All right. I'd like to introduce our next presentation here at Planet MicroCap Las Vegas, powered by the MicroCapClub. We have a fireside chat hosted by Lake Street with Backblaze.
I just wanted to introduce myself. I'm Eric Martinuzzi. I'm a senior research analyst in the technology sector at Lake Street Capital Markets. It's my pleasure to introduce Marc Suidan. He joined Backblaze in August of 2024, so coming up on two years this August. Prior to that, he was CFO for Beachbody Company and 17 years partner at PwC. I have a list of questions. I can go for probably an hour, but we've only got 25 minutes, so I'm going to pause at about the 15-minute mark and open it up to the floor if there's any questions. Otherwise, we'll go right back to my list. Let's start it out, Marc, with a description of Backblaze. What's the industry that you serve and what are the revenue segments?
Perfect. Good morning, Eric, and hello, everybody, and thanks, Robert, for having us here. Quickly speaking, Backblaze is called a cloud storage company. What does that mean? You've all heard of the hyperscalers, AWS, Azure, and Google Cloud. There's typically three workloads that go to these companies. It's compute, networking, and storage. We're a pure play on the storage part. Companies will come store their data with us, and they work with all the different compute providers or networking providers. We typically offer it at 80% below the hyperscalers. It's a incredibly innovative platform where the software layer intelligently manages data in and out in a way that we could deliver it at a lower and simpler way than everybody else. That forms a majority of our revenues. It's right around 60% of our revenues. That's the one that's growing 24% a year now.
The legacy business, which is pretty much flat to slightly declining, is a computer backup business that makes up the other 40%. It's backing up your laptop, and computers, about 400,000 customers on that. On the cloud storage one that we call B2, the one that's growing 24%, there's about 120,000 customers on that.
Okay. I'm picking up on that B2 cloud storage. That's what's interesting to me as an analyst. That's really the exciting part of the business. What has been the key driver of the growth? Is it just more people deciding that the cloud is going to be where my active storage layer is and that's reliable and the pipes are big enough? Or is it something AI related? Is it something blockchain or what's been the big driver of the growth there?
Yeah, if there's one common denominator with data, it just proliferates, right? There's a lot of data proliferation. If you do nothing, data growth is around 20% a year. There's a large tendency for all companies to either move things to the cloud, or if you think about all newly created companies are cloud native. In other words, they don't even have the on-prem solutions. Then you layer over that the AI boom that is happening. With the AI boom that is happening, the amount of data growth is pretty explosive. That data's got to sit somewhere. There's only three different kinds of physical devices it could sit on. It could either sit on flash memory or high-capacity drives or tape, and we operate a very large platform of over 330,000 high-capacity drives. It's our CapEx. It sits in data centers that we operate.
We're vertically integrated from that standpoint.
Okay. The company has talked about customers generating over $50,000 a year in annual recurring revenue, or ARR, that that now number is 187, and that that customer count in that 50K-plus range is up 51% year-over-year, and that was in Q1. How are you changing the sales, the go-to-market motion to address that kind of upmarket thrust, and what does the average deal size look like now, and what'll it look like two years from now?
Yeah. Eric, for those that don't know Backblaze well, Backblaze, for the computer backup, was largely a consumer-oriented business with a bit of small businesses on it. When it launched this B2 cloud storage platform in 2016, it's what you call product-led growth. It's a self-serve motion. The 120,000 are overwhelmingly every month, we get over 1,000 net new that come sign up and start using the product by themselves. There's no interaction with sales. The company launch a direct sales motion. Typically, when a company goes and launches new ways of selling, there's some new muscle to be built there. We've been aggressively building out that muscle. They're the ones who go deal with these larger customers, and as you note, the growth of that customer set is very high. We used to talk about $50,000-plus ARR per customer.
Now we're disclosing also what proportion of our ARR comes from them, and you could see that's growing even larger than 51% because the average deal size keeps growing, right? The average deal size of those customers is over six figures. We've been announcing, I want to say also roughly almost around $1 million AR customers about at a rate of one per quarter for the past five quarters. In one of more recent earnings release, we also announced an even larger customer than that, one that committed to over $15 million of TCV. We're continuously going after bigger and larger users. I would say the common denominator in these customers are one of two. They're either an AI late-stage startup that's growing really fast with a lot of data proliferation, or they're what you call a neocloud provider.
A neocloud provider is a broad term for a lot of these computing companies that are rising across the board, they're using us to offer what you call the warm layer of storage. Because in order for them to do it on their own, it's very hard to catch up to the intellectual property we've built. That's the one that's driving the bigger TCV list. We launched a new solution that's really focused on those B2 neoclouds, we've already signed up quite a few of them, we're talking with quite a few of the large ones.
Okay. The average deal size is rising, but you're not disclosing quarter by quarter on the B2 side?
About what the average deal size will be?
Sorry. Yeah, the average deal size.
Yeah. What we did two quarters ago is just until we get that motion in place that's repeatable and we're comfortable with it, we said we're not going to give guidance to how many of those customers we'll land per customer. In general, we consider a large customer somebody who spends over half a million dollars a year with us. We've been averaging one to two per quarter, there's certain quarters it could be zero. There's certain quarters it could be three. Last quarter was three, it was really nice, and that's why we did a really nice raise in our guidance for the year. Every quarter passes, if we could keep seeing repeatability, we'll start layering that into our guidance.
For the time being, our guidance assumes we're not going to be closing any of those for the rest of the year, which is a very prudent, conservative guide in a way.
Yeah. I do want to go there because you guys sort of stood up and said, "Hey, we're changing how we guide." Right? In 2025, we guided a certain way. In 2026, this is how we're going to guide. There was definitely kind of a whipsaw moment for myself as an analyst because I was modeling conservatively, because you guided conservatively, and then you pop the clutch first quarter out of the gate. Explain what went into the revision and guidance logic at the company.
Yeah. When we reported Q4 and we set the guidance for 2026, we said it's going to be $157.5 million for the year at the midpoint. When we reported Q1, we raised that to $162.5 million. Half of that raise was driven by a price increase we did that the market's taken really well to date, and the other half was from the three larger customers that we booked in Q1. Some of our larger customers using us more than we expected in Q1. To your point, we're off to a great start, the way we're guiding is one that's going to be a lot more prudent, just because what we experienced last year was the whipsawing where we said we're going to finish the year growing 30%, we finished at 24.46%, right? It didn't quite hit the 30 mark, so that disappointed many.
Now we're saying, listen, we're going to keep it exactly at what we could feel really secure, that we could predict and repeat. Everything over that we'll be layering over every time we announce. As it becomes more predictable in nature, we'll start embedding it, we'll be transparent as to what we're embedding in there.
Okay. Part of the business that I'd like to spend a little bit of time on is the computer backup, which historically had been the dominant part and now is the minority of the revenue stream. In the first quarter, it was $16.2 million. That was a 2% year-on-year decline. Is there a revenue level on a quarterly basis, on an annual basis, that you feel like this sort of melts to? How do you manage this CBU computer backup decline while investing in B2, the growth area?
Because we're investing a lot in the B2 growth, because that's a really fast-growing market, right? That market, that TAM grows at over 20% a year, and we feel we could outperform the TAM there, and that's what we're working on. In order to do that, we're using the legacy business, the computer backup, as the cash cow. That market, that TAM, independent of us, that TAM grows low single digits. We could either put a lot of energy and grow low single digits or put little energy and contract low single digits and use that cash cow opportunity to fund our B2 growth. We're taking a few actions to make sure it stays where we would like it to be close to flat. We've guided to it contracting more than that just to be more prudent there.
What I would say is that the business model there is one where the gross margin is not as healthy as the B2 platform. The B2 platform is the larger one and growing is purely consumptive, everybody pays on a usage basis. It's actually the gross margin is healthier there. On the computer backup, because it's a fixed price, consume as much as you want for a fixed price, relatively speaking, and data's always growing, it ends up being a lower gross margin. We'll keep monitoring that side of the business to make sure we have profitable customers, and thus optimizing the cash inflows we could keep fueling the growth on the B2 side.
I wanted to follow up. You made a comment about neocloud, just correct me if I'm wrong here, but this is my dumbed-down explanation of what neocloud is. Everybody knows about NVIDIA, and you want the best and fastest GPUs, and this is kind of neocloud equals GPU as a service in the cloud. What you are talking about with your $14 billion neocloud opportunity is that the storage associated with the GPU as a service?
That industry is obviously growing really fast, and you described it right. It's a GPU as a service, or it could be CPUs as a service. It's effectively that compute aspect of the business. GPU's been really hot in the past two years. This year, CPU also blew up and is really growing fast, in general, compute is a demand constraint. There's about 200 neoclouds. Neoclouds are companies that offer computing as a service. Most of them have to offer storage because in order to run a compute workload, data's got to feed into it, and that data's got to be sitting somewhere. There's different types of workloads. If you're about to really run it through the GPUs, it's got to be sitting in the flash NVMe-type, high throughput semiconductors, and those are very expensive.
Those are up about 10x year-over-year, versus if you look at high-capacity drives, they're also up year-over-year, but they're up about, let's say, 50%. Now you're creating a big difference in cost factor in a supply-constrained market of how much it costs to store that data. Everybody's looking for what you call that warm layer. The warm layer is basically something that moves around really fast but is the one right outside the GPU. Not the one by the GPU, but right by the GPU. When you're running AI, you got to aggregate a lot of data and schedule your time of when you'll book the GPUs, and GPUs are very expensive, and that's where that warm layer is where we have a huge competitive advantage, right?
The software layer that we've built over the 20 years we've been in business really optimizes well the way the data interfaces with all these drives. Most companies, when they launch from scratch, it's called the IOPS. They don't know how to interface with drives in a high throughput way, it ends up slowing things down. We have a really great competitive advantage and a great cost performance, it's creating a lot of good demand. For the neoclouds, we're talking with a lot of them, and we've already signed up four where they use us as their warm layer storage.
How does a neocloud contract differ substantially from a traditional B2 contract?
There's a few attributes that are different. Number 1, they could be, like in the case of our RPO, you could see a meaningful increase in our RPO since we've been doing this because they sign multi-year contracts because they want commitment and capacity. In the case of the largest one to date we've announced, they also had requirements that would help our platform also serve them better. The work we're doing with them would serve others better as well, that's why we said it'll take about a year before it has a meaningful contribution to our revenue. It's a scale game. The gross margin there won't be as attractive as a self-serve customer that just signs up and uses it. However, the sales and marketing as a % of revenue will be lower because they own the customer acquisition cost.
To date, most of them have been a white label solution. The neoclouds we've signed up, if you go to their website, you wouldn't know whether it's us or them directly providing it. Eventually, that could change. They don't all need to be white label. I think some will seek to put us in there, some could eventually turn to a sell with versus a sell to. Today, they've all been a sell to, meaning they're committing to the capacity, they're buying it, and they commit to paying that capacity to us, and they sell that capacity. We could do eventually more co-marketing arrangement or sell with them. There's a lot that could happen there, but it's giving us access to a lot more of the larger customers that have intense AI workloads.
Okay. I'm going to pause there and open it up to the floor if there's any questions from the audience.
I'm assuming you provide object storage here, right, with B2 platform. Most of the hyperscalers also provide that. How do you compete against them? Because Amazon, which has S3, the largest object storage there, what's the competition look like there?
Yeah. Just so I repeat it since it's live webcast, you're asking how do we compete with a hyperscaler, given what we offer is S3-compatible object storage. That's exactly what we offer, by the way. It's fully S3-compatible. If you Google, cloud tech engineers' worst nightmare is surprise billings. There's something called the egress. One of the most expensive part of this is not just the price you pay for the storage, but the price to move it around and all the surprise billing. We take a lot of pride in we have a very simple billing model. We bill $6.95, almost $7 a month per terabyte storage, and we rarely bill for egress because we allow a lot more egress to move around, and that's included in the price. It ends up being about 80% less than the hyperscalers.
For a lot of cloud-native companies that are looking to optimize their cost, you hear these days all over the news, not all tokens are equal, right? There's going to be tokens that have to be way more affordable than others. A lot of them are seeking to optimize their cost structure. There's this whole field called FinOps. There's softwares that analyze your cost. There's consulting firms that analyze it. It's very attractive if you could buy something for 80% less. That's how we win a lot of that business today. We publish performance metrics of us and all the hyperscalers, and it's available on our site. We publish it every quarter. The performance metrics at par. It's not like you're sacrificing performance to get a better deal. You're actually getting a better deal, and you're getting great, really, really solid performance. Go ahead.
What is it that lets you do same performance at 80% lower cost? What advantage do you have relative to the hyperscalers?
Yeah. Backblaze, like I said,
Could you repeat the question?
Yeah. He asked how do we get that competitive advantage into our pricing structure and model? I say there's a few things. We've been doing this for a while, and the software we operate is really smart in how it chops up the data, compresses it, how much you put in drives versus somewhere else. A lot of know-how there, and Backblaze actually created the first Storage Pod, right? Before, you had to always buy a server with storage. We're innovators in even how the hardware is structured in the software layer, and we've been publishing, we're the leading publishing organization around drive failure rates. Even the drive manufacturers love talking to us because we keep publishing these things quarterly.
Lots of intellectual property built into the software layer, in how you move around data and cut it up and compress it and encrypt it and so on. Sorry, you had a question?
Yeah. Maybe a follow-up to a couple of these questions as well. How long do you think you'll be able to have that pricing competitive advantage for? Also, another aspect to that would be, when you took price, how much more would you have, what was the customer response to that in terms of being, maybe under 80% less than AWS?
Yeah. Good question. Chet asked how long do we feel we'll have that competitive advantage in the pricing, and how would market react to price increases? I'll answer the first one. If one of those big companies is to take an 80% cut on how much they're billing for this stuff, they'd wipe out their operating profit, right? People have been talking about it for a long time, 10 years, never happened. Right? I think we're safe on that front. We also offer a lot of simplicity, right? With that simplicity comes other things and features that companies like. On the price increase, we did the price increase, it was effective May 1st. Today, things are going really well. We haven't seen any reaction to the marketplace. We announced it in March.
Online, if you read the people around it and read it, what they said, they largely said no surprise. It's expected, right? The cost of hardware is going up, and even Jensen at GTC said, "Listen, the new choke point for the AI world is no longer the GPUs. It's the capacity to store the data." I think the market's taking it really well, and a lot of the smaller companies that operate in this field, actually, within a month or two, followed suit and also took up their prices. It's good. We were the market leader there, and we haven't seen any detriment yet. Sorry, let me take that one because he's had his hand up a few times. Go ahead.
Marc, second question on flash memory. You mentioned it's up 10x. There's obviously a big memory crisis right now. Drive memory is only up 50%. Obviously, they're not dropping replacements for each other. They're different use cases. I was wondering from the perspective of the neocloud, to what extent they can use drives instead of flash in use cases like that?
Yeah. The question asked is how much could neocloud rely on drives versus flash, right? What I would say is out of the gate, when you launch as a neocloud, there was a big race to just, "Let's just build up." It's a lot easier to build up with flash memory. The reason why is the software layer that interacts with flash memory is a lot simpler because you don't have to think through the physical limitation of how do you communicate with thousands of drives. What's called the IOPS just operates way more smoothly with flash memory. Once you get into drives, everybody would like to operate with drives, but there's a lot more complexity there in how do you interface with drives to take data in and out. That's where we're seeing a huge competitive advantage for us.
If you're a neocloud provider, you have to operate with both, and I'll tell you why. The average price for flash storage per month is, I'd say, in the $80-$120 range per terabyte of storage. The average price for warm storage, I would say the market is between $12-$25. If you're a neocloud provider, you have to be able to offer both, or you won't be able to attract customers that would otherwise be leaving the hyperscalers. They need to offer it both ways. If you're going to offer the warm layer using flash memory, it's not going to be very profitable. Okay. Go ahead.
Can you just describe the length of the sales cycle for the neocloud? Today, I don't suppose much, but how is it? How long is it varied? Also, you mentioned you're working with, sorry, you said there are 200 neoclouds. How many are you with? If any, all of them, or a subset of that?
The question was, sales cycle for neocloud, and how many we're working out of with the 200. We've already signed up four, and we're probably talking with most of the top ones. The sales cycle is longer. It's a way more strategic sales cycle. Obviously, they got a lot at stake in making sure they drive the right customer experience from their side. They're going through a lot more validation. The testing is more extensive. Sometimes, there's a lot more engineering and cloud operations involved in the sales process to make sure we meet their needs. Out of the 200, we're off to a great start. We just announced this earlier in the year, and we already have four signed up, and we're talking with most of the large ones. The sales cycle is definitely going to be longer.
The really large one with the $15 million, there's a minute left, right before she is going to tell me that soon. The really large one we announced took a year. There's a lot of discussions. I would say the other ones, they'll range from three to six month. Longer than a startup who shows up and within 10 days is ready to sign up for a million-dollar spend. Longer than an enterprise classic sale, which could take 60 to 90 days. I'd say that's the average, three to six month. All right, one more question. Go ahead.
You got this big neocloud opportunity in front of you. What capacity do you have to serve them, and do you have concern about the pricing and delivery prices? Like, does that cause you to bring on additional capacity at a higher cost?
Yeah. The question is, how much capacity do we have, and are we concerned about the cost? Number one, we do have a lot of capacity. We list on our website, we're operating a multi-exabyte scale platform. There's not a lot of them. We got great relationship with the vendors, and we have accelerated CapEx to be able to fuel the growth based on the demand signals we're seeing. The price increases do impact our margins, but when you look at the recent price increase we just did compared to how much the CapEx is on depreciation as a percent of revenue, you'll see that it'll be margin neutral.
All right.
All right.
Okay. Well, Marc, thank you so much for entertaining our questions during the fireside.
Thank you, Eric.
Let's get on to your next one.
All right. Great to meet you all.