Western Digital to spin off flash business | Hacker News Reader

Western Digital to spin off flash business

theverge.com

87 pointsdm_me_dogs47 comments

So saddle all the legacy liabilities in with the legacy business and slowly suck out the equity, while the flash business grows? I can see how that's a win for investors at least. Are the flash products still going to be branded WD?

Yup. Seems like an increasingly common move by businesses. AMD did this with global foundries, for example.

Im not following the surprise or maybe incredulity of this sentiment. It's normal and healthy for businesses to shed poor performers and merge to create synergies. The food majors went through this in the 2010s and nobody was crying or implicating malfeasance. The market should be reacting, that's what makes it useful.

There's a difference between what is good for the public, what is good for the market, and even what's good for a business's long term outlook.

The issue with spinning off losers and acquiring winners is the prioritization of short-term gains over long term gains. Further, I dispute that it's motivated by "the market" and not the gambling habits of corporate leaders.

Take, for example, HP's purchase of Palm for 1 billion dollars. But then a few months later, they let go of basically the entire palm staff. Then they created their webos tablet, then they killed it just a few months after launch. Then they let webos linger until finally selling rights to it to LG of all places.

You certainly can argue "this is just the market being efficient" but was it? Or was it overpaid C level individuals making big bets, failing, and punishing their employees for their incompetence?

Some companies have done this recently (J&J, iirc), but the article didn't mention any liabilities WD was trying to shed, and both these businesses are viable, so I'm not sure if this was just a knee-jerk corporate criticism or there's something to this.

Spinning rust side is only living because of price collusion by the flash vendors. Price for 8TB of TLC generic is currently 170 USD. QLC is cheaper than that.

That's actually a valid reason to spin off a business, especially if you happen to know you'll cross the inflection point in ~5 years. It isn't the same as the "legacy liabilities" game, though. Some investors want the growth business, some want the stable business, some want a cheaply-priced declining business (seriously; look into anything that pays royalties).

Spinning rust is still useful for many purposes and will probably live on for decades more. We generate a ton of data and most of it ends up as read rarely or never. But it can live with slower access times on longer duration storage.

It’s honestly just like tape. There will be an HDD market for a long time that will be hidden in the whirl of data centers. Consumer HDD will slowly melt away, but the price/byte and durability of disk will keep it viable for quite a while. It’s still faster than tape (but not as cheap).

> Consumer HDD will slowly melt away

I’d say it already has. Does any consumer PC come with a spinning disk anymore?

Many consumer PCs will come with both a smaller capacity SSD and a higher-capacity HD.

Maybe 10 years ago, but I just checked Dell (for a single data point), and it looks like all their current laptops are SSD only, and for desktops, most configurations are SSD only, with only a handful of balls to wall storage configurations offering large multi-TB HDDs in addition to a SSD boot device.

With games weighing in at half a terabyte, they have to.

Nailed it. They’ve been spending goodwill like crazy with debacles like pushing out SMR junk under the Red name.

In my experience, their overall quality control has plummeted over the last few years. I’ve bought 3 Red drives for my home Synology. After about a year with each, the drive utilization graph will start creeping up toward a steady 100% while the other drives in the RAID are mostly idle. All drives have the same throughput. All drives have the same IOPS. And yet, the Red starts taking longer and longer to complete its requests. There aren’t any SMART errors or warnings. If I pull the drive and run WD’s diagnostics, it’s fine. It shows every sign of being perfectly healthy, other than being dog slow up until the day I replace it. I’m 3-for-3 with this happening. Maybe I’m the only person and they work great for everyone else. I won’t buy another, though.

I think we’re seeing WD turn into HP. People will still buy it for the name on the label, but those in the know will steer their friends to calmer waters.

I also had two Reds showing bad sectors after around a year, with a Synology. What "colour" drive do you usually buy now? Or even which manufacturer?

I've switched to only buying Toshiba, been very happy with them. They're also more silent as a bonus.

I've had good luck with Seagate Iron Wolves. They're a little noisier than others, but I've never had one fail. The Toshiba I've had in there for 3 years has also been fine.

After years of seeing many brands of drives go through good and bad years, I'm hesitant to recommend any particular manufacturer. Seagate had a rough stretch, but they seem to be pretty good right now. WD used to be great, and now I don't want anything to do with them. Anytime I look for a new drive, I check out the latest Backblaze report and hope for the best.

HP or HPe?

Basically the same segmentation of “good business” and “bad business” there too.

Which is which?

Are your drives WD Red, Red Plus, or Red Pro? Not that it should matter, just curious.

To add another anectode, I’ve been running a 12TB and a 6TB WD Red Pro (both CMR) in my Synology without issues (knock on wood) for 16 and 38 months respectively. I recently added an 18TB Seagate Exos and a 16TB WD Ultrastar, which are also both CMR and are intended for data centers (with twice the MTBF of Ironwolf or Red Pro). The enterprise drives are quite a bit louder, but since the NAS isn’t within earshot it’s not a problem.

The NAS runs 24/7 as a Home Assistant and Plex server, as well as handling backups from machines on my LAN. So far, so good, but I’m not kidding myself about eventual drive failure. I know it’ll happen at some point.

WDDA analytics are not turned on. Don’t need to see a pointless warning at the 3-year mark.

Feels like the HDD business is the more stable of the two. The competition in the flash/SSD market if fierce.

True, but it’s clearly the future, at least for mass market devices.

There’s still a place for buggy whip makers. The remaining ones can charge an arm and a leg because the few remaining ones generally compete on quality. I bet that’s what we’ll see for HDDs soon.

NAND is unable to replace HDD for long term storage.

Yes, and that's an important market. I suspect that it's tiny compared to the mass market devices I mentioned.

SanDisk's reputation is in the toilet though, thanks to things like this:

https://arstechnica.com/gadgets/2023/05/sandisk-extreme-ssds...

Western Digital's isn't any better though, due to their own mismanagement

https://arstechnica.com/gadgets/2020/05/western-digital-gets...

They need to come up with a good name for the flash business. Maybe something that indicates their products are like disks, but made from silicon instead of spinning rust. SiliconDisk? Maybe Sand-Disk? Or simplify that to Sandisk?

The answer is obvious. Eastern Digital.

Usually stock investors don't like diversification within companies. It is harder to value.

I think that they realized that there is not so much in common between the HDD and flash business. No added value to own the 2 business compared to separated entities.

With that, the WD legacy HDD business can continue as a dividend stock dying slowly. And the flash business can now pretend to crazy valuations based on a shinny future.

It might also free up the flash business to make aggressive acquisitions without worrying quite as much about anti-trust from having a larger position in the overall storage market due to the old HDD business (which is dominated by just three firms).

The stand-alone HDD business will end up as an acquisition in the not-distant future (Toshiba, Seagate, etc. somebody will eat it).

It would be such a great headline if they worked disk and/or disc into it somehow

> Western Digital spins Disk+Disc Businesses off

Or how it's a bad sector to be in.

Edit: I'm sure the editorial staff at The Register are brainstorming a headline at this very moment.

Something, something, disc-injury spin from disc/disk industry on worker injuries

> I'm sure the editorial staff at The Register are brainstorming a headline at this very moment.

They're "disc"-ussing it.

Curses ;)

They probably needed to prioritize getting the article out fast over coming up with good headlines since it’s a flash sale

I wish I could get paid, like as a job, to mint the headlines

Which one will get the RISC-V cores?

I will never waste money on a Sandisk SSD again. My last one stopped working out of nowhere. Much happier with my Crucial X10.

On the contrary, my Sandisk SSD I got around (roughly) 6 years ago is still happily working, despite being utilised for a lot of writing. It's gone through 3 house moves, and it's still going strong -- what a great purchase.

My 2 Crucial MX500 SSDs have been pretty solid too; nothing bad to report on them.

Re: your SSD, did you RMA it?

There is no brand that never produced a SSD with a fault. I had tried all major HDD/SSD brands, and I think every single of them failed at some point.

What is important is how they handle that. I had a faulty Sandisk and it was replaced with a working one without any problem or unnecessary question. Some other brands annoy you with stupid questions, puzzles and other tricks trying to avoid doing a replacement. And some just ignore any requests (ex. Samsung, from my experience) to replace a faulty SSD.

I've never had a problem with Intel, but they sold that business recently.

they had a whole series that would irretrievably die after 1700 idle hours due to a firmware bug.

This summer I bought two Intel 2TB nvme drives for $65 each. Not the newest tech but I've also never had a problem with an Intel SSD. I still have my original 40GB from 2012 in an old laptop and it works just like new.

The company that bought their SSD division (Solidigm) supports the Intel branded drives, and their utility had firmware updates ready to go for the ones I bought.

I'd probably buy a Solidigm if I needed a new drive and these firesale Intel drives are gone.

It's the Koreans - SK Hynix bought it and is using it as their consumer drive division. The P44 Pro is supposed to be a more reliable and cheaper competitor to the Samsung 990 pro...but that unit is also not doing so hot w/ the memory downturn/post-covid correction

> Not the newest tech but I've also never had a problem with an Intel SSD.

I used to run Intel (enterprise) SSDs in servers. I didn't run enough other SSDs to have a sense for whether the failure rate was abnormally high, but it was definitely not zero (it was significantly less than the enterprise hard drives we were running). And the failure mechanism was sudden disappearance from the bus.

My message to you is not to avoid Intel SSDs, just to make sure you have a backup in place, because when they fail, my experience is there's no warning, they just disappear. This seems to be the common failure mode for all/most SSDs, even though they're supposed to fail read-only. Hard drives usually give warnings as they fail, but not always enough warning to rescue data.

Lets not forget killing the drive instead of going read only, and later lying about this being intended outcome despite opposite statements in official documentation.

https://techreport.com/review/26523/the-ssd-endurance-experi...

"After a reboot, the SSD disappeared completely from the Intel software. It was still detected by the storage driver, but only as an inaccessible, 0GB SATA device.

According to Intel, this end-of-life behavior generally matches what’s supposed to happen. The write errors suggest the 335 Series had entered read-only mode. When the power is cycled in this state, a sort of self-destruct mechanism is triggered, rendering the drive unresponsive."

> and I think every single of them failed at some point.

I understand that you probably are alluding to the fact that you never scientifically measured this experiment and that this is anecdotal. But this is pretty loose verbiage when we're talking about a product that is guaranteed to fail at some point or another. These are a wear item.

Oh, yes, you're right, it's more like anecdotal experience. I didn't specify that, but I actually meant the time period when I don't expect them to break, which is about a year after the purchase. The device is still under warranty so I'm saying that some brands make it easy to replace it, some do the opposite.

Reply on news.ycombinator.com