Intel to Shut Down Optane Memory Business, Retire 3D XPoint Memory

[btarunr]

Intel's pioneering 3D X-point Memory, which sought to bridge the gap between non-volatile Flash memory, and volatile DRAM, stares at an untimely demise, as Intel plans to wind up both its Optane Memory business, as well as further development of 3D XPoint. The industry's reception of Optane Memory has been lukewarm; while cheap NVMe SSDs have driven Optane out of the client segment. Intel in its Q2-2022 Financial Results release, announced that it has initiated the winding down of the Optane Memory business, and that the company is incurring a $559 million "Optane Memory Impairment" charge this quarter.

3D XPoint faces technological competition from the latest crop of 3D-stacked Flash memory, which is achieving over 200 layers of density; while the latest generation of PCI-Express Gen 5.0 controllers enable data-rates in excess of 10 GB/s, and certain enterprise-relevant features of PCIe Gen 5. In a statement released to AnandTech, Intel says: "We continue to rationalize our portfolio in support of our IDM 2.0 strategy. This includes evaluating divesting businesses that are either not sufficiently profitable or not core to our strategic objectives. After careful consideration, Intel plans to cease future product development within its Optane business. We are committed to supporting Optane customers through the transition."



View at TechPowerUp Main Site | Source

 

[R0H1T]

Well looking at their financials I won't be surprised if they have to spin off their fabs as well, that chips subsidy can't come soon enough!

 

[TheLostSwede]

So many crying over this.
Yes, Optane was interesting, but it was way way way too expensive to be competitive.

 

[hat]

So many crying over this.
Yes, Optane was interesting, but it was way way way too expensive to be competitive.

Definitely not attractive in the consumer space, at least.

 

[TheLostSwede]

Definitely not attractive in the consumer space, at least.

Yet so many investing in it, because Intel.
It's obviously superior in terms of IOPS compared to current NAND based flash, but at 3-4x the cost, it's just not going to sell.
Personally I believe Intel messed up by pushing it for caching as well, no-one really asked for that.

 

[Why_Me]

Better the memory than the SSD's.

 

[LabRat 891]

Well, this is unfortunate news. Been a big fan of PCM. Was hoping to see prices come down as volume increased. Even had held out hope for NVPCM not-RAM coming to desktops. Nope...

Any chance of BAE Systems' Ovonyx Memory coming to the market? Basically the same thing as Optane (both were developed off the same patents).

 

[R0H1T]

If Intel can't make Optane work, they really should've licensed it, I doubt anyone else can come close!

 

[LabRat 891]

If Intel can't make Optane work, they really should've licensed it, I doubt anyone else can come close!

I agree. However, Optane is at least 3 levels of 'not all credit goes to Intel'.

1. The inventor and original patents holder of the particular kind of Phase Change 'crosspoint' deserves some credit. (Micron and Intel's 'work' extends from patents once held by the solar energy company owned by the inventor)
2. Before Micron and Intel got to the tech, British Military Industrial Contractor BAE Systems developed Ovonyx PCM off the previously mentioned's patents. (I wonder if this is related to Intel not licensing the technologies?)
3. Micron was instrumental in materializing Optane / 3DXpoint. Both on the engineering side, as well as manufacturing. (Their divestment and 'distancing' from Optane were sadly the beginnings of Optane's death throes)

 

[R-T-B]

Better the memory than the SSD's.

It's everything optane, SSDs included.

 

[Deleted member 185088]

I feel this is the most innovative thing from Intel, the one area where no one comes even close, yet they messed it up with stupidly high prices.

 

[R0H1T]

I agree. However, Optane is at least 3 levels of 'not all credit goes to Intel'.

If you're saying that about developing the tech then yes I agree but the downfall of it is mostly, if not totally, due to Intel.

yet they messed it up with stupidly high prices.

Yes, it really needed more "competition" ~ imagine if Toshiba hoarded everything about NAND 4 decades back!

And they had what 7-8 years to figure that out

 

[LabRat 891]

If you're saying that about developing the tech then yes I agree but the downfall of it is mostly, if not totally, due to Intel.

Yup. You got it. Was about the development and 'getting to market'.
I also agree that its failure is squarely on Intel's shoulders, and that 'the giant' Intel had the best chance at market penetration and proliferation of PCM.
(I'd say Micron 'had a part', but [with no evidence] I just *know* Intel did something to sour their relationship.)

On the bright side: as long as there's not more weird legal troubles surrounding 3DXpoint-Optane PCM, 'workaround' and competing PCM IPs will have a chance at the market. Thankfully, 3DXpoint/Optane/Ovonyx PCM isn't the only prospective option for Phase Change Memory in general.

 

[Chaitanya]

Yet so many investing in it, because Intel.
It's obviously superior in terms of IOPS compared to current NAND based flash, but at 3-4x the cost, it's just not going to sell.
Personally I believe Intel messed up by pushing it for caching as well, no-one really asked for that.

In consumer space Optane as a solution looking for problem to solve.

 

[R0H1T]

That's a cliched trope, it could've worked if Intel was willing to let go of margins but no they wanted their premium product competing against established players, especially in the NAND space! Caching works & it works brilliantly when implemented in a practical way, for me that was perhaps its biggest area/use case.

They were marketing Optane as an alternative to NAND & DRAM in some cases, without putting in the legwork to make it more accessible/universal & cheap(er) as compared to the incumbents.

 

[Tomorrow]

Yet so many investing in it, because Intel.
It's obviously superior in terms of IOPS compared to current NAND based flash, but at 3-4x the cost, it's just not going to sell.
Personally I believe Intel messed up by pushing it for caching as well, no-one really asked for that.

They should have pivoted it to enthusiasts not regular consumers. If Nvidia can sell $2000 GPU's then Intel could have sold $1000 SSD's.

If Intel can't make Optane work, they really should've licensed it, I doubt anyone else can come close!

Indeed. It could have given it a chance to succeed at least in the enthusiast space. I mean WD made 10K RPM HDD's viable because enthusiasts bought those even tho the capacities were low.

 

[LabRat 891]

Yet so many investing in it, because Intel.
It's obviously superior in terms of IOPS compared to current NAND based flash, but at 3-4x the cost, it's just not going to sell.
Personally I believe Intel messed up by pushing it for caching as well, no-one really asked for that.

There's more to Optane than just the IOPS. It has extreme write endurance, indefinite read endurance, and theoretically decades-to-centuries of data retention.

Intel, like most 'Big Players in Industry' are 'disconnected' from the wants and needs of their consumer base. Related: The marketing they used to try and push 'cache modules' as 'system memory' in OEM PCs was awful.

I'm unabashedly a fan of the technology; I otherwise wish Intel to 'pound sand'.

 

[DeathtoGnomes]

If you're saying that about developing the tech then yes I agree but the downfall of it is mostly, if not totally, due to Intel.

Yes, it really needed more "competition" ~ imagine if Toshiba hoarded everything about NAND 4 decades back!

And they had what 7-8 years to figure that out

Intel assumed they could corner a market, Intel being Intel as usual is no surprise, but what of its Ngeedia brothers?. If Intel had put even half the cash into Optane marketing as they do into unofficial 'other' leaks, and took a step back from the koolaid, Optane could have been a much more popular product.

 

[Gungar]

There's more to Optane than just the IOPS. It has extreme write endurance, indefinite read endurance, and theoretically decades-to-centuries of data retention.

Intel, like most 'Big Players in Industry' are 'disconnected' from the wants and needs of their consumer base. Related: The marketing they used to try and push 'cache modules' as 'system memory' in OEM PCs was awful.

I'm unabashedly a 'fanboi' of the technology; I otherwise wish Intel to 'pound sand'.

And lowest latency.

 

[TheLostSwede]

There's more to Optane than just the IOPS. It has extreme write endurance, indefinite read endurance, and theoretically decades-to-centuries of data retention.

Intel, like most 'Big Players in Industry' are 'disconnected' from the wants and needs of their consumer base. Related: The marketing they used to try and push 'cache modules' as 'system memory' in OEM PCs was awful.

I'm unabashedly a 'fanboi' of the technology; I otherwise wish Intel to 'pound sand'.

The tech wasn't ready for consumers and Intel forced it in ways no-one had asked for. The Optane/QLC SSD they made was a true abomination.

Yes, the tech itself wasn't bad, it just took too long to get to a point where it made sense from a usage perspective and by then, Micron had already thrown in the towel.

 

[LabRat 891]

And lowest latency.

I did indeed forget that part. Important, since it was a huge part of its marketability as a supplement/replacement for giant RAM pools in server infrastructure.

On that note, Intel keeping it exclusive in an attempt at regaining lost server-space market share hurt its adoption. (I'm not sure about the engineering hurdles to having Optane platform-agnostic and 'DDR#-intercompatible')

 

[ExcuseMeWtf]

It's a shame it came to that. We're not getting consumer grade NAND alternative with superior lifetimes anytime soon, are we?

 

[dgianstefani]

What a shame. Even as small 100-200gb OS/Program drives they are fantastic products. The 118gb for example can be had for under £100 new. Latency, IOPS, practically infinite endurance etc. Brilliant product killed by economies of scale.

 

[bug]

Yet so many investing in it, because Intel.
It's obviously superior in terms of IOPS compared to current NAND based flash, but at 3-4x the cost, it's just not going to sell.
Personally I believe Intel messed up by pushing it for caching as well, no-one really asked for that.

I believe they hit some technological barriers they couldn't clear.
And yes, nothing will reach those IOPS for quite a while, sorry to see it go.

Also, fanboyism has nothing to do with buying into Optane. Early adopters or tech-enthusiasts were around since forever. Remember when CD-ROM drives were >$1,000? Some people bought those, too.

 

[ZoneDymo]

Sigh, capitalism at its finest.