News Solidigm pulls out of consumer SSD market with discontinuation of drives – Storage company shut down consumer division over a year ago

Why sell consumer SSDs at a loss when you can sell DC SSDs at massive profit margins?
Of course, you don't want to sell them at a loss. That's obviously a deal-breaker, but the companies in consumer market aren't generally taking a loss on them. I think one big advantage is that it lets them bin their NAND, reserving only the highest grade for their datacenter markets. In general, it should give them more return-on-investment, since most of their controllers, firmware, etc. can be shared across both consumer & datacenter product lines.

2022-2023 was quite an exceptional period, for SSDs in general - not just consumer! I even bought a datacenter Solidigm SSD in mid-2023 for about half what it was selling for, a year later.

Intel used to be one of my go-to brands, for SSDs. Last year, I even bought a couple SK Hynix consumer drives (P31 Gold). I was looking forward to what the merged entity would produce. This is disappointing.
 
Of course, you don't want to sell them at a loss. That's obviously a deal-breaker, but the companies in consumer market aren't generally taking a loss on them. I think one big advantage is that it lets them bin their NAND, reserving only the highest grade for their datacenter markets. In general, it should give them more return-on-investment, since most of their controllers, firmware, etc. can be shared across both consumer & datacenter product lines.

2022-2023 was quite an exceptional period, for SSDs in general - not just consumer! I even bought a datacenter Solidigm SSD in mid-2023 for about half what it was selling for, a year later.

Intel used to be one of my go-to brands, for SSDs. Last year, I even bought a couple SK Hynix consumer drives (P31 Gold). I was looking forward to what the merged entity would produce. This is disappointing.
Disappointing but not really all that surprising. When Intel sold off its SSD division and Solidigm was formed, it was always going to be an uphill battle. Intel didn't see a future in the market segment. Could a separate entity do better? Apparently the answer is no.

RIP, Solidigm. I'm still sad that the promises of 3D XPoint never panned out as well.
 
  • Like
Reactions: snemarch and King_V
I think it's a shame. The few times I've looked, I hadn't been able to even find Solidigm consumer nvme drives in stock. I've kind of settled on mostly WD Black lately as Samsung has been a little overpriced IMO and the controller firmware issue from last Gen affected me.
 
Disappointing but not really all that surprising. When Intel sold off its SSD division and Solidigm was formed, it was always going to be an uphill battle. Intel didn't see a future in the market segment. Could a separate entity do better? Apparently the answer is no.
Intel needed money and Gelsinger wanted to focus on CPUs, GPUs, and AI. That's different than "not seeing a future".

Clearly, there are companies having success in the consumer SSD market. Also, Intel didn't only sell off their client SSD business. The same business unit also made their datacenter drives. The only part SK Hynix didn't get was the Optane business. In fact, the Solidigm datacenter drive I bought last year actually arrived with Intel branding, since apparently there was still some older stock in the channel.

RIP, Solidigm.
They're still around. Just not doing client SSDs, any more.

I'm still sad that the promises of 3D XPoint never panned out as well.
Yeah, that got wound down before the sale. I assume if there were any buyers to be had for Optane, Intel would've sold it instead of shuttering it. They never turned a profit on it, though. I'd have loved to see the Optane CXL drives they were allegedly working on.
 
  • Like
Reactions: King_V
I hope someone is able to get some more details on the situation. I think I mentioned it in one of the SSD review threads here that I was surprised when I went looking for the software for my P44 Pros when I put my system together and found the drives discontinued. This all seemed to happen under the radar. I think the biggest loss here is the custom driver development (someone over on the TPU forums even has it working on other drives).

I think given the information in this article it's likely they'd laid off everyone working in the US in 2023. Allyn Malventano who'd worked for Intel and moved over to Solidigm started working for Phison sometime late 23/early 24. He had been one of the people behind the custom driver push which was something Intel hadn't been interested in doing. It doesn't particularly seem like there's going to be a big client push from Phison, but they are hitting datacenter.

NAND client drives are a low margin business and Solidigm had been acquired for the enterprise business (SK Hynix had no real presence here). Now the real question is what does this mean for the client facing SK Hynix drives. The P44 Pro used the same hardware as the P41 SK Hynix drive and there is a P51 though it still hasn't been released to market. SK Hynix also has an OEM client division which is where these drives all originated from. Now there's Klevv (under the same corporate umbrella) who uses SK Hynix NAND/DRAM but multisources controllers. Klevv doesn't seem like it will be going anywhere and I doubt the SK Hynix OEM business will either, but the retail client division it's hard to say.

I've talked multiple times about the death of 3D XPoint and how bad that was from a technology standpoint. Rather than rehash it though there is something new I thought of while pondering it's death reading this thread. While there is no doubt it was a money sink it might have actually survived if Intel had been getting EUV machines in volume pre-pandemic. I've always assumed part of its death can be attributed to the manufacturing process being driven by Micron who likely used industry standard tools. This meant that while it's nowhere near as complex as a CPU/GPU it would likely have meant a non-insignificant amount of work to spin it up on Intel's own fabs. During a transisition time like DUV to EUV it might have made sense to manufacture it on early EUV nodes since it should have been simple and high yield. It also might have made sense to do the work to produce it on DUV to keep those fabs working while designing a new DUV node with industry standard tools.
 
it might have actually survived if Intel had been getting EUV machines in volume pre-pandemic. I've always assumed part of its death can be attributed to the manufacturing process being driven by Micron who likely used industry standard tools. This meant that while it's nowhere near as complex as a CPU/GPU it would likely have meant a non-insignificant amount of work to spin it up on Intel's own fabs. During a transisition time like DUV to EUV it might have made sense to manufacture it on early EUV nodes since it should have been simple and high yield.
Assuming it could continue scaling in 2 dimensions, a key question is whether they could keep scaling the number of layers. If not, then even a bump in the planar density wouldn't be enough for them to keep pace with NAND, in the long run.

IMO, there are just too many unknowns about the underlying technology and whether it really had the potential to scale much further. I assume they looked at the trajectories of the respective technologies and only saw 3D-Xpoint falling further behind. If it had been only a short-term setback, then I'd have expected they could at least find a buyer. Even now, they should be able to sell their patents and someone else could take it up, if it had legs. It makes zero sense for them to just sit on viable patents, now that they're not in the storage industry at all. That just underscores my feeling that the technology must have appeared unable to compete with 3D NAND, ever.
 
Last edited:
  • Like
Reactions: JarredWaltonGPU
Saw someone on the thread over at TPU regarding this news that they'd added the P41 to the Solidigm driver (double checked and they did this back in June) so perhaps there will be further development under SK Hynix branding.
 
  • Like
Reactions: bit_user
In my opinion, there are too many SSD brands out there. Many which don't sell that well unless they go aggressive on pricing. So while it is good for consumers, there is very weak business case for companies to justify fighting tooth and nail in this segment.
 
Saw someone on the thread over at TPU regarding this news that they'd added the P41 to the Solidigm driver (double checked and they did this back in June) so perhaps there will be further development under SK Hynix branding.
Do you have a link to TPU thread,
 
Hmmm, so then will my P44 Pro now go up or down in value? All other gen 4's keep going up in price which makes no sense. Paying more year after year, for old tech that is yet another year older...... It's just ludicrous! Ahhh but the kids love it 😁.
 
All other gen 4's keep going up in price which makes no sense. Paying more year after year, for old tech that is yet another year older...... It's just ludicrous! Ahhh but the kids love it 😁.
980 Pro, 2TB:
https://camelcamelcamel.com/product/B08RK2SR23

OTwq0nf.png
 
  • Like
Reactions: bit_user
Last I checked.... $169 is a much "bigger" number than $99. That is not 'down', that is up! Unless numbers work different where you're from?
 
Yep, I use the same site. And it's the same basic story across the board. We were in a good place not long ago, then stupid happened.
I'm not sure I'd call too much supply a "good place" though it certainly was benificial for buyers. It should have been very obvious this wasn't going to last because NAND manufacturers couldn't sell what they were making and paused/lowered production.
 
  • Like
Reactions: bit_user
I'm not sure I'd call too much supply a "good place" though it certainly was benificial for buyers. It should have been very obvious this wasn't going to last because NAND manufacturers couldn't sell what they were making and paused/lowered production.
What should be obvious is paying more over time for older tech. Production numbers can be easily manipulated, simple truths cannot.
 
What should be obvious is paying more over time for older tech. Production numbers can be easily manipulated, simple truths cannot.
Except that you're not "paying more over time" since the prices of every drive have dropped (many significantly) from where they started. The bottom of the market dropped out which dropped prices significantly and prices went back up when manufacturers got rid of excess inventory. This is an extremely straightforward example of supply and demand.
 
  • Like
Reactions: bit_user
Except that you're not "paying more over time" since the prices of every drive have dropped (many significantly) from where they started. The bottom of the market dropped out which dropped prices significantly and prices went back up when manufacturers got rid of excess inventory. This is an extremely straightforward example of supply and demand.
If you cannot read and understand the numbers right in front of you, then you are more than welcome to pitch a tent and camp out on moot point hill. I have no interest in nonsense. Logging out!