Who wants Oxide Computer and Why?

Hey everyone. I'm Brian Cantre, co-founder and CTO of Oxide. Great to be with you this morning.

Thanks for joining us. Uh, it's, it's a fun space to be in because we're actually making real stuff, right? Um, and as, as Steve outlined, we took a real, in this case, literal clean sheet of paper.

This was not a napkin, but it was damn close. This is an actual sheet of paper. And when we, we started the company, um, we had been confined by the state of the industry, which had become commoditized, had become ossified with different companies all delivering these different layers of the stack.

And when you tried to get them to work together, you, the customer were often the one who was performing that integration. And you've got fingers pointing at one another. I mean, I, and I honestly, in many cases, you would have vendors blaming one another and the actual uncomfortable truth is they're often both, right?

They are often, actually the problem is not crisply it with one vendor or another. It's actually at that integration point and where it's like, actually your behavior is reasonable taken on its own. And you other vendor, your behavior is actually reasonable taken on its own if you need me as your customer to tell you that.

Um, but when I combine them, I have a pathological unusable system, and you are treating me like it's my fault. And, you know, you are the, uh, you're the only customer seeing this. Yeah.

And the number of times that, that, that, you know, if you are, if you've been a large customer, it's extremely frustrating to hear. And I, you will never hear that out of oxide. You'll hear that we have never seen this before, though.

We want to, because it feels like it's an excuse. You're the only customer seeing this feels like an excuse. It's not Investigate your problem.

We started with a queen sheet of paper and we knew that with a queen sheet of paper we were gonna be solving a bunch of problems. All of these problems had accrued over decades in this industry. Going back to going back literally to the dawn of the pc.

'cause we were trying to run personal computers in our data centers. We enterprise users, hyperscalers weren't doing that. Hyperscalers had long since figured out that this doesn't work.

They had designed their own machines, des their own machines fit for purpose, and they kind of, uh, befitting each of them. Um, they kind of acted with their own idiosyncrasies. Uh, Google, um, built their own and then wrote a book explaining how awesome it was.

The, the warehouse size computer. So you can read a book and appreciate Google's brilliance. You're like, do I get to buy one?

He's like, no, no, no, no, no. But look, but you can admire my great work here. You're like, okay, Google.

Yes, it's great. It's a, it, it's amazing. Terrific.

Um, you, uh, Amazon, you could, you could rent it, right? Amazon was very closed lipper about what they'd done. But, but you, you could rent it.

You actually, you could not buy it. Um, Facebook. Now meta, interestingly enough, kind of did the, uh, the most interesting motion where the technologists inside of Facebook said, you know, this is kind of not right.

That we, we developed this great innovation that the world doesn't actually have. And they started the open compute project as an, as an initiative to share those innovations. So you could actually go, as Steve and I did, to OCP summit, as we were contemplating oxide.

We went to OCP summit. And you look at these machines and like, this doesn't look anything like the machines that I have to run these commodity servers. This looks like, why don't I get to have these?

And it's really basic. It starts, it starts, but certainly does not end with the physicality of the system. The first thing you notice on these systems is they've got a DC busbar based design.

So you've got a power shelf that rectifies from AC to dc and you run direct current DC up and down the spine of the machine. And those sleds blind made in to that direct current means there's no AC power supply per one U2 U. And once you see this, you can't unsee it.

Mm-hmm. Once you are aware of this and aware that literally every hyperscaler, Microsoft meta, Google, Amazon all run with a DC busbar, BA based design. This is not like new thinking.

This comes outta the telco world. This happened for a long time. Anyone running at scale runs with a DC based busbar.

Not the, not enterprise computing. Enterprise computing has got AC power supplies per one U per two u. It's cords everywhere.

It's fans everywhere. It's heat everywhere. It's inefficiency.

And that is the beginning. That is the most visceral thing we were seeking to disrupt. But we knew we were gonna be.

When you take a clean sheet of paper, when you take this thing, there are a whole bunch of things that we knew we would be able to go do. Yes. Be able to do what the hyperscalers had done.

We also knew we were gonna be able to leapfrog, leapfrog the hyperscalers and develop yet new differentiators. Give you a very concrete example of that. We assumed you looked at the open compute project.

The, you looked like that Tioga pass, circa 2019 DC power in the back cable networking cables out the front. We assume we were gonna do the same thing. You look at those a, a picture from, you know, Google, when they're flexing on you, they're providing you that, those, those pictures of the dcs and it's all like lovely pastel colored cables, but it's cabling coming out the front.

That's what we assumed we were gonna go do. But we, we have the luxury of a clean sheet of paper with that clean sheet of paper. We actually had more latitude.

And we were actually talking to a connectivity vendor, and they're like, outta curiosity, why are you guys doing the cabling at the front? We're like, because we ha because everyone does it that way. I mean, the hyperscalers do it that way.

It's like, yeah, the hyperscalers do it that way, but not if they were taking a clean sheet of paper today, they wouldn't, what would they do? They would blind mate into a cable, back plane, cable, the back plane blind meet the network connector in. Like, can, can we do that?

Like, oh, absolutely, you can do that. Why don't they do it? Well, they're afraid to do it.

And that, that of course is like catnip, like, oh, they're afraid to do it. Okay, well, you've gotta go do that then. That's great.

That cable back plane was a bet. The company moment, it has removed the cabling from the system. You take a sled, there is no cabling in that sled like that, that cable, the, the, the back plane is cabled.

That sled goes in, locks in no cabling, because where there's cabling, there's mis cabling. That was, again, it was the beginning. Another bet the company moment.

There is no bias in the system. BIOS basic input output system. This is an idea.

All caps must be screamed. This comes from C-P-M-C-P-M predates dos. Dos is from my childhood.

As my children are quick to remind me, I'm old. This has been around for a really, really long time. The bias is this lowest layer of platform enablement that initializes the computer before the computer, before the actual operating system runs everybody on X 86.

And sadly, many people in arm run with a bias with A-U-E-F-I bias. We knew that that thing was a huge problem. Wanted to rip it out, wanted to do our own low level platform enablement.

This was something that a MD thought to be impossible. Why? Because Google had tried and failed.

If Google has tried and failed, it must be technologically impossible. And it was really, really hard. And it was another bet the company moment, we got it working.

We, we don't have a bias in our system. We, we boot holistically. First instruction is our instruction.

We do all that low level platform enablement, initialization, and bring up the entire stack in front of it on top of it, including all the distributed system, the control plane on top of it. There are a lot of these. When we had a queen sheet of paper, we had a lot of opportunity to do things that people hadn't done for, for each of the things that we did was not hard to find technologists to say, yeah, of course this is the way you would do it if you got a queen sheet of paper, but you never get a queen sheet of paper.

We got a clean sheet of paper. We didn't wanna spoil it. What this meant is it's a little hard to be crisp about the differentiator of oxide.

What is the differentiator of oxide? It's like, well, it's complicated. And we had a board member, the, the great pure Lamont, you don't know Pierre.

Pierre is a truly singular individual. Pierre was at Fairchild back in the day. Uh, Pierre hired Andy Grove at Fairchild.

So yes, Pierre just celebrated his 95th birthday. Um, we are the, the last company that, that Pierre helped found Pierre was on our, our, our founding board. I, I like, I would like to say that we are the last of the fair children.

If you know this idea in the, in the, in the sixties and seventies of companies that were spun out of Fairchild, um, oxide is the last of the Fairchild. Pure is great. Peer is infamously tough.

And p says to me, we had many direct conversations. Peer says to me, what is the differentiator for oxide? Like, well, it's complicated.

No, that's actually what I wanted to say in an earlier and maybe the first or second board meeting, Pierre had banned my use of the word complicated, didn't wanna hear. It's complicated. So I had to use all these euphemisms for like, well, it's, it's nuanced.

It's, it's multifaceted. I'm like, well, Pierre, no, this is not what you want to hear. But there's not one differentiator.

There are a bunch of different differentiators, and it depends on what your biggest problem is. We, we have absolutely differentiated on power and efficiency. We've differentiated on reliability, on operability for sure.

We've differentiated, just as Steve was talking about, about the, the time to deployment. Oh, that was a load star for us. We want that rack to arrive.

We want it to be come outta the crate, the importantly sled ship in the crate. We want that thing to be spun up and have developers on that thing at in hours, not days, weeks, months. So there are many different differentiators, and that's what you get when you get a queen sheet of paper.

That's the good news, right? Well, okay, well, what vertical are you targeting? It's complicated.

It's a general purpose machine. It's, there are a bunch of different verticals for whom this is really valuable. So part of what we honestly did not know when we started the company, we knew we were building it the way we wanted it to be built.

We had experience deploying at scale, at joint, and then very much at Samsung. We knew these were pain points that existed, but you don't necessarily know the specifics of like, who's gonna grab whom and when you, it took us four years to build this thing. Three and a half, uh, it took us.

So, and actually I, I really should stick to three and a half and not four because, so we, we took our first capital in in 2019. We shipped in in June of July. June is when we shipped.

July is when it arrived, uh, of, of 2022. Of 2023, excuse me. Um, so it, it, it, it took about three and a half years.

The reason that three and a half years is important is because when we first started the company, we said it was gonna take two to three years to build. And Pierre, quite famously, when we were originally pitching him, said, which is it? Two or three?

And I'm like, eyeball the eyeball with him. And I'm like, two to three. He's like, which is it?

He's like, okay, this is like, anyway, love Pierre. It's great. Pierre asks very direct questions and gets ambiguous answers from me and manages to not, you know, manages to still stay in the room, which is great.

So it took us three and a half years to build when we, when it takes you something three and a half years to build the world can change a lot from underneath you. And we knew that the world that on-prem compute was not gonna go away. We felt very confident that the demand for running elastic infrastructure on-prem, we knew that that was not gonna go away.

We knew that was gonna grow over time. And that has grown over time. It's a little bit amusing to me when venture capitalists that passed on us.

Send me these like, public reports showing the growth of on-prem usage. Like, why are you sending this to me? I should be sending this to you.

Like you passed on us because you didn't think this market existed. Uh, uh, we knew the market existed, absolutely knew the market existed and absolutely knew that there would be increase. You know, there was a question earlier in Steve's session about, about sovereignty, data sovereignty.

We, we knew that that trend was gonna continue. We knew that there were, and especially knew that there were economic drivers. And the economic driver is still like the driver.

I think, I mean, if you were to really, if Pier were to really get me down, like what is the differentiator? Ultimately it has to be around doing more with less. But there's a lot of different ways in which that happens.

So we didn't know what was gonna happen necessarily. We knew that there was gonna be demand for this, but it's like, you just don't know. You just don't know what's gonna happen.

Uh, when we first started the company, one of the earliest advisors of the company is like, you've gotta run ESX on this thing. It's like bad news. That's not gonna happen.

We're not running ESX on this thing. Let us re-explain to you what we're doing. We are doing our own hypervisor, our own control plane that our own distributed system to provision the primitives of our system to pro provision compute, storage, networking.

So we are not gonna run ESX on this thing. And he's like, I think you're making a huge mistake. Well, needless say couple years down the road, he's like, you guys are geniuses.

And like you were so like, did you know this was gonna happen? Like, well, no, not exactly. Didn't really predict the, the acquisition of of VMware by Avago.

I insist on calling them Avago because Avago has taken on the skin of Broadcom to give themselves like legitimacy, sorry, this is private equity. The ticker is still avago, I'm gonna call them Avago. So they're just a, uh, point or principle I guess.

But we, no, we did not predict that. Um, we did not. But with that happening, the amount, the, the customer pain is definitely predictable.

And we've gotten a lot of people who are all of a sudden newly interested in oxide because of that. But we, we, but we didn't know exactly what was gonna unfold. We also knew that there would probably something just like we take a clean sheet of paper, you know, that there's probably something that you're not considering, like that cable back plane, an opportunity that you're gonna have that no one else is gonna have because you took a clean sheet of paper.

And so we know that like, there's probably gonna be a market that we can't really predict right now. That's gonna be a huge tailwind. If you look at successful companies, they have this, I mean, it's famously with Nutanix, this is a VDI, right?

I was at, I was at Sun back in the day. Um, I was at Sun for kind of exactly the, the latter half of its life was there from 96. And I went down with the ship in, in 2010.

Um, but the, if you look at Sun in the eighties, EDA right? EDA was explosive for those workstation companies. So there, there was a, a market that they didn't necessarily predict, but it was a huge tailwind.

So we thought there might be one of those markets, but I'll be candid with you that I'm surprised a little bit about what that market is and the differentiator that they really care about. So to get, to give you a little bit of context for this, uh, when we were first announcing the, the series A, uh, we were working, you know, with a PR firm as one does, and, and they're talking to reporters as one does, they're like, this really needs to be an AI company. It's like, what do you mean this needs to be an AI company?

Come on. And they're like, well you, what, what is your AI story? Like, what is it?

We're not our a we're we're making this general purpose computer and AI is obviously important and we, we saw that, but that's like being, it's the accelerators that are targeting that clearly Nvidia and then a bunch of of other folks. It's like, well, what's your accelerator story? We get this a lot.

What are you guys doing about GPU? Okay, what are we doing about GPU? Well, the GPU, as I tell people, there are two doors.

One is labeled partner with Nvidia, and the other is labeled compete with Nvidia and what they have in common. Those doorknobs are both hot. There's a, there's an inferno blazing on the other side of both of those doors.

Those are both very complicated doors to go through. And we've been like, you know what? We've got enough to go do on general purpose compute, general purpose compute, networking, storage.

We absolutely are keeping an eye on what we wanted to do with an accelerator based product. But we, it's, it's a bit of a gory market and NVIDIA's executing well and so on. So it's like not, not really rushing to go partner with Nvidia or compete with Nvidia.

We're, we're happy to kind of be off to the side. Alright, well, okay. They, you know, they kind of throw up their hands like, well they can't get a reporter saying it's not ii alright, fine.

But then some of the big AI companies came knocking. It's like, well, okay, what's this about? And as it turns out, and again, this is like, you know, Steve was describing having the national labs having kicked our door down.

Sometimes, you know, it's a little bit thick for, we're a little bit thick. I'm a little bit thick to under, we try to explain to these, you know, large AI companies that like No, no, we're, we don't have an accelerator. Yeah, yeah, yeah.

We know that these, you think these models are, are all GPUs? Well, I guess not. Uh, so if you go to you, you go to your LLM of Choice chat, GPT, what have you, you see searching the web when you type something in, you think A GPU is searching the web?

No, that's A-C-P-U-U-C-P-U is searching the web. Mm-hmm. You think that when they're training their next gen model, someone's gotta go out and crawl the web, pull that in.

Think the GPU is out there crawling the web. No, you think A GPU is doing all the processing it needs to, to, to do kind of traditional map produce, like data pre-training on that? No.

C you're like, oh, right C that's what we make. They're like, yes, yes. That's what you make.

And by the way, when we look at, we, we companies with a lot of compute and the large model companies have a lot of compute. Like, I've got some bad choices right now. I'm in the public cloud and it's killing me.

Like I, what I, I, this is economically ruinous to be in the public cloud, so I know I'm gonna need to go on-prem, I'm gonna need to own my own compute. Especially when you look in the even not so indefinite future, I need to own my own compute. Okay.

So that was, I I mean, again, I'm a little embarrassed to be like, oh, right, okay. Yeah, I guess that makes sense. And you realize that like all of the trends, I mean, probably was true two years ago that there was much less of a CPU component.

But if you look at all of the trends towards those large models around test time, compute around, you know, you, whenever you are using these models, and I love it when they write a program for you and you can have them show their work and they're like writing a program for you. That program is ex all that is happening on, on A CPU. There's a lot of CPU involved.

Yeah. So kind of bouncing off of that, um, you, you guys have talked a lot about energy efficiency. Yeah.

If you are doing something with AI and you have lots of GPUs, that's gonna be, you know, the, the majority of your, uh, energy usage, um, how much does, and if, if you're using your platform to uh, you know, do these other things that you're talking about. Yeah. Um, how much does that factor in?

Or does it just kind of get overwhelmed by um, you know, how much compute is? Yeah, I mean I think that, and again, it kind of depends on who you're talking to. If you're talking to like an enterprise user that's kind of pulling in some GPUs that's already power constrained.

Our ability, and this was actually when we, we were kind of talking about this in, in kinda the 20 22, 20 23 time. The, the reason they were interested is like, Hey, I can actually go take my traditional rack of compute and I can now get a lot more bang for the buck. I mean, in terms of performance per watt, we are the, the, we are, uh, definitely a, a step function compared to what one would run from a kind of Dell HP super micro.

So the, for those folks yes, absolutely. For the large model folks, I think that they are, I mean, their power right is, I, I guess my question is like, does it become substantial enough that it's like, Hey, if you switch over to us for the CPU center, right? Can you fit in X amount more gpu?

For sure. And, and, and if you were to tell me, Hey, with no other data, hey, the, the, the large model companies are really interested in oxide, I would have that same, I'd be, oh, it must be because of the power. It's like, well, sure, like the power.

That's actually not what they're most interested in. And what they're most interested in did surprise me. What they're most interested in is something else that we took a clean sheet of paper on, which is the security of the system.

So we put a true root of trust in the system, really took a, a, a page from what, um, what the iPhone did with its root of trust, with the, what you see in like the Chromebook model took a, put a true root of trust with secure silicon, with the idea of being able to attest the entire stack of software. So the, in the kind of the PC world, part of the problem is that you've got this kind of verified boot in the bias, but it kind of ends there. And I actually wanna run, I need to run an entire control plane on top of this.

There's a whole SBO I've got associated with that. I need to attest this, this whole layer of, of software. And then I wanna pull that in to my own control plane, my own being a user of oxide.

So I know what I'm running on. I've got confidence in what I'm running on. And that is overwhelmingly what has drawn them to oxide.

And you know, at one point, because I, I felt like, you know, it just felt like, okay, I can see why they're interested dioxide, but you know, how, how differentiated are we really gonna be? And on power, we're certainly differentiated on power observability. We are extremely differentiated, namely, like we can, we can tell you how much power is going to your fans, for example, the oxide rack, very little.

But we can demonstrate that. Um, you don't get that outta traditional systems. In traditional systems.

You actually don't get that kind of granularity. But that's actually not what they cared about. What they cared about was that that ability to do attest station of the entire stack.

And what they're very concerned about, because this is like some of the most important stuff we've developed in terms of these models, they're naturally very concerned about very sophisticated attacks about supply chain style attacks. You may remember the Bloomberg did a piece in 2018 on the implant at Supermicro. It's a very controversial story because the implant itself has never been produced.

Uh, they, they wrote a story in 2021 reiterating their belief that the implant exists. Uh, they made the somewhat dubious decision when they ran this story in 2018 to put a picture of a, a, a finger holding a piece of electronics. I think it's like a 6 0 4 resistor, honestly.

But the, but they wanted to do that to kind of make it more realistic. It's like you've actually made it less realistic. 'cause no one knows what this implant actually is.

But it stoked a lot of concerns about what would an implant look like? What would it look like to infiltrate the supply chain? Especially if you look at the traditional BMC, the baseboard management controller, which ultimately also eliminated, replaced that with our own service processor and own software.

That baseboard management controller is a completely proprietary piece of silicon with proprietary software and coming from a source that you don't control. And if, if the BMC gets rooted, you control heaven and earth on the machine. So the kinds of questions that they were asking us that these large model companies are asking us, you look through their diligence and you're like, oh, you're asking us questions about where our sources of non-volatile memory on the actual board.

'cause you're concerned about an attacker being able to hide out in some of the non-volatile memory on the board. Well, we can answer that question really quickly because we designed the board, we designed it from first principles. We can tell you exactly where all this stuff is.

And when you're like, if this is what you care about, oxide is gonna be extremely differentiated. Because I can tell you as a Dell customer, an HP customer, a super micro customer, it's literally impossible to get answers to those questions. They speaking for Dell, we were a big Dell customer.

Like they actually don't know. They actually don't know where all the source is. Not volatility are, yes.

Oh, Go ahead. Yes. No, I, okay, so I'm, I'm gonna come from a business standpoint because Sure.

A lot of what you're saying is way over my head in terms of not as smart as you. That's for sure. Um, you talk about that there are some very specific value propositions that you bring in as a wide variety, and I can understand that.

Do you have, can you talk about d you know, a rack that you, comparable rack that you would see from Dell or somebody else versus what you have? That would be one in terms of a cost efficiency. Yeah.

All those efficiency issues. Second piece to that is, I hear what you're saying on a, the value proposition of what you're bringing in. I am sure the numbers are big, very big differentiation between Dell RAC and urac.

My fear would be a single vendor. My fear would be I am, I'm getting back to the mini computer, the mainframe computer, et cetera. Sure.

Yeah. That I am, I'm tied to you in that situation. And that's dang scary.

Although you're tied to a vendor no matter what. There We go. You're answering your own Question.

I I am, I'm answering that question, but so I need to go beyond, so I need to get beyond, it can't be just a 30% value proposition. It has to be bigger than that. I mean, it depends on who you're talking to and why, but Yes.

Yeah. But bigger than that, if I'm looking at how do I do the sovereign, if I have a sovereign issue, if I have a, an energy issue is a completely different, can be a completely different one. So I'll stop there, but rattling and you can answer.

Yeah. I mean, so a couple of things. One, um, you're absolutely right that we ultimately, we all put faith in single providers, right?

And that has risks. Yeah. And so, so give me, I I understand that.

So, so in order to get beyond that risk, How do you mitigate those risks? The number has to be significant. It's kind of like clients that we, I had for, you were talking about this morning, clients we had over in Europe buying Huawei.

And the reason why they bought Huawei is because Huawei came in and was over 30% different cost. Right. Selling to them.

Otherwise I would never have looked at, they would never have been able to buy them. Yeah. And I can assure you that Huawei, that that a 30% delta end price would not make Huawei attractive to, uh, large provider.

I do understand that. I mean, so Understand where I'm going with this. Yeah, Yeah.

I mean, so I think that there are, um, a couple of things. One, in terms of of totally get the, the, the single vendor risk. Um, and you kind of get the, you know, it's like, God, this is, are we, are we reinventing many computers or we reinventing mainframes?

And we're mindful of that. Um, I think that a big difference is that we live in an, in an open source era. And so we get this really crisp, clean, open source revenue model, namely, like unlike these companies have to go, you know, relicense themselves to figure out how to make money.

We make money 'cause you buy the thing. So what that allows us to do is be very transparent. So we uniquely have opened up all of our software.

So if you are running Dell, HPE Super Micro, you're actually beholden to a single vendor called American Mega Trends. A-M-I-A-M-I is actually the vendor that's doing the proprietary layer that is common to all of those systems. There's only one company that is gonna give you a machine.

There's only one way to boot an X 86 box without having a MI involved. And that's to run with oxide. We give you all of that source code U uniquely, uniquely, a AMI ISS not open source.

That bias is very proprietary. Is that an OCP project or be related in any way? AM is ours No OCP project or is the a MI is definitely not an OCP project.

No, no, No. I know they're open. No, um, it, it's open source, but we, I we don't, we haven't made it a, a, a formal OCP project in part because like we are open sourcing it for slightly different reasons.

Well, I'm just wondering if it's part of a broader community or if it's an oxide community open source. Um, I, I mean, we have open sourced our entire stack, and at any given layer you will find its own community. Okay.

So one of the things that we did, we opened, sourced, you know, we, uh, we had the, the, uh, we needed our own operating system. We needed an operating system for the microcontroller that we were gonna use to replace the BMC with a service process. We wanted to use existing open source things that were out there.

Um, but ultimately we couldn't find one that fit our goals. So we fittingly named the operating system Hubris. So Hubris is our operating system that runs on the microcontroller.

The debugger, by the way, for hubris is called humility. Um, and those are all open source. Those are being used in lots of interesting places.

So the, yeah, Okay. It's just, it's, it's a question of there's open source in community development. There's open sources in transparency mm-hmm.

CB and, and I would say that we are, the priority is open source for transparency, but because we solve problems that people are really interested to have solved, we've had a lot of open source as community follow on from that. And, and it's not, on the one hand, it doesn't completely mitigate single vendor risk, but it gives people confidence that like, okay, I can actually see this whole thing, which again, I can't get out of anybody else. Nobody else allows me that transparency.

And this is at the absolute brainstem of the system and it's of, uh, so it's not a 30% kind of a thing. It is a, like, this is an existential risk versus this is a risk I can manage kind of a thing. So Yes.

It seems like the, the competitive for you is not the traditional OEMs, but it's more the OCP community that is building systems based. You would think. You would think.

Um, it is. So, and, and, and that's actually kinda what we would have thought for sure. When Steve and I went together to OCP summit, again before we started the company, we're like, let's go buy one of these racks.

And you go to like, okay, who do I need to to go to buy one of these racks? And they're like, sorry, which, which part of Facebook are you in? Which, which group am I talking to?

Because if you wanna build an OCP, there's not a thing you buy. You gotta go ODM direct. And if you've gotta go O-D-O-D-M direct, like your numbers need to be huge because now you are doing your own oxide.

And so there's not a, and we would bump into other enterprise buyers who are kinda like, oh, I wanna go check out this OCB thing. You're like, wait a minute, this is not, I, there's nothing I can buy here. There's, because I've gotta go like, sure, I can go talk to Quanta.

I can go like, and I can, you know, if you wanna go talk to Quanta about the kind of numbers that you need for Quanta to actually build a box with your specification. But there, there are not board designs coming outta OCP. They're rack designs but not board.

So the, you might think that, but it's actually much more of a OCP is a is great love OCP, but it is, it is by no means products. It is hyperscalers collaborating on how they do their vendor management. So, so I get the barriers to entry for manufacturers on that side.

Um, I'm also very interested in the barriers to consumption from the customer side. Yeah. Sounds like you guys are going after enterprise, not necessarily MSPs, which would seem Not, not MSPs.

Yeah. Right. Which would seem to me to be a natural way to go, but I won't get into that right now.

But, um, but especially with the move from public clouds to something less public. Um, but the entry point for your product seems like Iraq. Iraq, right?

Yep. So, um, and I run into this with some other types of vendors as well. How, you know, like if I'm an enterprise, getting into your product involves, uh, pretty much buying into the management model.

I have to learn the APIs, I have to integrate, I have to learn your automation. Now you may, if I'm already a cloud savvy organization, I know a lot of this stuff and probably very little, I'm gonna imagine there's gonna be very little to to, to learn, which is great. Um, but I still have to buy into your, your management plan, your management Absolutely.

Your constructs, right. Which is, that's an not only an investment, but I'm putting a ring on it, right? Yep.

So, um, if I'm in, I gotta be in mm-hmm. And if I'm, and so, 'cause I'm not going to, the, the idea is I'm not just gonna buy one rack. I'm gonna go all in eventually, and I'm gonna have racks either Yep.

All in there or everywhere. And you're very happy if that happens. Right?

So, um, so if I'm an enterprise that's, it's like you're at, you know, if I'm first getting to the cycle, that level of commitment upfront is hard to uh, yes. It's, it is hard to, uh, it's hard to obtain. So, yep.

And of course you're not just gonna throw in, maybe you do at this point in your world, but, you know, I wanna start, you gotta throw in a rack for at a, at a large customer, maybe you would invest in that, but not for Nope. At scale you can't uhuh. So I see on the website you've got a trial.

And so that trial is that you get to, we, we've got a, we've got a rack in a co-location racks in, in a co-location facility. And we've designed multi-tenancy in to the, the, the very core of the product, which we didn't do for this reason, but it ended up being a huge benefit, is that we can now spin you, you wanna try the oxide rack? Great.

We'll spin you up as a tenant on this colo rack. Right. You can hop in there and, and you, you can try it out.

Yeah. And I'd be great. I didn't, I I Glen deck from Equinix actually.

Yeah, There you go. Yeah. I don't know if you would Well, so, and your question is so great that we've got an entire presentation right now that's gonna address exactly this.

So my, my colleague Matthew Sanabria is working on exactly the problem you're talking about because this is a very important problem Yeah. About how do you integrate in to the environment that people have. And this honestly would've been much harder a decade ago, but the, and this is, you know, you can correct me if you disagree, but I think the, the momentum behind Kubernetes in 2015 was actually around people getting optionality around their cloud vendors.

Just to the point earlier about having a single vendor, like people felt very trapped by AWS in 2015 Kubernetes people lung at, 'cause it's like, this is gonna be my multi-cloud story. Mm-hmm. And so that has given us, now we are now a full decade after that.

We've got a lot of software substrate that we can go integrate in that makes it super easy for customers to use our product.