Backblaze, Inc.BLZENASDAQ
Loading
Operator: Good day, everyone. Welcome to the Backblaze Fourth Quarter and Full Year 2025 Earnings Call. Just a reminder, this call is being recorded. I would now like to hand the call over to Ms. Mimi Kong. Please go ahead.
Mimi Kong: Thank you. Good morning, and welcome to Backblaze's Fourth Quarter and Full Year 2025 Earnings Call. On the call with me today are Gleb Budman Co-Founder, CEO and Chairperson of the Board; and Marc Suidan, Chief Financial Officer. Today, Backblaze will discuss the financial results that were distributed earlier. Statements on this call include forward-looking statements about our future financial results, the impact of our go-to-market transformation, sales and marketing initiatives, cost savings initiatives, results from new features, the impact of price changes, our ability to compete effectively and manage our growth and our strategy to acquire new customers, retain and expand our business with existing customers. These statements are subject to risks and uncertainties that could cause actual results to differ materially, including those described in our risk factors that are included in our quarterly report on Form 10-Q and our other financial filings. You should not rely on our forward-looking statements as predictions of future events. All forward-looking statements that we make on this call are based on assumptions and beliefs as of today, and we undertake no obligation to update them, except as required by law. Our discussion today will include non-GAAP financial measures. These non-GAAP measures should be considered in addition to and not as a substitute for our GAAP results. Reconciliation of GAAP to non-GAAP results may be found in our earnings release, which was furnished with our Form 8-K filed today with the SEC. You can also find a slide presentation related to our comments in the webcast, which will also be posted to our Investor Relations page after the call. Please also see our press release or presentation for definitions of additional metrics such as NRR, gross customer retention rate and adjusted free cash flows. And finally, we will be participating in the Citizens Technology Conference on March 2 in San Francisco. Thank you for joining us, and I would now like to turn the call over to Gleb.
Gleb Budman: Thank you, Mimi, and welcome, everyone, to the call. We finished 2025 with solid fourth quarter results. Revenue came in line with guidance and adjusted EBITDA margin reached 28%, doubling over the prior year. We also delivered adjusted free cash flow profitability for the first time as a public company, a major milestone demonstrating the inherent operating leverage in our business model. For the full year, total company revenue grew 14% year-over-year with B2 Cloud Storage growing 26%. Today, I want to focus on 3 things: first, the strength and durability of our core business; second, an update on the meaningful progress of our go-to-market transformation; and third, how we're positioning Backblaze to take advantage of the AI opportunity. Let me start with the core of our business. As data creation accelerates exponentially, Backblaze addresses a large and growing market where long-term demand for scalable, cost-effective storage compounds over time. Our business compounds within that market as we add new customers and retain them for an average of 9 years. B2 net revenue retention of 111% reflects consistent expansion within our installed base, reinforcing durable, long-term growth. We've proven our ability to grow in that market, delivering an annualized growth rate of 21% since IPO, being a cash generating business is an important financial milestone. Year-over-year, we meaningfully improved profitability, demonstrating how we are building a sustainably durable company, one that can invest in growth while maintaining financial strength. Now let me talk about our investment in growth and the progress on our go-to-market transformation. While we didn't achieve our budgeted Q4 B2 growth rate, we made meaningful progress and have positioned ourselves for success. More importantly, the underlying fundamentals of the business remain stable and the investments we've made position us for durable growth going forward. Excluding the highly variable growth of the large AI customer we previously mentioned, we stabilized on a baseline of around 20% B2 revenue growth in each of the last 5 quarters. Now we've shared our goal of moving up market. We ended the year with 168 customers generating more than $50,000 in ARR each, up 35% year-on-year. The ARR of this cohort increased 73% year-on-year to $26 million of ARR. We're very proud of this upmarket progress. We've also launched 3 key initiatives: number one, increasing awareness. We launched Flamethrower, our start-up program designed to engage high-growth companies early and establish Backblaze as their long-term storage infrastructure partner; number two, driving greater pipeline consistency. We're upgrading our top-of-funnel systems and scaling demand generation programs to drive higher velocity sales motion; number three, expanding revenue within our installed base. We are implementing processes to proactively identify and capture additional share of wallet across our more than 119,000 B2 customers. People are the cornerstone of our success, and we continue to strengthen our leadership bench to support these initiatives. We have already hired the co-founder of an edge compute company to drive our Flamethrower program, a business systems leader for our systems work and a head of customer success to build out that expansion effort. We will keep up-leveling our leadership and talent. For instance, we are also in the final stages of hiring a sales development leader to drive pipeline and a revenue operations leader to drive tighter coordination and accountability across the entire go-to-market organization. Scaling into this next phase requires even greater execution discipline. To support that, Elias Mendoza joined us as strategic transformation leader. He previously served as partner and COO at private equity firm, Sirius Capital and held leadership roles at IBM and Morgan Stanley. In these roles, he has helped companies drive strong strategy to execution. Under his leadership, we also established a go-to-market advisory committee of operators who have scaled enterprise and platform businesses to a billing and revenue and beyond at companies such as Okta, Snowflake, ZoomInfo and Carta. Their role is to bring pattern recognition, pressure test key decisions and provide external perspective as we scale. We have made meaningful progress in our go-to-market transformation, and I'm excited about the team we're putting in place to drive it forward. Now let's talk about how we're positioning Backblaze to take advantage of the massive AI opportunity ahead. We all understand there's a lot happening in AI today, but sometimes the scale is still hard to fully comprehend. I saw a report recently that capital spending on AI as a percent of GDP by just the hyperscalers in 2026 is forecast to be 5x larger than the entire spend to create the U.S. interstate system, 10x larger than the Apollo Space Program. AI CapEx spending accounted for 92% of all U.S. GDP growth. It's hard to hyperbolize AI. With AI, a big focus is who's disrupting and who's getting disrupted. We believe Backblaze is one of the disruptors, participating in this infrastructure replatforming as a storage backbone for the next wave of cloud infrastructure. So while like any major new innovation, there will be market volatility. We are firm believers in the long-term growth opportunity and are leaning into it. We're doing that with 2 growth vectors. Number one, on the supply side of AI, neoclouds and other AI tooling companies are building the platforms for AI workflows. Our opportunity is to be the storage backbone of those platforms. And number two, on the demand side of AI, companies are using AI to build everything from anomaly detection to zonal forecasting. These companies are using and generating large data sets. Our opportunity is to be the storage of choice for their developers and use cases. And we are uniquely positioned to be the glue between these, creating a virtuous cycle. Developing a platform that can deliver massive performance with large-scale data sets while providing that cost efficiently is a significant technical challenge. Backblaze has done that, and AI is driving an increasing need for this technology. On the supply side, roughly 200 neoclouds have sprung up, and industry estimates project that market to reach $237 billion within the next 5 years. These companies provide GPUs as a Service. And most will need cloud storage to fully serve with their customers. We've already signed multiple of these multibillion-dollar neoclouds with not only 6- and 7-figure deals but our company's first 8-figure TCV deal, an over $15 million deal. And we believe all of these have material upside potential, and we're in discussion with half a dozen others. By our estimates, neocloud storage for our solution alone represents a $14 billion opportunity by 2030. To further pursue our neocloud opportunity, this morning, we launched B2 Neo, a high-performance white label storage offering specifically designed for neoclouds. Developed in collaboration with our neocloud customers, B2 Neo allows neoclouds to offer a top-tier storage solution without the massive capital costs or years of engineering required to build a storage back end from scratch. On the demand side, the growth in AI developers is exponential. GitHub disclosed they were adding, on average, a new developer every second. Hundreds of AI companies and countless individual AI developers already use B2. For example, one of our customers uses AI to generate audio. They just launched a year ago and already have multiple petabytes with us signing a 6-figure annual deal with us. As they add new users and those users generate more audio, that data grows exponentially. Our self-serve platform, where we added 12,000 customers this year alone is a great enabler for this class of AI developers who just want to get going. We launched our start-up program called Flamethrower and a developer relations initiative to ensure developers are building with Backblaze. To drive our road map forward for the AI opportunity ahead, we strengthened our product and engineering leadership. Dan Spraggins joined as SVP of Engineering, and Rhett Dillingham as SVP of Product, bringing deep experience in AI and high-performance cloud infrastructure. We also added Russ Artzt, Co-Founder and former Head of R&D at Computer Associates, as an adviser. Together, this team strengthens our ability to scale the platform for larger, more complex AI-driven deployments. We entered 2026 with a strong and growing business, a rapidly improving go-to-market motion and a tremendous AI opportunity with a targeted B2 Neo offering and a strong product team. AI is reshaping how data is created and scaled and storage sits at the center of that transformation. Across neocloud platforms and AI native developers, we are building the foundation for the next generation of data infrastructure. Durable growth and massive AI potential are the hallmarks of our opportunity. With that, I'll turn the call over to Marc. Marc?
Marc Suidan: Thank you, Gleb, and good afternoon, everyone. We grew revenue while achieving adjusted free cash flow profitability in Q4. This is a significant milestone and an important step forward in our profitability journey. This progress was not driven by short-term cost actions but by the inherent leverage in our operating model as revenue scales. For the quarter, total revenue was in line with guidance at $37.8 million and adjusted EBITDA exceeded the high end of our guidance by approximately 600 basis points. In the fourth quarter, B2 revenue grew 24% year-over-year, up from 22% in the prior year. This is modestly below the range that we outlined last quarter. We delivered record bookings this quarter. As Gleb noted, we closed our largest contract in the company's history with over $15 million in total contract value. We're excited about this 8-figure deal. This deal validates the product market fit at scale. We don't expect to see meaningful revenue in 2026 as we complete certain development work. In 2027, we expect this customer to contribute over 300 basis points to B2 revenue growth. This customer helped drive our RPO, up 60% year-over-year to $66 million. In-quarter B2 NRR was 111% compared to 116% in the prior quarter. The sequential decline reflects variability from the large customer that we mentioned in our past 2 earnings calls. Factoring out that one customer, the underlying retention and expansion trends remain stable. Moving to the income statement. Q4 gross margin was 62%, flat sequentially and up from 55% in the same period last year. Adjusted gross margin was 80% compared to 78% last year. Margins remained stable despite higher data center costs, reflecting continued efficiency in our infrastructure and disciplined management of our operating model. Looking ahead, we anticipate some pressure on gross margins driven by increased costs. In response, we are proactively launching a gross margin optimization initiative focused on structural improvements across pricing, packaging and infrastructure. Our Q4 adjusted EBITDA margin was 28%, doubling year-over-year. The adjusted EBITDA outperformance was primarily driven by nonrecurring items, including variable compensation alignment and office restructuring savings. Excluding those onetime items, adjusted EBITDA would still have been above the 22% high end of our guidance. Adjusted free cash flow was positive $4 million in the quarter, representing a margin of 11%, exceeding our outlook of being adjusted free cash flow neutral. We ended the quarter with $51 million in cash and marketable securities. Based on our current operating plan, we expect to fund our growth through operating cash flows and capital leases. We do not anticipate a need to raise additional capital. We will continue to evaluate opportunities to optimize our capital structure over time in a disciplined manner. To improve accountability and further align management incentives with shareholders, we are shifting part of the compensation to performance-based stock units. These awards are tied to clearly defined performance objectives. Turning to our guidance for the year. Our objective is to provide a clear incredible baseline that reflects the most predictable portions of our business. While pipeline activity remains healthy, larger customer wins in usage-driven workloads can introduce variability in timing and revenue recognition. To maintain forecast discipline, we have derisked our outlook by excluding large swing deals and anchoring guidance on opportunities with more predictable demand characteristics. For our customers with high variable usage patterns, our assumptions reflect contractual minimum commitments rather than potential upside consumption. Our outlook is, therefore, based on continued expansion within our existing customer base and steady adoption of B2 across core use cases consistent with recent operating trends. We believe this approach provides a prudent and reliable foundation for the year, while preserving upside as deployment timing and usage visibility improve. For the first quarter of 2026, we expect revenue to be in the range of $37.6 million to $38 million, with adjusted EBITDA margins in the range of 18% to 20%. For the full year, we expect revenue to be in the range of $156.5 million to $158.5 million. Full year adjusted EBITDA margins are expected to be 19% to 21%. We expect adjusted free cash flows to be roughly neutral for the year with normal quarterly variability. Due to the difficult comp from last year's large variable customer, we expect B2 year-over-year growth in Q2 and Q3 to be in the range of 12% to 19% and approximately 20% for the full year. To wrap up, over the past year, we made meaningful progress towards becoming a Rule of 40 company, with our combined B2 revenue growth and free cash flow margin improving from 9% to 35%. As we look toward 2027 and beyond, we believe Backblaze is well positioned to grow efficiently. Our platform is already built. Our infrastructure scales with discipline and incremental revenue increasingly translates into profitability and cash generation. This capital-efficient model allows us to pursue the massive AI-driven opportunity ahead while maintaining financial discipline, expanding margins over time and building a durable self-funding business. With that, operator, let's open it up for questions.
Operator: [Operator Instructions] We'll take the first question from Ittai Kidron from Oppenheimer.
Ittai Kidron: Solid numbers, and thank you very much for derisking the outlook for the year. It's hopefully a very smart move. Gleb, I wanted to dig of course, into the neoclouds and the large deal. First of all, just from a big picture standpoint, are demand patterns any different? Can you explain how Neo -- your B2 Neo cloud solution, how is it different than B2? And what way are the demands different, the pricing different, the margin different? And if you could elaborate also why this deal is going to take a year before we started seeing revenue, I would appreciate that.
Gleb Budman: Yes. Thanks, Ittai. All good questions. So one thing I'll say, first of all, is our pursuit of the neoclouds is one part of the business pursuit. There are about 200 of these neoclouds. We do think it's a large and important opportunity for us, right? The -- just our part of the neocloud opportunity, we view as about $14 billion, so it's important. And we are really well suited for it. The hyperscalers are not key competitors here because they are competing with the neoclouds as opposed to being vendors for them, the way that we are. So it's a good opportunity, which we're well positioned for. In terms of what B2 Neo is, it is a white label offering. So B2 is generally sold directly to the end customer. B2 Neo is a white label offering that they can build in directly into their service. It is -- it provides a lot of the same functionality that B2 provides. It's high performance. It's low cost. It's durable. It's scalable. But it also provides them the ability to manage that storage on behalf of their customers through APIs, with API integration, single sign-on, et cetera. So it's really leveraging all of the technology that we've built over the last 17 or so years for the company and then layering on top of that technology to make it simpler for them to integrate natively and make it easy for them to manage and offer that storage [ upmarket ]. So that's what B2 Neo is. Now in terms of why it's going to take a year for this one neocloud provider to start seeing the benefits of it, it's a combination of work we need to do and work they need to do. So they have an existing storage offering that they're going to be switching to use B2 Neo instead. And so it's, basically, we have some work to do to make it so that it's even easier and more robust to automate and natively integrate for them. One thing I'd like to make clear is all the work that we're doing for them is useful for other neocloud providers and also other companies but not required for most. So we have multiple new clouds that have already signed up that don't need this work, and we think that there's a large number of them that won't need any of this work. But the work that we're doing is broadly useful for others as well.
Ittai Kidron: Okay. Appreciate it. And then, I guess, first of all, the TCV, $50 million, that's great. But can you tell us the duration of the contract? And is the margin profile of this business -- as you ramp up the neocloud, Gleb, is there potential upfront costs hit to you as they ramp before margin normalizes on these businesses?
Marc Suidan: Yes, I mean, Ittai, this is Marc. I can take that question. We do have to accelerate some capital expenditures. That would impact that and other things happening in the market would impact our gross margin by a few hundred basis points to help us prepare for this because it's obviously a large deal. You need to have the capacity in place.
Ittai Kidron: Okay. And then lastly, on computer backup, Marc, can you comment on the expectation? I mean this business is -- the number of customers is now declining here. I guess help me think about the framework for this business for '26. How do I -- how should I think about the quarterly cadence and the annual cadence of this business? Is there a different long-term outlook for this?
Marc Suidan: Yes. I mean, I'll start off by the coming year, Ittai. We see this business declining 5% year-over-year. Currently, in Q1, that's more like a minus 3%. That builds up throughout the year and makes an -- averages out for the end of the year at a minus 5%.
Ittai Kidron: Okay. And longer term, is there any -- should we just continue to expect this business to slowly decline?
Gleb Budman: What I would say you, Ittai, on that one is we have programs that we've put in place and are putting in place to stabilize the business. We would like to get it to a place where it is flat and possibly even slowly growing. We don't think this is a fast-growth business, as you know, but it would be good for it to not be a declining business. But it's a little too early for us to have confidence in those programs getting to that place. So for this point, we're estimating it at that shrinking rate, but we are putting effort into getting that to be flat to slightly growing.
Operator: The next question will come from Jeff Van Rhee from Craig-Hallum Capital Group.
Jeff Van Rhee: Congrats on the free cash flow. Great to see it. A couple for me. Maybe if you could just start in terms of B2 coming into Q4, came in a bit below expectations. Just expand a bit more on what missed there. And then as you're looking at the annual number, I didn't catch what you had guided it for in Q1, so if you could just fill in the gap. I think we can back into it, but maybe you could just share it. So what happened in Q4? And what do you think in Q1?
Marc Suidan: Yes, Jeff, good to hear from you. This is Marc. So on the Q4 '25, we were expecting, when we set our guide, quite a few deals to close in November. They came in very late in the quarter, so they didn't benefit Q4. That's why we've adjusted our guidance philosophy going forward, where we said, going forward, we're going to factor out the swing deals because they're less predictable in timing of closing. So that feeds into the guide going forward. And we said for B2 year-over-year, it will be 20% in 2026. The ranges that we provided of 12% to 19%, a lot of that has to do with the comps of that high variable customer in 2025, so Q2 would be the low end of that range, and Q3 would be about the higher end of that range. And overall, the year would average out to 20%. Does that answer your question?
Jeff Van Rhee: Yes, I think it does. And so the growth is, if I do the quick math, maybe in Q1 looks like it's 9%, if I have it right, on year-over-year and you're decelerating to 8% for the overall year. So it actually looks like maybe you're assuming some deceleration in the year. I'm sure there's a little bit of lumpiness from the large customer, but generally speaking, you had some pretty good momentum in sort of Phase 1 of the sales build and build out. And it sounded like you felt like you had some early good signs on Phase 2, but the numbers are painting a picture of deceleration. So just help me reconcile the 2.
Marc Suidan: Yes. I mean the deceleration that you're seeing is largely driven by that one monthly customer. If you go to Slide 21 of our earnings deck and you factor out that one customer, you could see that pretty much we've been stable around the low 20s. So if you recall, factoring out any price increase, B2 growth rate has always been growing but decelerating for 5 years. We've managed to stabilize it in the low 20s. So now with this new guidance philosophy, we're seeing 20% year-over-year, and that includes the lumpiness that I described in Q2 and Q3. But -- then with all the Phase 2 changes we're doing, Gleb could elaborate on that. That will then afterwards come drive benefits.
Gleb Budman: I think also, Jeff, I think you were talking about the whole company, not just B2, right? And so part of what's driving that is that computer backup was growing, in part, due to price increase before and it's -- as Marc said, we expect it to shrink about 3%. So it's putting some downward pressure on the overall company in Q1. But on the GTM transformation, I think some of the things that we look at is, in terms of progress, there is progress that we're making in terms of actions, things like we've hired the VP of Revenue Operations. We've made material progress in moving the systems forward and expect that work to be largely completed at the end of this quarter. We've gotten pretty far down the path with some sales development leaders to bring in. We've made a number of kind of improvements. And then you can also see some of the outcomes like the 73% growth in ARR from customers over $50,000 and this 8-figure deals. So I think we've made progress on the GTM side. Obviously, we all want more work to be done there.
Jeff Van Rhee: Great. Maybe just one last one if I could. On the large neocloud win, can you just expand a bit on what the competitive landscape looked like there, maybe the finalists, the kind of 2 or 3 that it came down to at the end of the day and if there were specific features, capabilities that were the deciding factors for your win there?
Gleb Budman: Yes, it's actually -- it's interesting because this neocloud, they have their own storage. They started realizing from their customers that the source that they had wasn't going to provide what they needed for this next phase of evolution. And so they started thinking about how to handle that. A number of their internal engineering and business leaders were actually familiar with Backblaze from prior roles in other places, and they knew that Backblaze had a really strong reputation for providing a great storage platform, that it was trusted. Basically, we built a moat around this idea of high performance but predictable economics and low-cost storage. And so we were at the top of their list for consideration. Now when they went and evaluated, they wanted to make sure -- because they were going to be basically placing their brand on the line for saying they're going to use us for this underlying platform for all of their customers, so they wanted to make sure it absolutely worked. They did pretty detailed technical due diligence and then chose us. So the why, I think, came in part because we had established a lot of credibility over many years that we are a great storage platform, and then we met their technical requirements for both performance, scale, affordability and openness.
Operator: Your next question today comes from Mike Cikos from Needham.
Michael Cikos: If I could just come back to the gross margin comment, this expected headwind that we're up against, I guess it's a bit of a two-parter here. But when I think about the headwind we're facing this year, is that really tied to customer success initiatives or deployment in advance of recognizing revenue from this large neocloud agreement that we're talking to today? Or is there potentially an ongoing presence or multiyear factor we need to consider when evaluating corporate gross margins on a go-forward basis?
Marc Suidan: Yes. Mike, it's Marc. There's a few factors in there, right? First of all, data center cost and equipment have gone up. That, combined with us needing to accelerate some CapEx, does reduce our gross margin this coming year by a few hundred basis points. That's why we said we're doing that gross margin optimization initiative to look for opportunities to offset that. Now in terms of business model, when you go after a white label, large-scale solution like that, generally speaking, the gross margin will be a bit lower and the OpEx will be lower as well because you have to spend less on sales and marketing. So it nets out to the same economic model for us, but that's the P&L benefit if that makes sense.
Michael Cikos: It does. It does. And then I just wanted to come back again to this derisk guide that we're talking to here. And appreciate the commentary in the prepared remarks. But just to better understand, these swing factor deals or the idea that we're only going to underwrite minimum contract commitments from customers, is that really tied to the neoclouds when thinking about those swing factor deals? Or is it maybe the move upmarket? Anything else you can provide that's creating that dynamic? And then second -- go ahead. Go ahead. And I just have a follow-up.
Marc Suidan: Look -- okay. I'll answer this one and then you could ask your next question if you want. So moving upmarket -- I mean there's different sides of upmarkets. But when you look at the average deal size of those 168 customers, it has grown quite a bit. But I think the even larger ones, and let's call larger ones $500,000 in ARR and greater, they do take longer to close. So there's less predictability for us to factor that into our guide. So that's why we factored them out. Doesn't mean they won't happen. It's just harder for us to guide on them. So I think it's less around the neocloud. I mean the neoclouds are big deals, too, and they have similar attributes, right, where you got to take longer to do the technical feasibility and make sure you went over the POCs. So that's what's driving that side of it.
Michael Cikos: And then, I guess, the final follow-up on my side, but for those, let's say, $0.5 million plus deals that you're signing, can we start bifurcating the extent to which those sales cycles are longer versus a more typical run rate business? And then final piece, but for the calendar '26 guide, is there any way you can give us some pointers as far as the NRR that you're thinking about when we look at this calendar '26 guide? And that's all on my side.
Gleb Budman: Mike, in terms of the bifurcating the size of the deals and the length of time, when we look at those, they certainly are longer sales cycle ones, but it's interesting because they're not dramatically longer. So some deals like the 8-figure deal that we talked about, that did take the better part of a year, in part, because they had to look through their own systems. They had to understand what it would take to switch out to a different system, what integration that would require, et cetera. A number of the other neoclouds didn't take anywhere near that long, and many of the other larger customers, especially ones that are $50,000, $100,000, $200,000 actually moved quite quickly. But certainly some of the largest of those deals, they did take, call it -- so some of them took 6 months or so to close, whereas we've seen a lot of the deals close in sub-90 days.
Marc Suidan: Yes. And then I could jump in and discuss the NRR outlook. Due to the lumpiness of that large customer in '25, we factored out any usage above their minimum commitment level and our guide for '26. So assuming that that's what materializes, the NRR, just like the revenue growth rate for B2 and just like the overall growth rate at the company will be lower in Q2 and Q3. NRR could go down to closer to 100% for 1 or 2 quarters. But our overall growth rate of 20%, which is where we should be finishing the year and year-over-year overall should equate to an NRR that's closer to 110%, so pretty much where we are now, plus or minus 200, 300 basis points.
Gleb Budman: Mike, one thing actually, I'll mention also on NRR that I think I find quite exciting. The -- we have a broad base of customers, but we're leaning in heavier to the overall AI customer type, not just the neoclouds. And we have hundreds of those customers that are using us for AI workflows specifically. We've seen a growth rate of 75% in the number of those AI customers. But one of the things that I find even more exciting is that the growth rate of those customers is about 3x faster than the growth rate of our average customer. So as we sign up more of these AI customers, we see the opportunity for NRR to go up over time as well because they are generating data at a faster rate than your average customer.
Operator: Up next, we'll go to Jason Ader from William Blair.
Jason Ader: Wanted to first ask about your comment, Gleb, that most neoclouds don't have storage. I think that's what you said. I just wanted to understand why that might be. And then also your comment that the 8-figure win was with the neocloud that did have storage, but the storage wasn't going to handle what they needed, maybe just if you could elaborate on why it wouldn't be able to handle what other customers needed.
Gleb Budman: Yes. Thanks, Jason. Both good questions. So with these 200 neoclouds that are -- that have come up, they almost all started with GPUs, right? So the need that happened was for these AI use cases, they needed the GPUs first. The second thing that they need is they need a place to keep the data to feed these GPUs. So initially, they set up data centers. A lot of them set up data centers that were more specifically designed for GPUs, which are very power hungry. Oftentimes, they want liquid-cooled environment. They don't need nearly the square footage in the data centers that they need. They need more power in the space, et cetera. So they built these providers focused on the GPU opportunity. What they realized then is customers who want to use the GPUs need a place to keep the data. They needed the place to keep their data to build the models. And then they needed the place to keep the data when they're doing inferencing for the outputs. And so what some of them have done -- many of them have not done anything on that front yet. They've just stood up the GPU side of things. But what some of them have done is said, okay, well, we can do something, and they -- some of them have used open source projects for -- to stand up their own infrastructure, where some of them have set up a storage infrastructure using flash systems. The problem is what they found is the flash systems are incredibly expensive to operate. And so for large-scale data sets that becomes very quickly unaffordable. The open source tooling is difficult to manage. You have to have experts ongoingly working to tune it, operate it, et cetera, and they're really not designed to scale to exabyte scale. Most of those open source projects were designed for potentially handling a single enterprise's scale. And so once they started seeing some movements and success, they start reaching the limitations of those projects. So the opportunity for us is that there are these 200 providers. They've built up the GPUs. They're starting to realize that they need storage. They're not going to get that from the hyperscalers for the most part because those are their direct competitors and the solutions that they have are either really expensive, really complicated or don't scale.
Jason Ader: Got you. Okay. And then the neocloud that you announced or that you talked about, the 8-figure one, can you say if that is a publicly traded company?
Gleb Budman: They are a publicly traded company, yes.
Jason Ader: They are. Okay. Great. And then last one for me. Just, Gleb, what's your confidence level that you could win additional deals like the one that you announced on the call today?
Gleb Budman: I mean, I'm very confident that we can do additional deals. The timing is obviously always uncertain, but this is -- it's not like this neocloud is the only neocloud that we have won. We've got others that are 6 figures and 7 figures already. Those that we have already signed at 6- and 7-figure deals, I think they themselves have the opportunity to become 8-figure deals because, as they roll this out to more of their customers and more scale, they're big enough that they could become 8-figure deals for us themselves. And we're currently in discussions with about half a dozen other neocloud providers that are somewhere in this same scale of size of organizational opportunity. So timing is obviously a question for us, but our ability to be a good fit for these kind of customers and the discussions we're in give me a lot of confidence.
Jason Ader: And I may have missed it, but did you say how the duration of that 8-figure win was?
Gleb Budman: That one's a 3-year deal.
Jason Ader: Three-year deal. Okay.
Operator: Eric Martinuzzi from Lake Street Capital Partners has the next question.
Eric Martinuzzi: Yes. You mentioned the revenue impact from the 8-figure transaction really doesn't start to hit until 2027. Is that -- based on your answer about the 3-year duration and over $50 million, is that to say then that we're a small amount, maybe the end of 2026 and the bulk of it split between '27 and '28?
Marc Suidan: Yes, that's correct, Eric. And for now, honestly, we're not factoring anything into 2026 for that.
Gleb Budman: By the way, Eric, you said -- I just want to make sure that -- it sounds like you said $50 million, it's $15-plus million, 1-5.
Eric Martinuzzi: Got you.
Gleb Budman: I look forward to a $50 million deal in the future, but we're not there just yet.
Eric Martinuzzi: The other thing I wanted to ask about was your comment regarding the adjusted free cash flow. You talked about it being neutral for the year. And I'm just wondering, given the investments you're making to have the infrastructure in place here, it seems like it's sort of front-half loaded. Is that to suggest then that the adjusted free cash flow positive, we're Q4 for sure and potentially Q3? Is that the right way to think about it quarter-by-quarter?
Marc Suidan: Yes, Eric. I mean, generally speaking, the first half of the year is our cost base increases. It starts kicking into Q1. And our OpEx lines honestly should not be really increasing that much other than maybe around 500 basis points, not as a percent of revenue, just off the dollar baseline from last year on a non-GAAP basis as it relates to just basic inflation, salary raises and so on. Other than that, we're keeping our OpEx model pretty tight. I spoke about the gross margin being set back by a few hundred basis points. So when you combine all those factors and accelerating some of the expenditures to prepare for these customers, that's why we're free cash flow neutral for 2026. It is lumpy during the year. Usually, Q2 is also where we have the least of our computer backup renewals. So Q2 is usually the worst set, and the second half of the year is in better shape. And that would be a nice improvement from the minus $5 million for 2025 as a year and the minus $20 million in 2024. So I think we're pretty well set on exiting the phase of cash burn, and our aim is to stay here and get better.
Operator: [Operator Instructions] Up next is Zach Cummins from B. Riley Securities.
Ethan Widell: Ethan Widell calling in for Zach Cummins. I guess start with neocloud and with there being a high portion of leverage there to AI and HPC. How would you define, I guess, the incremental revenue opportunity or overlap, whether it be like customer base or function or revenue opportunity versus B2 Overdrive?
Gleb Budman: Yes. Thanks, Ethan. It's a good question. So B2 Overdrive was initially actually developed because we heard from customers saying they wanted to use high-performance storage, high throughput storage that would enable them to send their data to the neoclouds when they needed them or to other hyperscalers, for example. So B2 Overdrive is not a white label offering. It's designed for end customers to actually use themselves. B2 Neo is specifically designed as a white label offering for the neoclouds to them themselves offer storage to customers. So they're largely serving different sides of the market but both serving the needs of AI and HPC type use cases.
Ethan Widell: Understood. That's helpful. And then the large TCV deal, can you clarify whether that was from an existing customer? And generally, is the revenue upside from existing customers there based on increasing usage?
Gleb Budman: So the $15 million-plus TCV deal is a new customer, completely new to us. However, what I would say is if you look across the $1 million-plus deals that we've had over the last year-ish, it's roughly half-half. Half of them are net new customers to us that came in, evaluated, considered, tested and then signed a 7-figure deal with us. And the other half are customers that started off small. Some of them started off self-serve. Some of them came in as just smaller sales deals, got familiar with the platform, liked the platform and then expanded into 7-figure deals.
Marc Suidan: Yes. And Ethan, this is Marc. What I would add, if you look at Slide 17 of the earnings deck, it breaks down the new versus expansion from the existing, and it's certainly half and half. So it's pretty well distributed because the self-serve product-led growth is about half of that as well. And then the larger direct sales customers half. And each one is -- kind of breaks out into a half by itself of what is expansion versus new logo. So it's basically that's why if you look at the stacked bar, it's like 4 quarters, it's pretty well diversified in terms of how it comes through.
Ethan Widell: Got it. Well, I appreciate the color.
Gleb Budman: Yes. Maybe one other piece of color just to add in terms of -- so one of the things we look at is -- as a forward-leading indicator is pipeline. And in 2024, we generated about $15 million of pipeline, and in 2025, we roughly doubled pipeline to about $30 million. Our aim with our continued GTM transformation is to get to a run rate of about double of that. So with our industry-leading win rates, pipeline transfers into ARR quite efficiently. And so we're not there yet, but that's -- we made, I think, meaningful progress in '25 and aim to make more meaningful progress on that in 2026.
Operator: The next question is from Rustam Kanga from Citizens.
Rustam Kanga: Marc and Gleb, congrats on the RPO acceleration. Just building on another question that you answered, Marc -- Gleb, where you kind of mentioned that B2 Overdrive versus B2 Neo are serving 2 different sides of the market. And as we sort of think about the build-out of the pipeline for B2 Neo, is it fair to say that these opportunities are going to be anchored towards larger deals, albeit maybe not as large as this one that you've just put it up in the quarter, but is it fair to say that this is kind of the larger opportunity? And is that likely to sort of lead to higher ASP engagements as you look towards this opportunity?
Gleb Budman: Yes. It's a good question, Russ. So one of the ways I would look at it is the market for the neoclouds, if you take just the hard drive-based storage opportunity inside of those 200 providers, that market is estimated at about $14 billion in the next 5 years. So with 200 players representing $14 billion of opportunity, every single 1 of those deals on average is going to be a large deal. So the short answer to your question is, yes, the B2 Neo deals, we see as large opportunity deals. The ones that we've signed so far are 6- and 7- and now 8-figure opportunities on those. Some of those, I imagine, they start smaller just as they start getting familiar with it, but I think all of them have the opportunity to get quite large.
Rustam Kanga: Great. That's helpful. And then just kind of thinking about the investment cycle for next year, is there any sort of relative color that you can share with us in terms of the level of CapEx investment that you guys are thinking about for '26?
Marc Suidan: Yes, Russ, this is Marc. Good to hear from you. Our CapEx will be higher next year. As a percent of revenue, when you look at our PP&E at the end of the year, it should be in the high 20s percentage of revenue. We typically finance our CapEx through capital leases, and we're fully set up to do that. And that would be the principal lease payments on a statement of cash flows, which is around mid-teens of revenue, right, because you're buying today but financing over 5 years over a growing revenue base. That mid-teens, I mean, over the past few years, has actually improved from our side as we continue to optimize our cost of capital.
Operator: And everyone, at this time, there are no further questions. I would like to hand the conference back to Gleb for any additional or closing remarks.
Gleb Budman: Thank you. We have a strong and durable core business, made meaningful progress in our go-to-market transformation and have a tremendous opportunity in AI. We drove growth while becoming adjusted free cash flow positive. We launched B2 Neo and signed multiple neoclouds, including this $15 million-plus deal. We also launched Flamethrower, our program for high-performance start-ups. In just the last few days since the launch, it's exceeded expectations, growing faster than the kickoffs at other leading companies that are a leader for that has driven. We had about a dozen start-ups that have applied, been evaluated, accepted and given credits, including ones from Andreessen Horowitz and Y Combinator, and we've bolstered our team overall to take advantage of this tremendous opportunity. I'm really excited about the year that we have upcoming together. I want to thank our employees, our customers and our investors for taking this journey with us, and we look forward to chatting with you next quarter. Thank you.
Operator: Once again, everyone, that does conclude today's conference. We would like to thank you all for your participation today. You may now disconnect.