Z PLUS DEADPOOL!!!
ZizzyDizzyMC
Hosting Provider
Just sent this out to all existing customers, so posting here for anyone else:
Hello,
First off, apologies for the inconvenience of closing up the (public) shop. This is a pretty tough decision but I'll explain below why.
TL:DR: I'm closing up all public sales and existing services will run out at the end of march, payments will be suspended - no further payments by the end of next week.
TL:DR Why: Ram prices. NVMe Prices. HDD Prices.
And if you'd like a more reasonable explanation below:
Z Plus LLC started off hosting individuals and projects relating to MLP, and our infrastructure got large enough that we had a lot of spare capacity to handle hundreds of clients that had nothing to do with ponies. So we went public, and because I liked low end deals and made a lot of friends in the market, that's what we did. Margins are small / non existent. Prices are cheap because the infrastructure is already there - we're just selling excess capacity to make a bit back on mostly hardware that would already be running.
We got semi popular, I've had a couple hundred clients thoughout the past few years, dealing with issues and fixing things up. Uptime's been pretty good since I sunk a new-car's worth of cash into rebuilding our infrastructure. And that's great, and then came along the Sloperators. Who absolutely NEED AI.
The real reason we're closing isn't because I'm tired of dealing with support tickets, or because we're not making money off of this side of the business (we aren't I assure you). It's because the entirety of everyone getting this message is running on a VERY nice AMD Ryzen 9900x server with 192GB of DDR5 ram. Why's that you ask? If this server has a bad ram stick it legitimately makes more financial sense to power it off, than replace it.
There are not enough clients to cover the cost of 1 48GB ddr ram stick in 6 months. Don't get me started on 4TB NVMes. Corpos have started slorping up consumer drives too, and with crucial out of the pool and skhynix out of the pool it's gotten worse.
My love for making good deals, upholding promises, and providing what people pay for far outweigh the downsides of tickets and potential headaches. So Z Plus will come back, eventually. But for now, if something went wrong I could not reasonably afford repairs, and it's not fair to you, the client.
Some of you have considerable amounts of data stored, if you need extensions on getting data out of our network please reach out to me and I'll keep the lights on to make sure you're set. If you are a client hosting open source software, or fandom projects please reach out - Z Plus will continue to provide support and VPSs in support of projects at no cost. If you are a client using Z Plus IP Transport and you've got this email, you can disregard the email as there are no changes to our IP Transport.
Also, if you've read all the way though, thank you. I appreciate your patronage and regret the situation that the world is in at the moment.
I'll be slowly disabiling the payment methods and no new orders will be processed, once done everyone's servers as is right now should continue running, without payment, till we shut down that section of the infrastructure.
The billing instance will remain up, as this is used in support of all of the projects we host for free as part of our ongoing open source and fandom support initiative.
Despite the grim news, we're not actually deadpooling. We own all of the hardware and are under no duress or hurry. I legitimately can't justify the expense of repairing the infrastructure if something goes wrong, basic fixes like a single bad ram stick are more money than Z Plus takes in as revenue within 6 months. This was tolerable when a ram stick could be covered in 3 months of service.
It wouldn't be right by the consumer, client, for me to continue operating in such an unsustainable fashion. Before the AI price spikes were were operating pretty much net neutral, and I did it for the love of the game really. I enjoy the process, and love the community. I'd hate for something to go wrong and THEN be like "Oops all your apes gone" in regards to people's data (VM or storage wise).
I'll be back with better offers sometime after the market figures itself out, perhaps I might end up with some of those Sloperator machines after the bubble pops... No sane data center on earth will let you colo one with them after that bubble bursts.
Regardless, thanks for being with me though this, and apologies for the bad news.
Comments
Understandble. My best wishes @ZizzyDizzyMC
Thanks for the clear words.
I believe in good luck. Harder that I work ,luckier i get.
We stop the bleeding so we may yet survive
blog | exploring visually |
When are the rescue boats hosts coming?
If you want information, feign ignorance reply with the wrong answer. Internet people will correct you ASAP!
It’s OK if you disagree with me. I can’t force you to be right!
A deadpool notice posted by the provider themself, that’s a first.
We are known to be "dangerous". affbrr
Zis iz zad to hear :-\
☰ Black Friday Storage — AMD EPYC VDS (ref), up to 4TB NVMe ／ NVMe + big HDD VPS (ref) from $2.29/TB/mo ／ Storage/GPU dedis (ref)
If you want information, feign ignorance reply with the wrong answer. Internet people will correct you ASAP!
It’s OK if you disagree with me. I can’t force you to be right!
Thanks, it was fun while it lasted!
Out of curiosity, any stats on faulty RAM? (Or fault rates in general?)
Basically this, I'd rather remain solvent so my clients can have the best service possible. We're not in dire straits, but we are at a crossroad where major changes needed to happen and temporary (expect at least a year or two) closure makes more sense than running fast and loose with people's data.
If luck holds true, ram prices will correct in 2 months to spite me which means I saved all of your gaming rigs /s
On average I'd say we lost a stick of ram in our fleet of 10 machines about once or twice a year. Buying enterprise ram and not overclocking leads to really low failure rates. It's the statistical average of DDR5 failure rates in particular and the lack of availability for a matched stick that made continuing service without replacements on hand extremely risky. I've had more friends with dead DDR5 sticks than I've had with ddr2-3-4 combined, and that's just the past few years. Coincidence? Maybe? I feel like ram manufacturers have been pushing their DDR5 chips a lot harder than they ever pushed ddr4 chips out of the box, leading to premature failures. I actually explicitly lowered the frequency on our box when I built it knowing that the ram (at the time) was still around $700 but now that same ram is approaching $3k
Honestly it's best to hear it from the horses mouth rather than the pony's ass. It's not a true deadpool because I didn't disappear overnight and run off with $ of yearly boxes and then blame a random "partner" in romania. I'm also not popping up another random host in a month either. No one will get adverts to their registered emails, and no spam calls.
We'll actually continue to operate the "Free" side of Z Plus, I have a patreon and subscribe star that people donate to, that pays for the "free" services. I'm OK with doing free services, for open source and fandom projects because it's free. There is no obligation to that, but with paid services I expect to be able to deliver consistent service and equipment prices put a serious burden on being able to do so in case of failure.
I love that the MLP provider is somehow more responsible and ethical than half the hosting market.
Sorry to hear this, can't have been easy to reach this decision, best of luck for the future.
TierHive - Hourly VPS - NAT Native - /24 per customer - Lab in the cloud - Free to try.
FREE tokens when you sign up, try before you buy. | Join us on Reddit
What's MLP?
We're the source, no cap. Address us: We/Our/Ours.
https://lowendspirit.com/discussion/comment/221016/#Comment_221016
Sorry to hear that man. Fuck AI. Fuck all those big corporations behind this AI boom.
Don't.
My pronouns are like/subscribe.
Be the example / change you want to see.
This is an excellent reminder for hosts to own your equipment because if you lease - your lease providers are definitely going to be feeling the pinch as dead equipment piles up. They can only absorb so much before they too either fold or raise prices by 50-150% for your leased units.
Lost it at this one.
It wasn't, I was considering sending out a notice about not being able to repair in case of disaster... I determined that really wasn't the legacy I was looking for and it'd be a hard thing to come back from. Best to stop public (paid) services on a good note and continue providing free services as usual. People who get free service typically don't mind the potential for 5 8's uptime when it's $0 to them and it's funded by public donors.
I did however do something with what I offered. My offers showed that there was potential in low bandwidth high storage vps's, and I'm happy that larger hosts like host_c had picked up and offered such plans. So even if it's not from me, others can still take advantage of a niche (paid) service that fits their needs.
I asked our vendor add more ram to the UAT server few days ago. They told me, hell no, you need to pay it yourself.
Action and Reaction in history
What's a UAT server?
Despite the grim news, based on my knowledge and experience, you’re likely the first provider to publicly announce a withdrawal from the market for these reasons — and it’s highly unlikely you’ll be the last.
I genuinely appreciate your transparency and your reasoning. As a provider myself, I’m dealing with the exact same reality: today’s parts market is a complete mess. We’re talking real, sustained +200% price increases on critical components, not temporary fluctuations, or at least, not something that will blow away in 2 moths.
As a remember:
If anyone needs a reminder of how long these distortions can last, look at Chia (XCH). When it launched in 2021 with its “proof of space and time” model, it triggered a massive surge in demand for high-capacity HDDs and SSDs. That drove shortages and sharp price increases across multiple markets. While the peak impact was in mid-2021 and prices eventually normalized later that year, the disruption itself lasted well over a year — long enough to seriously affect anyone relying on storage pricing stability.
GPU mining locked up consumer and enterprise GPUs for years, not months — pricing never truly returned to pre-mining levels, and availability remained artificial long after the initial hype should have died down.
I’ve said this before on OGF and I’ll say it again here: this is only the beginning of a long and painful road for many small- to mid-sized providers. And no — even larger providers won’t be immune forever, depending on how long this drags on.
What I respect most is that you chose to act before things implode, not after. That’s the responsible move. You limited the blast radius, protected customer data, and avoided making promises you couldn’t realistically keep. That alone deserves respect. Personally — you get a hug from me for that. This is the right way to handle things given the current circumstances. 👍
People often say “just raise prices” or ": you should have accounted for this at the time of offering" but that argument ignores reality. You can’t reasonably forecast or market around a 200% increase affecting multiple core components at once, especially when those components are fundamental to the service:
That’s not a tweak — that’s a structural break.
To the rest of the community: stock up on popcorn and inventory. This won’t be the first announcement like this, and it definitely won’t be the last, unfortunately.
Host-C | Storage by Design | AS211462
“If it can’t guarantee behavior under load, it doesn’t belong in production.”