this personnel arrangement in his conversation, which seems to imply the company's layout direction in the memory field. The relevant statement also echoes Intel's previously disclosed XBM patent, which is said to eliminate the silicon intermediate layer used in HBM, indicating that Intel is looking for a different way of memory integration than traditional routes.

However, there are still many variables regarding whether Intel will truly return to the memory manufacturing market. Intel used to be a memory company in its early years and was a successful memory manufacturer after its establishment in 1968, but withdrew from the market in the 1980s due to the rise of Japanese manufacturers. Afterwards, the company attempted to re-enter the relevant field through products such as NAND, Optane, and RDRAM, but ultimately failed to sustain its efforts.

Nowadays, despite the significant increase in net profits of 3D NAND and DRAM manufacturers due to tight supply and demand, Intel still needs to face huge capital expenditures, research and development investment, and time costs if it wants to fully return to the storage market. In the current context where the OEM business still needs to make full efforts, the outside world is also continuously paying attention to how Intel balances resources between its core business and the new storage track.