Microprocessor giant Intel (NASDAQ:INTC) announced a new memory technology dubbed 3D XPoint back in 2015. The technology is a type of non-volatile memory that the company says can be used as either extremely fast storage technology or as a substitute for traditional DRAM-based memory modules in some enterprise server applications.

In late 2015, Intel told investors that with its 3D XPoint memory modules, it had the opportunity to go after the data center memory market, which the company projected would be worth about $34 billion by 2020.

A wafer of Intel's 3D XPoint memory chips, a technology that the chipmaker hopes to deploy as both storage and memory.

A wafer of Intel's 3D XPoint memory technology. Image source: Intel.

Clearly, Intel can't address the entirety of the server DRAM market with 3D XPoint memory modules -- the technology just isn't suitable for all applications. Fortunately, Intel's data center group (DCG) chief, Diane Bryant, went into some additional detail around the true size of the opportunity for Intel's 3D XPoint memory modules and the assumptions that Intel used to quantify that opportunity.

An $8 billion opportunity by 2021

Intel says that by 2021, the opportunity for its 3D XPoint memory modules in the data center could reach $8 billion. That's clearly much smaller than the $34 billion-by-2020 figure that Intel had previously put out there, but in the context of the roughly $17 billion in revenue that Intel's DCG segment generated in 2016, it's a significant opportunity.

"With 3D XPoint, it is a very compelling value proposition to the end user because we can deliver twice the memory footprint capacity in a system at a 40% lower cost structure [compared to DRAM]," Bryant said. She characterized Intel's expectations in this market through 2021 as "pretty conservative," noting that the company assumes "single-digit server attach rate of 3D XPoint memory out through [Intel's forecast] horizon."

In other words, of all the servers sold in the market over that time, Intel expects just a "single-digit" percentage of them (so anywhere from 1% to 9%) to include 3D XPoint memory modules.

She further clarified that Intel isn't going after the entirety of the server market with these products, but is instead targeting the narrower market of servers that "are deploying 512 gigabytes -- half a terabyte -- of memory or above."

When will the revenue begin flowing in?

Intel had originally planned to start shipping 3D XPoint memory modules alongside its upcoming Skylake Server processors, which are expected to launch in the middle of 2017. However, the company admitted that the volume sales ramp of these products has been pushed out to next year.

What this means for investors, then, is that we shouldn't see significant, if any, 3D XPoint memory module sales in the company's DCG revenue this year. In fact, Intel explicitly said that one of the reasons that it took down its DCG revenue growth target through 2021 was the pushout of some new products, including 3D XPoint.

The good news is that if this technology is as compelling as Intel claims that it is and the company's got customers lined up and ready to buy it, then the ramp-up of 3D XPoint (as well as other technologies) should lead to an acceleration in DCG growth beyond the "high-single-digit" percentage that the company expects to see in 2017 in 2018 and beyond.