Backblaze, Inc. ($BLZE)

Ladies and gentlemen, thank you for standing by. My name is Abby, and I will be your conference operator today. At this time, I would like to welcome everyone to the Backblaze First Quarter 2026 Earnings Call. [Operator Instructions] Thank you. And I would now like to turn the conference over to Mimi Kong, Head of Investor Relations. You may begin.

Thank you. Good afternoon, and welcome to Backblaze's First Quarter 2026 Earnings Call. On the call with me today are Gleb Budman, Co-Founder, CEO and Chairperson of the Board; and Marc Sudan, 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-saving 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 most recent 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 on 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. We will be participating in the Needham Technology, Media and Consumer Conference on May 12 in New York. I hope to see many of you there. Thank you for joining us, and I would now like to turn the call over to Gleb.

Thank you, Mimi, and thank you, everyone, for joining us today. Q1 was a strong quarter. We beat revenue and adjusted EBITDA guidance, ending the quarter with $38.7 million in revenue, up 12% year-over-year with B2 growing 24%. We more than doubled our average sales deal size and drove 72% year-over-year growth in our $50,000-plus ARR cohort as we continue to move upmarket and are on track for our first full year of free cash flow positivity as a public company. What excites me most about Q1 goes beyond the numbers. AI is making storage increasingly important, and our organization is gelling and executing better than ever to capture that opportunity. This is evidenced by more than 1/3 of all new bookings coming from AI and the number of AI customers using our platform growing by 76% year-over-year. We entered 2026 saying we would build a more scalable, more predictable growth engine that serves the AI opportunity. Q1 started to show what that looks like. In AI, we are seeing demand from 2 parts of the market. One is companies building the infrastructure and tools that enable AI. The other is companies using that infrastructure to bring AI into products and workflows. We are winning in both. On the infrastructure side, there is a major replatforming happening in the market. For the first time in about 2 decades, the traditional hyperscalers are not the only place companies are building. They're also building on the neoclouds. Synergy Research estimates that the neocloud market was $25 billion in 2025 and growing to about $400 billion by 2031. In order for these neoclouds to support their customers' AI workflows, they need to offer cloud storage. Some neoclouds have offered cloud storage built on Flash. It was fast and it worked. But as these platforms have scaled and AI workloads have grown, the economics have become increasingly difficult. Flash is now about 10 times more expensive per terabyte than hard drives. It works well for use cases requiring the lowest latency for smaller data sizes, but it becomes unsustainable at exabyte scale. As a result, neoclouds are now actively looking to introduce a cost-efficient hard drive tier time patch to manage both performance and economics across their infrastructure. At Backblaze, we built an Internet scale file system to optimize performance per dollar out of hard drives and thus believe Backblaze is ideally positioned to provide exactly what these neoclouds need. We've seen support for that belief not only from the multiple signed neoclouds where we provide this for them already, but also the active engagement we're having with many of the top neoclouds. We estimate our opportunity to support neocloud at $14 billion by 2030. And with the success we're seeing, we are aligning resources internally behind that opportunity. In addition to neocloud, we're seeing a significant opportunity for us supporting other AI infrastructure. For example, we are also seeing strong demand from companies supplying large data sets into the AI ecosystem because they need a place to store large data sets efficiently, but also be able to move them where they need to go rapidly. One recent example is a train data provider serving AI use cases that selected B2 to store large volumes of video data. A hypergrowth company, it was experiencing rate limits and bandwidth constraints with its existing provider and needed a solution that could scale quickly. Backblaze won on both economics and technical fit. The deal closed in just 11 days at nearly $1 million of ARR, underscoring how quickly these companies move when infrastructure becomes a constraint and how well Backblaze is suited to the infrastructure side of the AI opportunity. The other part of the AI market we're seeing is companies using infrastructure like ours to bring AI into their products. As AI models move from text to multimodal, incorporating video, audio and images, the volume of data required to train and run those models grows by orders of magnitude. This is not a future trend. It's happening now and is creating significant and growing need for storage that can handle it economically and at scale. With the Generative AI customers we have today, we are finding that price and performance get us in the door. But it is the experience that keeps them and grows them, transparent pricing, responsive support and a team that works with them rather than just selling to them. These customers are scaling fast, and they do not have time to manage infrastructure problems. With Backblaze, they don't have to. A good example from Q1 is an AI-powered video creation company that selected B2 to store data used to train its models. The customer had been running into cost and performance issues with its existing provider. The platform was difficult to manage and the economics were not working at its scale. Backblaze offered the best performance per dollar and a platform that was easy to use and easy to scale. The initial deployment represents nearly $0.5 million of ARR and creates a clear path to expand into higher-performance workloads over time. These customer wins are just examples of where we won in Q1 and are reflective of opportunity we have in pipeline going forward. It's clear that whether customers are building AI infrastructure or using AI in their products, they are scaling fast, their data is growing exponentially and they need infrastructure that is performant, open and cost efficient at scale. That is the moat we have spent 19 years building and AI is making it more valuable, not less. To be the leading storage platform for AI, we are also meeting developers where they already work. We are embedding Backblaze into the AI ecosystem by integrating directly into the tools developers already use. For Hugging Face, which has 13 million users and over 2 million models, we shipped a tool that lets teams store and share model caches on B2. For ComfyUI, which recently raised at a $500 million valuation, we built a plug-in to support Generative AI workflows. For CVAT, which is used by tens of thousands of computer vision teams, B2 is now integrated as a back end for training data. And for MLflow, the most downloaded tool for taking AI projects from lab to production with 16 million monthly downloads, B2 has now been added as an integrated artifact store. So the AI opportunity is making what we do increasingly critical. We're also stepping up to meet it. One year ago, we began a meaningful transformation of our go-to-market organization focused on 3 things: increasing awareness; driving greater pipeline consistency; and expanding revenue within our installed base. In Q1, we delivered progress on all 3. On awareness, the Flamethrower start-up program is gaining real traction. We have now welcomed approximately 100 companies in under 3 months, half the time it would typically take. We've been added to the a16z Founder Resource Program, the Launch Startup Showcase, and the Startup Grind conference, all of which expand our reach with venture-backed start-ups. On pipeline consistency, we have completed our core go-to-market systems upgrade, giving our team better visibility and a stronger foundation for a faster, more disciplined revenue motion. And within our installed base, pipeline sourced from existing customers has nearly doubled year-over-year, reflecting our growing ability to land and expand with our customers. To accelerate this next phase, we welcomed Anuj Kumar as our Chief Revenue Officer. Anuj has scaled go-to-market for cloud infrastructure and enterprise storage at NetApp, VMware, Red Hat and SUSE. He brings the pipeline discipline and execution rigor this phase of our growth requires, and we believe his leadership will be a meaningful complement to the upmarket momentum we have already built. We also saw encouraging new customer momentum during the quarter across a range of data-intensive use cases. That included a health care data company who selected us for disaster recovery, a cloud gaming platform that chose B2 to store video across multi-cloud environments, and an audio streaming platform migrating from self-managed infrastructure to B2. These wins reinforce a broader point. Backblaze is winning where data is valuable, active and operationally important. And this is why I am excited about the opportunity ahead. The shift to multimodal AI is driving exponential data growth and the need for high-performance, yet cost-efficient storage has never been greater. The customers who are choosing Backblaze are exactly the kinds of customers that compound with us over time. We are stepping up to this opportunity with an up-level team, a go-to-market transformation well underway and a platform we have spent nearly 2 decades building and optimizing. AI is making everything we have built more valuable, and we are becoming the storage infrastructure that powers the AI economy. With that, I'll turn it over to Marc.

Thank you, Gleb, and good afternoon, everybody. Our first quarter results reflect the strategy that we have been executing against. We exceeded the top end of both revenue and adjusted EBITDA guidance. The Q1 outperformance reflects stronger sales execution and the EBITDA beat demonstrates the operating leverage in the model. Let me walk through the quarter and then cover our outlook. We finished Q1 with revenue of $38.7 million, above the high end of our guidance of $38 million. The beat was broad-based across both B2 Cloud Storage and Computer Backup, with B2 remaining the primary growth driver. B2 Cloud Storage grew 24% year-over-year to $22.4 million and ARR grew 28% year-over-year, reflecting the underlying strength and momentum of the business. The Q1 revenue outperformance was driven by higher customer data consumption on the B2 Cloud platform and Computer Backup coming in slightly more favorable than our forecasted decline. On bookings, which primarily affect revenue in future quarters, we closed multiple large deals for a strong quarter. We made several updates this quarter to improve the calculations of our ARR and RPO metrics. I will briefly walk through those changes as I cover the results. ARR increased by more than $5 million sequentially to $158 million with B2 growing 28% year-over-year. This quarter, we updated our ARR methodology to improve comparability across periods, and the change is defined in the earnings presentation posted on our Investor Relations website. Under both the new and previous methods, the sequential ARR improvement is approximately $5 million. We ended the quarter with 187 customers contributing over $50,000 in ARR, up 51% from a year ago, reflecting continued strong progress upmarket. We also updated our RPO methodology this quarter and described the change in our earnings presentation. The change is aligned to our peer group and RPO is now a more important metric as we continue to move upmarket, signing both annual and multiyear customer commitments. Under the updated methodology, RPO increased by $6 million sequentially and by $31 million from the prior year period. Our gross customer retention metrics remain very healthy with customers continuing to use both our B2 and Computer Backup solutions for 9 years on average. Beginning this quarter, our reported net revenue retention reflects an in-quarter methodology, which we believe provides a more current view of our customer expansion and retention trends. In B2, net revenue retention was 110%, up from 105% a year ago, reflecting continued expansion within the customer base. As a consumption business, B2 benefits from both the organic customer data growth and the cross-sell, upsell sales motion. Q1 gross margin was 61% versus 56% in the prior year. The year-over-year improvement shows strong operating leverage continuing to kick in as we tightly manage costs and also from the extension of the useful life of our fixed assets. Total operating expenses were $29 million in Q1, roughly flat compared to Q4 and improved by approximately 600 basis points from the prior year as a percentage of revenue, reflecting strong operating leverage as we maintain our focus on cost management. Q1 adjusted EBITDA was $10 million or 26% of margin, up from $6 million or 18% in the prior year, reflecting strong operating leverage as revenue scales. Sequentially, margin declined modestly from 28% in Q4, primarily reflecting the onetime benefits we referenced in our last earnings call. Adjusted free cash flow was negative $1.8 million in Q1, reflecting earlier payments in the quarter. We are also pulling forward a portion of 2027 CapEx into 2026 in response to strong demand signals. Even with that pull forward, we continue to expect adjusted free cash flow to be positive for the full year with improvement weighted towards the second half of the year. We have the capital in place to support the growth that we are seeing. We currently have more than $100 million in capital leasing capacity with approximately half of that utilized. Based on our current operating plan, we expect to fund growth through operating cash flow and capital leases, and we do not anticipate the need to raise additional capital through follow-on equity offerings. In fact, we plan to continue to focus on reducing our dilution through our modest stock buyback and our net share settlements for RSU brands. Looking ahead, we introduced updated B2 pricing and packaging effective May 1. The change reflects the investments that we have made in our platform performance, our effort to further simplify pricing by removing API transaction fees and the rising cost of hardware and data centers. On a net basis, we expect the pricing update to be accretive to revenue and margins, and that will be reflected in our guidance. So moving on to our guidance. For the second quarter, we expect revenue to be in the range of $39.8 million to $40.2 million. On our last earnings call, we said B2 growth in the second quarter would be 12%. Based on this new midpoint, the B2 growth in Q2 will be closer to 20%, which is a big improvement. The Q2 outlook includes a partial quarter benefit from the May 1 pricing update, along with variable usage from customers that we have already actualized in April. We are not assuming the same level of variable usage in the second half of the year. Adjusted EBITDA margin is expected to be in the range of 21% to 23% for Q2. The sequential step down from Q1 reflects the timing of investments as we continue to build for growth. Turning to the full year. We are raising our full year revenue guidance to $161.5 million to $163.5 million, up $5 million from our prior midpoint of $157.5 million. That increase reflects 2 factors: stronger first quarter performance impacting the rest of 2026; and the benefit of the new B2 pricing and offering, each contributes to approximately half of the raise. We are also raising our full year adjusted EBITDA margin guidance by 400 basis points to a range of 23% to 25%, up from 19% to 21% previously. As a reminder, our guidance philosophy excludes individual deals greater than $500,000, high variable usage above contracted minimum and incremental upside from our go-to-market transformation. As these elements become more predictable and repeatable, we will incorporate them into our forward guide and communicate that transition clearly. In summary, Q1 was a strong quarter across the board. Revenue beat, adjusted EBITDA beat, B2 growth accelerating and bookings improving. We remain focused on executing on our AI opportunity by driving forward our go-to-market transformation and scaling our B2 business. We look forward to your questions. With that, operator, please open up the line.

[Operator Instructions] Our first question comes from the line of Mike Cikos with Needham.

Congratulations on the strong start to calendar '26. First question, I guess, is more for Gleb. But I just wanted to get more on the success that you guys are seeing with the AI customers following the go-to-market transformation initiatives we put in place. Can you just talk to the improved visibility you have for those AI customers in the pipe? And as you have more of these customers, I guess, begin to season, are you noticing -- is there a significant departure as far as cohort behavior or sales cycles? And then I just have a follow-up.

Mike, thanks for the question. I'll use actually the 2 customers that I referenced in my prepared remarks as a good example. So one of the customers came to us through the GTM motion that we're building, right? So the machine that we're building, the combination of outbound targeting, better systems, going to the AI events. And so we found them through that outbound process. The other one actually came to us as a referral from one of our existing AI customers who said that they were having a great experience. And specifically, they were saying that the combination of the performance that they were getting from our platform at the price point that they were getting was unmatched. And so they were -- referred them over. And so we're seeing more AI companies coming to us. We kind of feel like the market is coming our way, and it's really from both of these. Part of it is from the work that we're doing, part of it is from the referrals in the market coming to us. So that -- maybe that answers the first part of that. In terms of the cohort part of it, maybe you can answer question if I didn't completely get it. But one of the things we mentioned on the prior call is that we're seeing the AI companies growing much faster, about 3 times faster than our average customer, and that's just a function of their inherent data growth driven by their AI use cases. Did that answer the question you were asking?

It does. It does. And then for the follow-up, I think it might be more geared towards Marc, but I just wanted to double check on the B2 with the NRR of 110%, I know you said, hey, we drive that between 2 factors, right? You have the data consumption which grows each year, and then you also have the cross-sell, upsell. And I just wanted to see, can we unpack that a little bit more to get evidence of the go-to-market actually driving adoption, whether it is the cross-sell, upsell motion? What are the drivers behind that B2 NRR today if I'm trying to unpack consumption growth versus go-to-market initiatives to expand wallet and drive additional offerings into the installed base?

Yes. Mike, I'll start off by saying I think the best evidence of the GTM working is the RPO disclosure of committed contracts that change quarter-over-quarter. So for commitments of less than a year, it's up $3.4 million. You could see it on Slide 19 of our earnings deck. So I think that's the best evidence of the performance. Now that is made up of both, new logo as well as expansion sale. The expansion sale realistically does fluctuate. On the NRR, we did move to in-quarter reporting versus the trailing 4-quarter average, specifically to give you more visibility and to hold us accountable to explain what's happening. So there will be more fluctuation there from that perspective. So it's up to 110% from 105% a year ago because a year ago was a quarter where we did talk about one customer -- one large customer going away. I mean, since I joined, that was the only time we've had to reference that. So that's what drove that improvement year-over-year. But it generally fluctuates around 110% on a stable basis, but there's going to be some ups and downs and the expansion changes will be the biggest driver of that because the organic growth tends to be incredibly stable and predictable.

Congrats again on a strong start to the year.

And our next question comes from the line of Ittai Kidron with Oppenheimer.

Congrats, great solid numbers. I had a couple of things, maybe starting with you, Marc. Can you be a little bit more -- can you give us a little bit more color on the pricing update, the magnitude of this? How much of this you think you can capture? I'm just trying to think about your growth without the pricing update, how would you -- your outlook would have looked without it? I'm just trying to get my hands around that.

Yes, absolutely, Ittai. So in the $5 million raise for the year, half of it is from the pricing and packaging change. Half of it is from the strength of the business that we observed in Q1, so the booking -- the strong bookings in Q1. So if you look at that RPO number I referenced just a few moments ago, that kind of roughly equates to the raise from the organic health of the business for the rest of the year because that has no price increase in it. We continue to guide very prudently for the rest of the year. So the same philosophy we laid out last time, which is no large customers, no go-to-market benefits and not accounting for large -- variability of the -- of that large customer. What I would say also within Q2, I can give you a bit more color there, last time, we said Q2 would grow by -- for B2, we grow by 12% year-over-year. Now it's 20%. That difference is more anchored on the organic health of the business because the price change took effect May 1. So it's not a full quarter. And we obviously actualize some of the things we saw in April in the business. And on the price, I mean, we could elaborate a bit more on that price change. It's not a flat price change. It's a pricing and packaging change. So for instance, we are including now transaction API fees. So in the spirit of being the simplest billing model out there, we further simplified by no longer billing customers for transaction fees.

Got it. Okay. And then as a follow-up, maybe one for each of you. Marc, for you on the Computer Backup, the net retention rate is now well below 100%. So is this a model -- is this a business we should model towards decline or going forward? And for you, Gleb, on the go-to-market side, great to see the progress there. What else is left here? What is it that between now and year-end still needs to kick in, that hasn't from your perspective?

Yes. So Ittai, just to reiterate, the -- we're still thinking of Computer Backup as declining year-over-year 5%. And the NRR is going to be tightly tied to that because it's a subscription business, not consumptive, which would mean that B2 would grow 24% year-over-year. So the change in outlook is pretty much all on B2 and Computer Backup remains at a decline of 5% is what we're forecasting and guiding.

[indiscernible], thanks for the question about the GTM transformation, what's done, where -- and what we have to do still. So what I'll say is I think we've made great progress this quarter, and there's still a variety of things that we want to get further, right? So we hired Anuj Kumar to run that organization. I've asked Jason to -- who's with us to take on and focus most of the time on the neocloud opportunity. So Jason works for Anuj, and we see that as a $14 billion opportunity. So we're putting focus and resources on that specific part of the opportunity with Jason focusing on that. The awareness generation is off to a good start with Flamethrower, but it's only been a couple of months in. And -- so we've been moving faster than, I think, expected on that, and we've been invited to participate in some great organizations and partnerships with a16z and Startup Grind and Launch. But it's -- there's a lot of opportunity there still between that and the open source developer efforts that we're doing. There's still a lot of opportunity to make sure that everyone thinks of Backblaze as their first spot for their price performance storage. So there's a lot that we've done. There's still, I think, a lot of opportunity that we have. There's -- I am excited that we're seeing pipeline growth stronger than we've seen in the past. We're seeing more of our sales team hitting their quota than we've ever seen in the past. So a lot of the right things are happening, but we still -- we're always going to keep working on it.

And our next question comes from the line of Erik Suppiger with B. Riley Securities.

Congrats on a good quarter. Can you speak to what portion of the neocloud market you're either servicing or at least engaged with? And then where are you in terms of the hiring on the sales front? Are you adding additional salespeople at this point? Or where are you on that -- from that perspective?

Yes. Thanks, Erik. So there are about 200 neoclouds. We are -- we went to GTC, the NVIDIA's premier conference and had just a host of great conversations there at GTC -- there's some noise on the line. So the -- what I'd say is we're engaged with most of the top neoclouds as part of it. The part that we are servicing for them is this data lake layer, right? So if you think of the AI workflow, the GPUs themselves, there's the very low latency, high-performance Flash that you want adjacent to the GPUs. And then what you need is the place where you store all of the data, right? So you can almost think of it -- if the whole AI workflow was a laptop, you've got your compute, your CPU, you've got the RAM and you've got the hard disk or SSD. We are basically providing that hard disk layer. There's about half a dozen companies that provide that RAM layer. And then the base neocloud part is that CPU, GPU part. So we're providing that large scale, high performance, not the highest performance, but high performance per dollar data lake layer for them. And so we're the -- we're a white labeled provider for them. We're doing that, as we talked about in the last call, we've got the 6-, 7- and 8-figure deals that we signed for that. We have others that are in the works, and we're engaged with a bunch of the neoclouds at this point.

Do you -- Can you give us a sense of what portion of the neoclouds out there -- of the 200 that are out there that you're speaking to? Do you think it's a quarter? Any gauge on what penetration you've had?

So I think in terms of the conversations and engagement side, probably somewhere around that number. But I would say we're engaged with pretty much all of the top ones at this point and having different levels of conversations and some in POCs, et cetera, with them. On the sales side of it, we talked earlier in this year that there were a number of different roles we wanted to fill. At this point, I'm excited to say we filled the CRO role with Anuj, we filled the Rev Ops role. We filled the sales development role. And so we're -- we've got a really strong kind of build-out of that team now.

And the next question comes from the line of Jeff Van Rhee with Craig-Hallum.

A couple. First, just maybe, Marc, help me with the guide and the outlook. I'm trying to understand the progression here. So in -- at the end of February, Feb '24, you took roughly $4 million out relative to the consensus, and now we're putting $5 million back in. I'm trying to understand, in the Feb '24 call, was the May 1 price increase in B2 already contemplated in the guide?

Jeff, no, that was not contemplated in the guide. And in the $5 million increase we just did, half would be from the pricing, half would be from the organic momentum and health of the business we saw in Q1. And the change is really -- a lot of this is guidance philosophy we spoke about, just a lot more prudent going forward. That's what drove the change.

Did you -- if you take the final month of the quarter, March and then April, was there -- I don't know if radical is the right word to use, but did you see substantial improvement in close rate? Because it sounds like you're saying your conviction is coming both from improved bookings as well as usage. So I'm trying to understand how Jan, Feb bookings were weakish and then all of a sudden, March, April really killed it? I know you've made some process change over time to sales, but it was just such a quick snap. Is that -- maybe you can just help me dial it in there a little bit?

Yes, Jeffrey Van, this is [indiscernible]. Maybe I'll touch on and then Marc can also weigh in. So we certainly had a more back-ended quarter in Q1, and we've started off Q2 strong. So there's definitely [indiscernible] feel from the numbers that we're seeing, right? And then I think -- and we talked about like the $1 million-ish deal that closed in 11 days, that started and closed towards the end of the quarter. But it wasn't the only deal, right? The pipeline itself has been building strongly this April. And I think we're layering that on along with the execution that we're doing on our own side. So I think that that's kind of the, I guess, the conviction and emotion side of things based on the data and the execution. And then I'll let Mark, if you want to add anything on -- beyond that on the guide side of things.

Yes. I mean Q4 bookings -- back in Q4, bookings were good, but we wanted to hit that 30% growth, Jeff. To hit 30% growth, we'd have to be booking like $5 million a quarter. okay? So we weren't at that rate-rate yet, but it's been improving pretty much every quarter. And this latest Q1 is a further improvement and probably the closest we've gotten, frankly. And the demand signals are really strong. So the demand signals being really strong. We're feeling good about the outlook, but we're still guiding with that prudence. And we'll use some of that price change to also fund some additional CapEx, so we could have further capacity in place to handle that demand because we don't want to be in a position where we're declining any revenue opportunities.

Yes. Yes. Got it. Got it. And just to follow up on that last piece then in terms of the outlook for the year for CapEx for '26, I heard you referenced it, but can you just give us a number there, what are you expecting? And then also on the stock comp?

Yes. On the CapEx side, we're probably going to be around mid-30s as a percent of revenue. I would say there's 3 factors there. One, last quarter, we spoke about that large customer we got to service next year. So we need to get that CapEx in place now. Two, all the strong demand signal. And three, the general equipment cost is 30% higher than it was on a per unit basis from a year ago. So for those 3 factors, we're beefing up our CapEx plan for this year, accelerating it from '27 into this year.

Yes. And your thoughts on stock comp?

I'd say pretty stable. If you look at our headcount, I mean, generally speaking, year-over-year, our headcount is actually coming down. So we're continuing to drive more efficiency out of the business. So stock comp should be pretty stable in dollar terms. And as a percent of revenue, it does improve over time.

Jeff, the other thing -- one thing I would also just mention since you bring up supply chain and supply chain constraints and all that. What's interesting is we have to buy the equipment, right? So we have to spend more on some of that side of things. But the interesting thing is also we get 2 tailwinds from the supply chain being constrained. On the GPU side, when -- because the supply chain is constrained on the GPU side, customers are saying, "Well, I need to go and have access to wherever the GPUs are available." And so we regularly talk with customers who say, "I have to have my data somewhere, that I can send it to whichever neocloud has the GPUs available." And on the memory side, which is also obviously heavily constrained, the neoclouds that offer cloud storage have been building out often on Flash, and that becomes really expensive, especially now with the constraints there. And so it's driving additional interest from the neoclouds in working with us on that data lake here. So on the one hand, we have to deal with prebuying ourselves on the equipment side for CapEx. But on the other side, we have these 2 tailwinds to the business.

That's helpful. Congrats...

Jeff, just to step back on stock comp. I mean, if you look at the statement of cash flows, Q1, obviously, stock comp is higher as we settle some of our annual bonuses in equity as well. But you'll notice this year's stock comp was actually lower than last year's.

Congrats on the turn, guys.

And our next question comes from the line of Jason Ader with William Blair.

Just wanted to get a better sense on the neocloud. What are the size of some of these deals? I know you talked about the 8-figure deal that's coming in, I believe, next year. But maybe just some more detail on some of the other deals that you've landed or are in the pipeline? Are we talking about kind of household neocloud names that are contracting with you for potentially further kind of 8-figure deals? I mean, just, I think gauging kind of how significant an impact you might have from some of these neocloud opportunities would be helpful.

Yes. Thanks, Jason. So first of all, the -- we estimate that our opportunity in the neocloud market by 2030 is $14 billion. And that is just the data lake tier that we provide, right? So that's not the entire storage footprint. The deals that we have signed, the 6-, 7- and 8-figure deals that we signed, I'll say 2 things. One is you would recognize them, right? They are companies that you would know. And two is that all 3 of those are initial deals. So all 3 of them are ones where companies -- the companies look at it as the way to start, not the total opportunity. So I think there's -- frankly, I can -- I could see a path where the 6- and 7-figure deals could become 8-figure deals themselves. The 8-figure deal can certainly scale from where it is once it's ramped. So that's kind of a little bit of that side of the opportunity. The other conversations that we're in, many of them are -- assuming they move forward, are of that same scale. Some of the conversations are -- we may want to start with a 6-figure or 7-figure deal, but many of them, the scale of the opportunity is 8 figures at ramp.

Okay. Helpful. And then just on the -- I guess, the risk potentially that the neoclouds add a lower cost storage tier and then you guys are helping them for a little bit, but then they kind of in-source it?

I mean it's always possible, but it's a little bit like -- for the first almost 2 decades of Backblaze, one of the questions we were always asked was what happens if AWS ends up lowering their price to match you? And we're 2 decades in and that hasn't happened. I think that the challenge is it's not easy to build the type of IP that we have built up over the last 2 decades. It requires scale and expertise and a focus over a long period of time to get it really honed and right. And the thing for the neoclouds is they have a lot of things they need to do, right? There's opportunities around GPUs and GPU scaling and optimization and how do you make tooling better for inferencing and all kinds of things. Spending all their resources to try to replicate what we have built over the last 2 decades is probably not the best place for them to invest their own resources when they can -- when time to value is so much faster by using Backblaze.

Okay. Great. And then, Marc, for you, just a couple of quick ones. So the -- with the higher CapEx, are you still guiding for free cash flow positivity this year?

Yes. So the second half of the year, we're still guiding for that to be free cash flow positive. For the whole year, I mean, Q1 was minus $1.8 million. Q2 should be somewhere around neutral and the second half of the year should be positive. So net-net for the year, we should be neutral or very 1% of revenue, free cash flow positive despite the acceleration of the CapEx.

Okay. Great. And then just on the gross margin, just last question for me, sorry. It's sort of -- as I look at the last few years, I'm looking at just the -- not the adjusted gross margin, but the reported non-GAAP gross margin. It was like mid-50s for a few years. And then last year was 62% roughly, 62% in Q1. Can you just remind us of like what caused the kind of the significant increase in the gross margin and then maybe some puts and takes going forward on that gross margin line?

Yes. Sure, Jason. So if you recall about 1 year ago, we reviewed the estimated useful life of our fixed assets. And it turns out we're using all our fixed assets for typically 6 years onwards. So we moved all the depreciation to 6 years. So that drove a big benefit to gross margin. Second, all other lines, like if you think about all the labor or payment fees or everything else that fits to our cost of sales, we've managed really tightly year-over-year. So it's kind of staying flat in absolute dollars pretty much and improved as a percent of revenue. We're looking at everything within our gross margin now to further drive optimization. But I would say between the price increase which benefits gross margin, but the accelerated CapEx, which will push our gross margin down, it should stay flat around where it is now. We're not seeing -- we're not guiding to any major changes in gross margin through the rest of the year.

And our next question comes from the line of Eric Martinuzzi with Lake Street Capital Markets.

Yes. I was just curious about the timing of the price increase. I went back and looked it up, I guess it was, October of 2023 was the last time you raised the price on B2, and you really hadn't raised it since you rolled out the product back in 2015. So we're at about the 2.5-year mark here with the price increase. Was this something that you felt like, hey, we're delivering more value, we need to capture more value? Or was there competitive issues where competitors were raising price and kind of provided an umbrella for you to do the same?

Yes. Thanks, Eric, for the question. So we periodically reevaluate what the pricing and packaging of the offering should be. When we were looking at it, there were a few things that came together. One is that we've been investing more into the performance of the platform. More of our customers are using us in these hot use cases where we're driving high throughput, high IOPs. We've been -- we made egress free before. And one of the things that, that's enabled is not just that it's less expensive for the customers, but it allows them to actually run more frequent training of their models in the AI use cases. So it's actually unlocking their ability to innovate, but that makes it -- it's free for them. It costs us money to provide that. So we've been working to increasingly provide more and more value to these higher performance, more active use cases. And we also wanted to simplify the pricing by removing transaction fees. So the pricing and packaging combination along with, as Marc said, the underlying costs of the components, have been increasing. So taking all of that together, we decided this was the right time to do that.

Okay. And with regard to competitors, do -- I mean, historically, you guys have thrown out that, hey, we're 80% cheaper than Amazon. Does this -- obviously, you're raising, I think what I said was about a 15%, 16% per terabyte per month. Does that shrink that gap now? Or do you still feel like there's a big delta?

There's -- we are still dramatically more cost efficient than the alternatives out there. I was literally actually just talking to one of our account execs a couple of days ago, who was talking about a customer who has been ramping on our platform. And they said that they moved over a lot of their data and they're continuing to move over more of their data, more of their use cases because on the one hand, we're more affordable on the storage side, right? So just at the base level of storage. But where they were getting hit dramatically at their prior provider was that each time they egress the data out from their provider to one of the other neocloud providers, they were getting hit with massive egress fees, one. And two, the transaction fees were actually costing them 3 to 4 times more than the cost of the storage at their prior provider. So when you put it all together, they were more than -- 5 times more expensive at their prior provider, and they were literally wondering whether that was going to even be affordable for them to stay in business. So the scale of total cost of ownership that we provide on a benefit basis is still quite dramatic.

And our final question comes from the line of Rustam Kanga with Citizens.

Nice, clean set of results here. Just one on B2 neo. As workloads begin to shift more towards inferencing from training, will that lead to improving predictability and visibility? And then to that end, could you potentially share what percentage of neo business on average or even directionally represents inference versus training workloads?

Sure. It's a good question. And the first part of the answer is yes. As things move towards inferencing, it does make it easier to be more predictable. Today, more of the use cases that we're seeing are related to model building. And that makes sense because a lot of the data sets right now that are very large and that need to be moved, are related to the model building, and we're a great service for that. But I'll give you an example. In the Gen AI media space, I was talking about -- to a customer about their data flow. And the data flow is they accumulate a lot of data. They store that data, they annotate that data, then they find a GPU provider that is available and then they run iterative model building on the different GPU price. So they use us to store that large data set. And as they acquire the data sets, they use us to store more of those large data sets. They love the fact that they can store those efficiently and then send them quickly and for free to the GPU provider that they want. So they're using us in this whole model building process. Now as they do that, the other side of their business, the actual thing that they offer and they charge for is generating videos. That's all the inferencing side. And so they're looking at us for the outputs of all that video because every single time a user generates a new video, that video then gets stored basically forever and each version and each iteration gets stored forever. And so we become a great place to store that. And that inferencing side is a much more smooth and predictable side. So the short answer is, yes, it will be more predictable as we get more inferencing. Today, the bigger workloads that we see are related to model building because we're great for that. But we are seeing more inferencing start-up on our platform.

And that concludes our question-and-answer session. I will now turn the conference back over to Gleb Budman for closing remarks.

Thank you, everybody. Q1 was a proof point. We beat on revenue, beat on EBITDA, B2 is growing 24%. The deal size has more than doubled. The AI customer is up 76% year-over-year. We are not just riding the AI wave, we're building the infrastructure that supports it. We are key for the neo clouds, key for the AI builders, and we've had nearly 2 decades of optimizing performance per dollar at scale, which makes us ideal for the needs of AI. We raised guidance. We're on track for our first full year of free cash flow positivity as a public company, and we're picking up steam. Thank you to our customers, our partners, and thank you to our amazing team that's making all this happen. Thanks for joining our Q1 call, and we look forward to connecting on the next one. Bye-bye.

And ladies and gentlemen, this concludes today's call, and we thank you for your participation. You may now disconnect.