Innovation in the enterprise SSD market driven by Phison

I have a difficult decision this morning. I was trying to see what I should wear for this special occasion because, uh, am I a tech geek or am I a business person? So this is my result.

It's a t-shirt with a suit, right? So this is how I started my days at F Us. I'm the president of F us.

This is how I started with the F journey in the us. I started as a FAE, I go into the lab. Yes, I saw the stuff, making sure it works, run the performance, give it to the customer.

And then while I'm waiting for the product to run, I went over to the other room and present my roadmaps, right? So this is a style. We are engineering company, but everyone is passionate about business.

Every, uh, one is passionate about the things that they design, right? So I'm gonna tell you a little bit about bison story. So like, uh, like the introduction says, bison is everywhere.

Actually, every one out of five, any name, first storage device have some kind of Python technology in it. But we always have been under other people's brand, right? Uh, if you look at our revenue, $2 billion company, we are public listed in Taiwan.

I asked you, who knows, 5,005 years ago, nobody knows. Okay? But we actually started 25 years ago.

This is our 25th anniversary, and we started creating the first single trip, USV. We pioneered this technology. And fast forward today, we're gonna talk about from a 60 megabyte USV drive.

We remember that all the way to 1 22 terabytes of data storage today. That's how crazy it is. Um, look at the head counts, 4,500 head counts in the Hulk operation, but still only 5% focus on r and d.

So we use our partners to do marketing to go go to market for the storage side for the past 24 years. And we are just a engineering arm behind them. My lab is their lab.

They don't need to have a SSE team. They use us as their SSD team. What our business focus is this morning, you hear about adaptive ai, a very creative way to use a end storage to solve a insolvable problem of model scaling, right?

So nobody, you know, I think there was a one question about, Hey, how big is the market? And guess what? We go to every customer, they ask the same question, right?

But nobody is really looking at this as a problem because they can't afford it. They're not trying to, they, they take what they have on the public model, right? And they, they, uh, they make the best use of it, right?

But we're saying that, Hey, now you have another way to actually make your bottle better. And if you hear the chief, uh, avid, CEO says, own your own ai, but how do you own your own AI without your own data or model, right? How do you keep it on print, right?

So, so enterprise, uh, uh, AI is a, a, a big focus now, but oh, bread and butter is still storage. That's what's paid our paycheck today, right? And enterprise is our, uh, highest priority right now, because that's the biggest data growth today is, uh, smartphone scaling big.

Now, is laptop scaling big now? No. Right?

How many USB drive do you need? When's the last time you used a USV drive? Sd no.

Right? It's in the data center, right? It's the data AI at the cloud and AI at the edge.

So what's the percentage between, uh, consumer and enterprise right now? Uh, so three years ago, it's 5% out of the 2 billion. Oh, wow.

Okay. 2 20, 24 last year is grow to about 12 to 13%. So it is going to take over our company's revenue very quick.

Or I should say that if it's not 30, 40% of our revenue, uh, revenue, that means we're not catching up with the growth, right? So we are very much, that's why we have a local enterprise lab sitting in the Colorado. They have hundreds years of experience combined.

You know, all the people that used to do actually hard drive in the, uh, uh, validation for, you know, the, it's, it's a, it's a partnership with the Seagate, right? Mm-hmm. Uh, so, so they have a lot of experience, uh, for, uh, making sure that we capture this market.

Well, I gotta believe the margins are much better as you go into enterprise as well. Consumer tends to be very narrow Margin. Well, it depends if you only create a cookie cutter product that everybody has, yes.

How much, uh, 3 cents, 4 cents per gig, 5 cents a gig, right? But we're doing different today, and that's what we're gonna share with you. Uh, we are global.

We are global. So 4,500, uh, headcount total, we have our part of the r and d. And you're gonna see that, you know, engineering, you could see storage, right?

It's a controller, it's a nin how hard it is. Actually, the metric is in the framework. It's very difficult.

So one thing that we do very well is we have our core engineering in headquarter, while we have regional office in the us, in Malaysia, uh, in China, in Japan, and in India. So one of the thing that we have observed is that be able to be good at selling your storage on every regional is not just plainly created a lot of regional office, right? It's about how do we create, how do we work with the local partner and create like a jv?

So we did that with India. We announced it earlier this year. We create Malaysia office that's gonna launch next, next year, uh, sorry, next month, right?

Uh, my storage is focusing on design in Malaysia, manufacturing in Malaysia for the AI infrastructure in Malaysia. So Python, go to market on regional, you'll see Malaysia, India, China. How do we work with the local partner, either in the form of JV or other kind of arrangement, and create a local king that has local personality, local channel, you know, local network, right?

And, and then in the us we position as a technology, kind of a pioneer for the company. We talk with the, you know, what are the biggest tech company in the world today? It's all here.

Who get the whole, the, the fastest requirement? It's all here, right? Uh, our CTO office is here.

So that's why we wanna be, we want to be the pioneer, drive the direction of the company, and also have a local enterprise lab to serve the local customer. What About in Japan? Is that r and d or is that just, uh, sale local sales Office?

It's, it's, uh, FAE office and then business. Okay. Over there, right?

Okay. Let, let's look a, a little roadmap here. It's a very busy, I'm just gonna pick out a few.

One. So 2000, we started with A USB. Actually, I have a, a encounter with, uh, ex, uh, Intel, CEO, pat, uh, a GA singer.

And he say that he's the one to actually revive, uh, uh, a USV interface, uh, from, you know, alpha try to do the proprietary. And he's saved. And I walk up to him and I say, Hey, thank you.

Without you USV, we won't be here. Right? So that's how we started.

And then quickly, if you, uh, in the, in the four year, we became the fastest company that went to IPO in Taiwan, four years, and we are making money. Nice. Okay?

Uh, and then 2008, uh, 2004, same year, the whole world moved to the digital card, if you remember that, those days. Uh, and in 2008, we went into industrial solutions. Nobody like to do industrial solution is, is so much customization, is, is a high customization and, and high mix, right?

Uh, we do it, we do it behind our partner, and then the will go into the ultra book in 2010, right? Uh, and, and then all along, uh, we have, and, and then, and then eventually it went to the regular notebook, and then eventually it took over the entire notebook and then, and then the enterprise. So every single part of the journey of every product segment, because of our dedicated engineering, we always shoot for the best, fastest to market.

We always have been the leader of every single segment as we move on, right? Um, and one thing that I'll share on the next slide is, you're gonna hear this two words a lot. Today, worlds first, before 2018, the worlds first belonged to a Korean name manufacturer.

Oh, what's the latest, uh, gen three? Yeah, Samsung, when is the first, whatever, it's, it's all name makers, right? That all changed in 2019.

Who is the first SSD client? Bison, who is the first gen five SSD is, that'll change. Um, wanna highlight a few award?

So, um, we are in space. So because space has such a big demand in storage, no, come on. No.

A lot of data. No, no, no, not yet. It's put market, infinite market out.

But if Elon Musk wanna go to Mars, data needs to store somewhere, right? So when we took on this, we said digital version, uh, engineering, say K Ks, our CEO, why? Why, why do we want to do this?

Hey, this is a representation of our commitment on, we keep saying we could customize and we say no to space. Come on. Who can say no to for people who really wanna go to space?

Right? So we did it, right? 2000.

We went to 2001, uh, sorry, uh, uh, 2010 we went to March. So our perseverance last year, we have couple device on ISS. And this year, our SSC is in a lunar mission to the moon.

That happened two months ago, right? This is the very first air center that was a, that, uh, highway architecture on the moon, right? Um, and then, uh, you hear about a lot of adaptive that's, uh, uh, something new.

Uh, we also are being recognized a lot in the SSD. Obviously you hear more about, uh, us in the client. Mm-hmm.

But actually just yesterday, and you're gonna see a lot of review coming out of our pa gallery, X 200 SSD. There gonna be more review coming out. We're making some noise now.

Okay? Uh, also, you see on the left hand picture, there's a little, the mountain, the world's strongest man. Uh, that was last year.

Uh, super compute, actually. We broke the, uh, record in a good way and a bad way. What?

In a bad way. It's the first time they had to have a, a, a, uh, uh, officials coming and said, Hey, you guys are drawing too much cloud. This is calling a fire hazard.

So we have him lifted thousands of SSD 1 22 terabyte in one lift, and broke the highest density that ever lifted by a human being, right? So, thanks to my marketing team, that was a very creative idea, okay? Uh, on the right hand side, very small, I'll help you read it.

Automotive. So, have you guys heard about dual port drive? Dual port enterprise, okay.

Mm-hmm. Most. So let me tell you, all the drive, usually today, a lot of it is single port, but dual port drive was used more, but it's still here on the OEM side, right?

Dual port, right? Because it's for data protection, right? Um, so, uh, automotive, how many port do you think we need?

Automotive? It's just a car. It's not a data center.

It doesn't run and NVIDIA's a superpower, GPU or whatever, right? Right. Well, they also have a product, but they don't use that crazy product over here.

Mm-hmm. You are looking at a full core, four core, a, a port, sorry, or quad port. Quad port, SSD for the first time we work with a partner, right?

To release it to the market. So you're looking at a data center on wheel, and what they're doing is combining all the element in the car. You have one for a desks, the dashboard, every single put coming in here and managed by a single storage that they can orchestra.

Right now, there are a lot of AI applied to a car, automotive, uh, car. So, you know, by combining all the storage, right, you could do a better orchestration. And, and we, we are the first one that makes a corp, uh, carport, uh, SSD controller.

And we didn't make it for their center. We actually did it for the car. So you're gonna see more and more of these coming out.

It's called centralized architectural for automotive. We've been, uh, growing very healthily. You look at this, it's a good growth.

But one thing that is behind the scene every year, nan price erosion is about 40%. So think of how much we have to ship every single year to keep the growth. It's crazy.

It's crazy. And while we are in the enterprise, because even if the units are not growing, the capacity is growing like crazy, right? So we are able to hit the 1 billion in the first 10 years.

We hit another billion, 2 billion in 2020. And, uh, from 2021 to 2024, there's a big, you know, price, uh, uh, uh, you know, market dynamic. But we are still able to keep around 2 billion.

And we believe that going forward, the, the, the future is, is really bright. And, and the reason why is, unless you are the world's first, and you don't make the name flash, who cares about you. So this is always in our DNA, not only we are gonna keep selling more product, but we are going to be the world's first.

5 in enterprise. So we started enterprise actually four years, five years ago, but we only branded our product last year under pasi, what you see here, right? But we created right now the highest density data, drive, fi, uh, and then the Gen five SSD also fi we're the worst first.

Is there something like a corollary to Moore's law like you have with CPUs as far as in storage, the amount of storage increases every year, and how much the price goes down, or a rule of thumb? Yeah. Uh, so for, for storage side, right?

Uh, it used to be most company are trying to get to the next process, no, every 18 months. Mm-hmm. Right?

And by going to, uh, the next node, uh, I would say that not every single transition, but every two transition, you usually would double the density. So 36. So, so meaning that the, it's like a TikTok.

TikTok, right? You have one generation, you shrink the dye, and then another one, you shrink the dye further, and then you double the density. So right now, a single dye, a single dye density now is two terabytes.

So a little chip, uh, a single, you know, you can see the SSD, there's a lot of chip, but every chip actually is stack with a different dye. And each dye, even one dye, it's 256 gigabyte 1D, right? So this is why even cell phone is you're seeing the five twelve one terabyte, right?

Uh, so there is a 18 month, but I think that if you see the market dynamic right now, because name flash, there's many players, there's many up and down. So, uh, people are trying to kind of slow down the pace so that there's no, like oversupply situation. Mm.

Right? Yeah. And then the price with that, does it, is it going down for As, uh, yeah.

Yeah. So, so typically, uh, my door of thumb is that, uh, the, uh, in terms of the, the, the pricing go down is that usually, I would say every, uh, every, maybe every, every three years, maybe you will cut the cost by half. Yeah.

Yeah. So, uh, following to that question is, uh, do you have, uh, any progression of what the density per watt is, Uh, density per watt? Yes.

Uh, yeah. I don't have that detail information, but in terms of the density right now, uh, because of the, the single die, the two terabit nan, right? Each one.

So you could actually, uh, you know, today we're looking at 1 22, right? Next year we, we going, you're gonna see like 2 56. And all these devices had to be buy into a form factor that has a max limit wattage, right?

So for the U2 25 watt, you have to crank 1 22 today mm-hmm. 2 56, you still have to be below, uh, at the 25 watts. That's why there's a new factor that e three point s that you will see today that is 35 watts, right?

So in that example, you double the density, same voltage. Yeah. But I mean, even when you get to the 1 22 gigabyte, you need to do power shaping inside the device to be able to, to manage it.

So I, I was just wondering at the chip level, if you had plotted anything like this right? On what the, uh, what the density per watt is, because I'm guessing that that also goes up as well. Yeah.

That, yeah, if you're talking about the density per watt, I think that it will go along the density scaling a lot because every single generation, nobody can actually accept that you have higher power, but the performance, it doesn't like double every time, right? Uh, but it was, you know, for example of, uh, when we started the first gen four, the performance at that time, the interspace feed, I think it was like a 1200. And then by the time we go to gen five, it's 3,600, right?

So that's, uh, four years. But wattage for die to die from a, when it's a, a, a 60 foot, uh, gigabyte per D to 1 28 to 2 56, the wattage, you cannot ex increase too much. Okay?

Yeah. Uh, patent is, uh, to protect our intellectual property. And you will see the majority of patent is behind the n, right?

N applications, how we handle the hand, uh, n handling fee. Uh, and on the security side, uh, we have majority of the, the important region cover. Okay?

So what is different about Pfizer? And then that will fit into our Pascal story. So if there's one word that you remember today is customization.

And this is something that we do this a lot because we are behind the brand. So it is the DNA that we think that it is something that the enterprise market is missing today. Okay?

So we are customized ready, we have all the right ingredient in the technology stack because we're vertical integrated. So to make a name device, you have a controller hardware, and then the IP insider controller, and even the name maker, they actually don't design the IP within the controller. They outsource to different companies.

We did not, because to get to the time to market, you actually have to own all the technologies so that everything work, you know, perfectly orchestra perfectly. So we make from controller homeware hardware, we do the PCV board design, we make the entire device, we put, we even design the housing for our customer, the custom hits sink for the customer, all the way to labels shipping, logistic. So we are, you know, who, who makes the burger today?

Is it McDonald? No. Right?

But who is the most successful company today making burgers, right? So it's the business model. It's your whole, it's your mom's, uh, burger.

Yeah. So, uh, so, um, uh, you know, this is what's really ingrained in f is that we are vertical integrated. Uh, and then we have a lot of trusted partners that have, uh, been with us for a long time because what they get from us is not a product.

It's, it's a partnership. Let's talk about pascali. So why does the world need another enterprise?

SSD. Oh, that's why it is paska, okay? Artificial intelligence, security.

Ah, so paska come from, we are performance, we are artificial intelligence. I like that. We are security capacity.

We have the highest density today, acceleration, reliability, and innovation. That is paari. And our tag line is engineering imagination together.

So now when we try to think what define our personality is engineering is actually what makes the product works, right? But our customer's requirement, do we want to always tie to what we offer is a standard product? No.

We, what our brand's core value is that we want to plan our engineering and your imagination and working together. So do not tell us, hey, uh, what standard product you have, tell us what you need first, and we will see what product we have for you. This works, this works, this works.

And if none of them work, you need a little tuning on the power. You need to have a special command because you are going to the moon, like special data logging. Oh, sure, we'll do it.

Well, the other company may do it if you sign them a big paycheck. Or maybe you are the top five company, a large customer in the world, but small companies or special mission company to the moon, they don't have choice. They're settled with what you give of them.

So we want to have a engineering and new imagination working together. And like I said, it's a product. It's not just a product.

It's, it's a partnership. And that's the what it takes to win for us. I mentioned about the enterprise growth or added to that, uh, on the next slide, and one thing that is, uh, special of ours, everything that you see, uh, our enterprise drive, our, our, our high performance portfolio is our dual port.

And this is something unique. Uh, we have our Imagine plus we brand it our customization service under Imagine plus design service from firmware to hardware flexibility. And we are not just customizing products, we're customizing business model.

So let me ask a question. So with all the company that is trying to make an AI system, why did everyone try to design their own chip? I, I'm sorry, uh, I'm supposed to get the question from you, but, uh, I would like to, uh, discuss, discuss why did they all have their own asic They want something different for the ai, they want to save cost, or, you know, there's different reason, right?

And you have seen different ways to make the hardware the form factor, everything, but who is providing the flexibility on the storage side? No one. So, so what, what does the, when the box company want to do something different, what do they do?

So they actually have try to make a SSD team. They try to build SSD team, right? Many have company have done that, and almost everyone failed to do that.

You know why? Because when they really need to have their own SSD because of fresh shortage and all that, it makes sense to build it yourself. They start building and they take a three, four years to finish, and then by the time they finish, it's cheaper to buy at off the shop, right?

So it's this cycle. And the largest company, I'll probably name a few after our session. They, they, they, it, it all fell, right?

But when I say custom engine business model is, Hey, you know what? You like our technology and you have a good supply chain on an end. Hey, let's take your strength consign to me.

How be for you? Right? So I think something that you see, what we do differently is, yes, we have a standard product that everyone has, but we also have a capability to customize product and business model.

And our, our our fourth analysis, uh, you will see a lot of big brand products out there today. They're focused on the mainstream and they focus on the high end, the standard product. You will see some one point s or a lot of, uh, GPU servers.

You see a M two for the mainstream, for, uh, the, um, the boot drive. You also see the QOC drive. That's also what we consider mainstream, because it is the value, the price per gig.

And then you see the high end, which is, hey, gimme the fastest drive, right? What we bring to the market today is two additional, we supply the legacy, we will continue to ship SA Drive. Yes, I know it's old, you know, but we'll continue to ship that.

We are gonna be the last vendor that's gonna ship a SA drive. So Michael, You're welcome. Uh, yes.

Hi, Kimberly Bates. Um, and if I'm taking you down a path that you wanna go down, that's fine. Question, kind of question that goes beyond all this.

Yes. You guys are the big manufacturers. You supply so much.

Yes. What is the plan currently in terms of the tariff issue? Great question.

Great question. Okay. Uh, this actually, yes, everybody is having their phone next to a year for the last two weeks.

Yeah. Right? Everybody knows the rule, right?

Manufacturer in US or manufacturer in somewhere else to come in, right? Uh, there's a big movement, right? But who can actually make that change quickly, Right?

That's a big Question. So, yes. Yeah.

So I think that the core, the technology, the firmware and all that, you know, the intellectual IP design in the region, but in terms of how to ship to different region, in this case you mentioned about us, right? We have to be smart, we have to agile, we have to talk to, uh, manufacturer to, to, to, uh, actually manufacture the PCP board locally, right? Mm-hmm.

Uh, or, or in regions that has certain treaty through the us. Uh, and also, uh, in the interim, Hey, can we move the entire testing and everything like tomorrow? It's not easy, right?

So, but you know, we have, uh, very internal experts think of different logistic way and, and try to make it work, right? So it is a problem that every name, vendor, and actually, uh, uh, you know, there are letters issues that most of people are gonna pass a terrorist to their customer. Mm-hmm.

Right? But that would actually slow down the vendor a lot, and that's a problem, right? So we need to provide a customer with the option that, where they can, you know, minimize the tariff.

Yeah. Did I answer your question? I cannot say too much.

I don't think anybody can Answer any of the questions. I can't say much detail, you know, I'll get in trouble. Uh, we are in a so-called process of moving the manufacturing over here, it in process.

And we are commit to our customer that we will help you with this. And, and down that line, again, if I'm taking us down a line we shouldn't be going. Um, I know it takes about three to five years to bring a fab on, or a fabrication and that kind of stuff on That.

Um, may I interrupt? Sure, please. Uh, how about three months?

Okay. Well, that's a different game. Yes.

Very different Game. Yes. That's, that's, that's the, we cannot wait three months.

'cause by the three months, uh, my enterprise revenue cannot hit the 30% our revenue. Right? Okay.

So this, this is moving fast. Yeah. We have to act quick.

We have to act quick, fast. Interesting. Okay.

Right. Uh, and then I mentioned on the top of the pyramid, custom design. So think of as we have products, standard products, standard phone factor, and we are going to drive to lead the technology.

Uh, we announced a brand last year, and this, uh, very quickly, within a year, we announced the largest, uh, enterprise drive right ahead of many players right now. Mm-hmm. So we're gonna have the best looking forward product.

And on top of that, we are providing the customization. Um, um, due to the time limit, I'm gonna skip ahead, uh, in terms of phone factor, I mentioned about, uh, you know, so everything we have, right? A any single phone factor that we have from legacy, uh, so what, what do the, see that the trend that you are seeing is, there's a lot of udot too.

You know, that's like a same pizza box at the, uh, this is M two, right? 5 inch. It's try to be the shape of the hard drive.

Uh, and there are recent movement to move to the, uh, where is it? The, where is it? What I'm looking for, uh, the, the three s form factor.

You see E three, E three, uh, it's a thinner profile, right? And it has a short version, a long version, and there's a lot of movement moved through EE three because, uh, you could actually fit more drive into a two U form factor, right? Uh, and right now we are also embar embarking a, uh, a, a project, right?

Which actually take advantage of our capability. How do we replace a hard drive? How do we replace a hard drive?

There's so many out there, right? Can we do a drop in replacement? Can we have a what?

You know, the thing is, every generation know n people are trying to go there faster, faster, faster, right? But a lot of these, uh, AI data center, what they need is a capacity. And the code storage, they care about the capacity, they don't care about the performance.

So how do we design something that will hit the capacity point, super low power, so that we replace a hard drive, right? So we are under a lot of, uh, this is the, you know, the, our, our imagining process is how do we design something that we can replace that Are form factors proliferating or is there, 'cause it seems tough to have all these different, um, form factors and, and where they would have to go into, right? Are a lot of them being retired and new ones coming up?

Yeah. Yes. Like I said, so data, you see a lot of, uh, ADA in the more legacy platform, uh, people, IT department, if they need to replace their drive, they still need a place to buy.

Uh, UDOT two is the most common today for the the server. Which one is data center? This is the UDOT two.

Okay. Right? Uh, but then there's certain hyperscalers they have use some unique one because they're self sales consumption, right?

They, they can actually have a very unique phone factor because they're, even though they're the only customer that buy it. Uh, so this is, this is very popular. And then this is getting into the E three, uh, which is, uh, over there on the very right side, the E three phone factor is thinner, right?

It has 2, 2, 2 profile, actually the one, uh, uh, the, uh, the thinner profile and the larger profile. Um, so these are the main one. You're looking at this and the E three, and then you are gonna see the, uh, the GPU uh, server.

They have a little tiny little slots in the front. That's the E one s. E one s.

There's a shorter, and it's getting more, uh, pretty popular these days, right? So if you think of a GPU server that is running the actual GPU in the same server, you see a lot of one s. But if you're talking about the storage server only, you see a lot of the Utah two and then the E three today.

So the, it's kind of stabilized. Yeah. The D 100 p, what are those for?

Oh, this is, uh, uh, you mean the Real super thin and little one? Oh, This one U drive who drive? So some servers, you need a boot drive, okay?

So it doesn't need a high capacity, it just need to have the power fail, power protection and, and, you know, to run the server and you have a question. Oh, I think one thing, kind of a corollary to what you were asking is, okay, so you've got all these different sizes of drives. Are you creating the different sizes of drives because the hardware is changing?

Or do you have to, like all the hardware manufacturers now need to be changing, you know, their, their drive Sets phrase. Oh, it's the standards. Oh, yeah.

Uh, yes. It Doesn't look like there are standards, really. Yeah, I guess Well, well, actually they are standards, yeah.

That define this. Uh, every single one is a standard. Uh, even if they don't start as a standard, they become a standard.

Mm-hmm. Because one of the biggest thing about a storage, uh, a server vendor is not having a healthy supply chain. They need to have different sources of them.

So all these are the standards, and we have to have every kind of form factor cover, but something that I just mentioned about hard drive replacement and all that, right? Yeah. We are at the point of, you know, the reason why is this shape, and, and, and you're starting to see this kind of shape, because when we design SSD, everyone is like, okay, I need the shape of a hard drive, right?

So everyone look like this, okay. And then people say, Hey, why do I need to live in this? Eventually, it's all going to this box.

Why can't I design a triangle one or a round one, or whatever. Of course you're not gonna do that, but why, why can't I design something more feasible to this? So you're trying to see this, And that's kind of floppy discs.

The one, the thicker one. It's like a floppy disc form factor. Yeah, Exactly.

Remember the old, The old or what It is? Just that sort of square thing. Yeah.

Right? Like a lot of, uh, today, like a SaaS drive, but it is, it's still there. It's still good for zip, they called it Leon Neon line Hard drive, uh, is a faster one.

And also you have the code storage one, which is, you know, and, and you know, the, if you look at the, the hard drive company, they're laughing, they're still laughing. There's still a lot of hard drive, right? But then over time, right, as a SSD vendor, we have to think of a way through how to get into that.

And there is a way, so there are, uh, another form factor called E two that's under evolving. It's a new form factor that's coming. So usually the typical cycle is the interest in the, the industry recognize there's a need, and then people from vendor, the device, they will get together and then they will think of a phone factor, and then it takes about three, four years for it to be Yeah.

Takes time. Um, getting, oh, okay. Uh, so, so, uh, one thing that you, you you'll see differently is now not only you see all this product on the OEM product, but you could actually get our product in the channel now.

So we are in MAL lab, we're in, we're in a SI, we want to be available to the public. Um, a very, a very quick rundown, Chris will give a, a, a, a much, uh, detail. We have the high performance X series, we have the data center, that's the QLC drive.

So every, every time you think of a large density, uh, I don't know know if you re uh, uh, familiar with the QLC is the, the most, uh, value, uh, n technology. Now, uh, it has a largest density per wafer. So you have, uh, the, this, uh, lets belong to our data center series.

And then you have the ADA legacy, you have the boot drive, and then this morning we talk about the hundred drive drive per day for AI edge applications. Okay? Um, I, I mentioned, uh, before, uh, what do we do in, in imagine plus custom SSD design?

Uh, you know, we, we take actually name from everyone, uh, faster to market. And, uh, we want to, uh, you know, build the next, you know, imagination, right? For you, AI and enterprise.

Okay? We follow all the industry best practices, uh, a lot of testing going on, uh, and to meet all the demanding need of, of the enterprise. So six, 6,000, uh, testing hours.

So every drive, uh, has to go through this coing RDTs, right? And it's a very big investment. A lot of customer, they see this presentation I have in front of you, they was like, oh, great.

Oh, great. And then I said, can I take you to our validation lab? They see the lab and how much we invest, and they were shocked, right?

There are a lot of investment behind all this, uh, uh, uh, you know, in the making of these drives. So if it stops operating at 40,000 feet, how do you put it in space? Yes.

Uh, this is my last slide. It's, uh, because I'm a tech geek, so I need to speak some, some technical, you know, stuff. Thank you.

Of course. So I'm gonna tell you some, some, some magic behind the curtain. Every months.

So why are we able to be the times to market? Why? Mm-hmm.

Oftentimes in the development of a product, you get the nan the chip, you get it in-house, and then you take about three quarters to make the firmware, test the firmware, perfect the firmware, make sure it works about three quarters. So by the time you, the nan is ready to your product ready, sometimes three quarters. But actually, if you, a nan business is like a simi, the minute it's ready, you want to serve it.

You don't want wait, because your job is to sell the nan, right? Mm-hmm. How do we conquer this problem?

Whenever you have a nand, you take a three quarter throughout the product. This is the reason n emulator system. This is our secret ingredient.

What that does is, before we get the ing, before we get the n, we actually have a system. It's a computer system that emulate itself, is a n And the benefit of that is when my controller is in development, my phone was development. I already have NI can play with even before the n arrive in my house, right?

And also, another benefit is usually when you have all this, uh, failure mode with the n right? To make SSD work, actually, it's not that hard, but to make it run well and have all the checkbox click, it's difficult. So usually in the past, I would say, Hey, can you gimme a name?

I, I want to see that re retry issue. I want to get this issue that, uh, only happened in the high 10. I, I'll just keep getting all the name from, from the name vendors.

But then it takes a long time to get that one golden sample that failed in that condition. Guess what? Why don't we create the failure with our n system?

We create all the, so this net emulator is like a computer system where it, it is a n and it has all the failure system built into the library. And by the way, we are the only vendor that actually have, we can build product with the different vendors' name. So we have all the knowledge, right?

So that's something that is really valuable to us to get to the market. That's why something I didn't mention is when we had our first gen four SSD to the market, you know what? From the moment that we get a presentation like this, customers say, I want a gen four.

Fast forward nine months, only take nine months, right? We got a product on the shelf. On the shelf from a asic not created.

It is only on a concept. We build the asic, we build a ware system, put it together on the shelf to align with the AMD's Gen four launch in 2019. Typical cycle for this is two to three years.

We'll shake it to nine months. So, so let's, that's something that I wanna share. So, quick question.

You talk about this SSD uh, design validation. Yeah. And in the first slide that, uh, you, um, use your SSD for, um, um, missions, uh, in the, in the space.

Yeah. Are you following the same design process validation, or is something Yeah. On the space is completely new.

It's completely new. Completely new. Yeah.

Yeah. Uh, we, we had to actually ship to different special slides lab. Uh, the customer has to put some special, uh, shooting, uh, uh, radiation vibration.

Yeah. Yeah. Because the radiation Ation there, so Yeah.

So we don't expect you, uh, yet, right? To bring this standard product to the moon. But, uh, any any special design, the, the, the, and the other reason that we do is actually a lot of technology can average onto our standard product, right?

Yeah. That's why we did it. So you could see like from experience or like how robust our PCB design.

Uh, you know, we use industrial grade, all that. It, it could be actually baked into a standard product, but yeah, all, all the special kind of, you know, whether it's going up to the, the sky or, or, you know, a few thousand feet below, everything has to be kind of special validated on the customer side. Yeah.

And is, uh, my, my curiosity, is there a, a paper, a public paper talking about this, uh, uh, validation? Are you solved or Tried? Yeah, we, we have a, a block.

Yeah. We are actually starting to put a lot of the general flash 1 0 1 validation, uh, uh, industry standard practice. It's been around for like two years.

There a lot of material. I'll share some information with you.