Good example of optimization Intel scraps Optane and 3D XPoint

“Good example of optimization”: Intel scraps Optane and 3D XPoint direction

Intel’s earnings report revealed an interesting point unrelated to financial performance: During the second quarter, the company began to scale back Optane drive development and production. In addition, the document states that the company had to take over the associated “Optane Stock Decline in Value” $559 million in losses.

In other words, the company has started shutting down the Optane division, which was confirmed by Intel’s comment on the resource AnandTech: “We will continue to streamline our portfolio to support the IDM 2.0 strategy. This also includes evaluating the possible sale of companies that are either not profitable enough or do not meet our strategic goals. After careful consideration, Intel plans to end further product development within the Optane division. We will support Optane customers during the transition“.

As Intel Vice President Dave Zinsner noted:[Это] Two good examples of our ongoing portfolio optimization are the exit from the Optane and drone businesses. We continue to carefully review all opportunities to better direct our resources to the most valuable programs and increase the probability of success for each of these programs.“.

3D XPoint memory was announced by Intel in 2015 – the result of a joint project with Micron was positioned as a solution that combines the advantages of RAM and SSD. Bit-addressable memory is based on phase-change technology rather than electron capture like NAND. This allowed 3D XPoint to provide highly reliable drives – around one million rewrite cycles – and high speed, since data does not have to be combined into large blocks.

Based on the 3D XPoint technology, the company offered two product lines: non-volatile memory in DIMM modules for systems that require large RAM arrays (e.g. for database servers), as well as high-performance drives for servers and client computers, and high-speed ones -Caching modules for slower drives.

On the other hand, the uniqueness of 3D XPoint has been a big problem for Intel since the technology came out. Despite the fact that the solution offered good scalability at its core, Optane’s production costs were higher than NAND’s, so the drives turned out to be more expensive than high-performance SSDs. On the other hand, Optane DIMM modules offered more modest speeds compared to traditional DRAM at the same price. In other words, despite Intel’s attempts to combine the advantages of two solutions in one technology, 3D XPoint turned out to be inferior to DRAM and too expensive compared to NAND – Optane turned out to be very hard to sell.

As a result, Intel simply lost money for most (if not all) of the technology’s lifecycle: In 2020 alone, losses exceeded $500 million. Optane didn’t always show up in financial reports, but when it did, the numbers were always negative. Additionally, the company has accumulated a significant surplus of 3D XPoint chips — earlier this year, it was two years ahead, it said blocks & files. So the most recent quarter was marked by $559 million in Optane losses.

Thus, the closure of the direction was already a matter of time. It was decided to close the IMFT joint venture of Intel and Micron. The latter turned out to be the sole manufacturer of 3D XPoint and was subsequently forced to sell the company to Texas Instruments, which later repurposed it. Intel has since lost the 3D XPoint production center and apparently no longer needs it. Additionally, the company recently sold its NAND business to Korea’s SK hynix, which has turned it into Solidigm, meaning Intel is pulling out of the memory business entirely.

The decision to close 3D XPoint came at a critical time for the manufacturer. With the release of chips Sapphire Rapids Xeon The company planned to upgrade the Optane lineup to the third generation – the main ones were promised to be Crow Pass DIMMs with DDR5 support. Their development is now complete or nearing completion, but getting modules to market is becoming another burden for Intel: among other things, it needs to support customers over the long term.

With the departure of Optane, the company will shift priorities to CXL technology, which allows you to connect volatile and non-volatile memory to the processor via the PCIe bus – support for this solution is implemented in Sapphire Rapids chips. And Intel’s rejection of 3D XPoint will be a sad end to an interesting product line. 3D XPoint DIMMs were certainly an innovative idea, and 3D XPoint-based SSDs could offer performance that is unlikely to be replicated any time soon. So this is the end of an era for the SSD market.


About the author

Dylan Harris

Dylan Harris is fascinated by tests and reviews of computer hardware.

Add Comment

Click here to post a comment