Backblaze, Inc. Q3 FY2025 Earnings Call

Hello, and thank you for joining us. My name is Tiffany, and I will be your conference operator today. I would like to welcome everyone to the Backblaze Third Quarter 2025 Earnings Call. I will now hand the call over to Mimi Kong from Investor Relations. Mimi, please proceed.

Thank you. Good morning, and welcome to Backblaze's Third Quarter 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-saving initiatives, results from new features, our ability to compete effectively and manage our growth, our strategy to acquire new customers, retain and expand our business with existing customers, and the impact of previous price changes. 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 rates, and adjusted free cash flows. And finally, we will be in New York to participate in the Craig-Hallum Alpha Select Conference on November 18 and the Needham Tech Week one-on-one event on November 20. Thank you for joining us. And I would now like to turn the call over to Gleb.

Thank you, Mimi, and welcome, everyone, to the call. We delivered strong results this quarter. In Q3, revenue and adjusted EBITDA margin both came in above the high end of guidance, and we continue to be on track to be adjusted free cash flow positive in Q4. Overall company revenue grew 14% year-over-year, and B2 Cloud Storage delivered strong results, growing 28%. Now I'd like to take a step back and share how we see the AI industry evolving because AI is becoming central to both our customers and our opportunity ahead. AI is built on models, compute, and data. While just a couple of years ago, models were the domain of only a few large providers, today, they are widely available with literally millions of open-source options. So models are no longer a bottleneck for AI innovation. The next component of AI is compute. The GPUs that make up compute have become increasingly available, but access at scale remains challenging, and the market has also become quite fragmented with approximately 200 Neo clouds, such as CoreWeave and Nebius offering GPU capacity for it. That brings us to data, the key differentiator for AI success and where Backblaze helps companies win. Data sizes are exploding as AI expands beyond text to images, audio, and video, driving massive storage and performance needs. The teams leading in AI aren't just the ones with the most data. They're also the ones who can aggregate, clean, and organize it for today's workloads, move it to whichever Neo cloud they choose, all while preparing it for tomorrow's architectures. Backblaze has already built one of the largest, fastest, and most cost-effective storage clouds on the planet, offering throughput of up to 1 terabit per second and priced at about 1/5 that of traditional cloud offerings. Add to that, our free egress and universal data migration that allow customers to easily move their data from other cloud providers where they may feel locked in and then freely send their data to one of the 200 Neo clouds or anywhere else they need to go. Then layer on our team actively supporting customer success, and it becomes clear why hundreds of AI companies are choosing Backblaze. Highlighting this, Backblaze recently received industry recognition in both cloud security and technology innovation, including an award for B2 Overdrive. These honors reflect the strength of our technology and the pace of our innovation. Now let me share a few customer examples that bring this to life. This past quarter, we signed a new 6-figure deal with an AI start-up focused on large-scale vision language models. The team was small but scaling rapidly and racing to finish key projects. They couldn't predict how their data volumes might grow and needed a high-performance and cost-effective solution that could scale with them. They started self-serve and later engaged our sales team to access even higher performance. B2 Overdrive gave them the performance they needed, and our team made it easy, freeing them up to focus on their critical projects. Another AI customer in the surveillance space expanded their commitment by almost 10x to a 7-figure TCV deal. We won this customer away from a hyperscaler whose platform limited how they could automate their workflow. Our platform offers a valuable multi-cloud upload that streamlined their workflows. They expanded their commitment with us because we offered a predictable and cost-effective way to scale easily. Finally, we had another 6-figure win powered by Backblaze's white label deal with an app developer in the media space. This was another competitive win from a hyperscaler. The customer was growing fast and needed storage that could support millions of users. After being hit with a surprise roughly $100,000 charge for moving data with the hyperscaler, they came to Backblaze for predictable pricing and scale. Backblaze is now their storage platform, and they will join our 100,000 customers who collectively serve hundreds of millions of end users through our platform. Across all these examples, the reasons customers choose Backblaze are consistent: performance that rivals the biggest clouds, predictable and fair pricing, and people who make it easy to get things done. Whether it's a start-up training models, an immediate team moving petabytes, or an enterprise scaling to millions of users, the story is the same: faster, simpler, more affordable, and supported by a team that actually helps. Now I'll share how we see our growth opportunity. In Q3, we grew B2 28%. For Q4, we now expect B2 to grow in the range of 25% to 28% year-over-year. We're proud of our growth. But this is below the 30% target we set for ourselves at the beginning of the year. To help reach our growth goals, we're launching Phase 2 of our go-to-market transformation, focused on accelerating the velocity of both our self-serve and direct sales motions. For self-serve, we're proud of the consistent growth engine we've built. Now we're going to scale it by making it more frictionless for data-heavy AI use cases and by kicking off robust developer relations. These steps are aimed to deepen our connection with the developer community and make it even easier to adopt our platform. For direct sales, we're building on the progress we've made over the past year. We're adding talent and upgrading our core systems in order to better target customers and improve sales efficiency. To accelerate this, we're also partnering with advisers and a consulting team who has helped both Snowflake and Databricks in their go-to-market execution. We're reinforcing our sales and marketing engine and remain confident in our ability to deliver consistent, durable growth over time.

Thank you, Gleb, and good morning, everyone. We delivered a strong third quarter with results coming in above the high end of our guidance for both revenue and adjusted EBITDA. We have solidified our balance sheet, improved our path to being free cash flow positive, and accelerated revenue growth. We're focused on getting B2 to a Rule of 40, and we are on track to roughly triple our score this year. Now, let me walk you through the details of the quarter. Starting with revenue. Total revenue exceeded expectations. It came in at $37.2 million compared to the high end of guidance, which was $37.1 million. This represents a 14% year-over-year overall growth. B2 grew 28% year-over-year compared to the organic growth of 19% in the same period last year. This represents an acceleration of about 900 basis points this quarter. This improvement was driven by the first phase of our go-to-market transformation. As mentioned last quarter, usage from one of our larger AI customers has been variable as the customers' data storage needs have fluctuated for their business. Overall, industry-wide demand for data storage is expected to grow rapidly, and our platform is well-positioned to support those expanding needs. We're also seeing diversification within B2 across our core use cases, which are live application hub storage, backup, media, and AI-related workloads. In Computer Backup, revenue was flat year-over-year, reflecting the final roll-off of the price increase implemented in 2023. Turning to net revenue retention. Overall, the trailing 4-quarter company NRR for Q3 was 106% compared to 109% in the second quarter. Going forward, we believe it's clear to show in-quarter NRR versus the previously reported trailing 4-quarter average. As an example for Q3, in-quarter NRR for B2 improved to 116% from 109% in Q2, and this was driven by a large AI customer we discussed last quarter. You can see that dynamic on Slide 11 of the earnings presentation. Moving on to gross margin. It was 62%, up from 55% a year ago, reflecting operating leverage and the benefit from our recent useful life study. Adjusted gross margin was 79% compared to 78% last year. Overall, margin performance remained stable. Data center costs are generally rising. However, they are being offset by scale in labor, greater code efficiency, and lower amortized R&D as a percentage of revenue. Operating expenses were 71% of revenue, an improvement from 92% a year ago. This reflects the structural changes that we made last year through our restructuring and zero-based budgeting process. R&D spending held steady in dollars, but declined as a percentage of revenue from 33% a year ago to 30% this quarter. When you combine the R&D expense and capitalized R&D, the improvement is even more pronounced at 35% of revenue, down from 43% a year ago. Sales and marketing came in at 24% of revenue, down from 36% a year ago. We're focused on improving efficiency in our go-to-market model, including reallocating funds to strengthen our top-of-funnel programs and investing in operational go-to-market talent, all while maintaining the same cost discipline you've seen from us over the past year. G&A was 17% of revenue, down from 23% last year. We continue to drive efficiencies across our corporate functions and general administrative spend. Operating results. GAAP net loss was $3.8 million, a 70% improvement from a loss of $12.8 million in the prior year. On a non-GAAP basis, net income was $1.9 million compared to a loss of $4.1 million last year. Adjusted EBITDA margin reached 23%, almost double the 12% from a year ago and a quarter ahead of our outlook. That performance reflects continued financial discipline and the operating leverage we've built into the model. In Q3, adjusted free cash flow was negative $3.5 million, improving by roughly $0.5 million year-over-year. As we discussed last quarter, we used $2.5 million of our line of credit to fund capital expenditures outside of the U.S., so we can cut our international borrowing rate in half. Balance sheet. We ended the quarter with $50 million in cash and marketable securities, largely unchanged from last quarter. We believe the balance sheet remains strong and provides flexibility to support continued growth and investment. As we announced last quarter, we initiated a modest share repurchase program. In Q3, we repurchased $1.2 million of shares as part of our ongoing work to manage equity dilution. Guidance. For the fourth quarter, we expect revenue in the range of $37.3 million to $37.9 million. We're slightly widening this range to account for the variability we see in certain large customers. B2 growth in Q4 is expected to be between 25% and 28%. We continue to focus on driving operating leverage and remain on track to be adjusted free cash flow positive in Q4. To close, our financial transformation is progressing well. We're reducing equity dilution through our repurchase program, on track to achieve positive adjusted free cash flow in Q4, and continuing to build operating leverage towards GAAP profitability. While we haven't yet reached our 30% B2 growth goal, we expect the next phase of our go-to-market transformation to drive stronger growth. Together, these efforts are building a stronger, more efficient company, one that's on path towards operating at a Rule of 40 profile. Just looking at B2, we started the year with a Rule of 40 score of 9 and are on track to roughly triple that in Q4 of this year.

Your first question comes from Jeff Van Rhee with Craig-Hallum Capital Group.

Just a couple here. In terms of the sales evolution in Phase 2, talk to me about how you were envisioning the sales sort of evolution. I think you commented last quarter that most reps are more than halfway through the ramp. Are the existing reps ramping the way you had expected? Or was there something you experienced in the sales process that led to you sort of implementing the second phase?

Thank you for the question, Jeff. In the first phase, our goal was to move upmarket, and we have certainly achieved that with multiple 6- and 7-figure deals announced throughout the year. Phase 2 will focus on enhancing our execution velocity, which shifts our attention slightly. We aim to continue moving upmarket and supporting larger deals while also concentrating more on advancing opportunities through the sales funnel in a more explicit and active manner. Most of our representatives were hired at the beginning of the year or earlier, so they are mostly past their ramp-up periods as part of the onboarding process.

Sure. Can you provide more details on the B2 situation? You initially set a goal of 30% for the fourth quarter, and I heard some comments about that, but could you elaborate further? What specific factors did you anticipate that didn't materialize? And when can we expect to reach that 30% threshold?

So from our last earnings call, two key developments occurred. First, we had a significant customer mentioned in the previous call whose performance in Q2 was partly influenced by variability from a large AI customer. As we look towards Q4, this same customer is adjusting more than we anticipated. The second factor affecting us is that several larger deals as we target higher markets are taking longer to finalize. This combination has slightly altered our expectations for Q3. That's why we're initiating Phase 2 of the go-to-market transformation. We're excited about our progress in moving upmarket, as it demonstrates that the platform can accommodate larger customers and use cases. However, we also aim to establish more core predictability in our business. Previously, we had many small deals which provided strong predictability, enabling us to go public with numerous transactions in the $300 to $2,000 range. The downside was it made it harder to achieve rapid growth. The larger deals and our move upmarket allow us to accelerate growth, but we also want to boost our core base of smaller deals to ensure consistency.

One last quick one, if I could, on the data variability, Gleb, you mentioned several times on the call. Just talk about how that variability is maybe different than what you had expected from these customers, obviously, as you're getting AI workloads, some new experiences here as to how people are going to use you both in that moment as well as over time. Just a bit more about the variability on the data usage would be helpful.

Yes. The AI use cases are obviously all evolving very rapidly, and we're trying to support the customers where they are. So the different use cases that we see across AI are things that we are supporting and learning along with the customers. This particular customer that we talked about, as their business goes up and down in terms of the needs for their data, that we support them with that. So the way that we work with customers is many of our customers are pay-as-you-go. They enter a credit card and they just sign up and go. Some of our customers, we sign contractual commitments with. And for a number of the contractual commitments, they'll often sign a contractual commitment and then have the ability to scale up above that. And with some of these larger AI customers, they'll sign a contractual commit and then they'll ramp up even faster than that because their business is growing faster than they expected and then sometimes they need to do some pullback. So that's what we're seeing with some of these. The nice thing, obviously, is we have 100,000 customers on B2, we're distributed across a variety of use cases. So it's not all variable AI customers. We still have a lot of core predictability in our business. But we do want to lean in with the AI use cases because we do see that as transformative for the industry and where we see a lot of opportunity ahead.

Your next question comes from the line of Eric Martinuzzi with Lake Street Capital Markets, LLC.

Yes. I'd like to touch on - I know you haven't given a guide for 2026, but I just wanted to learn your expectations for the two sides of the business going forward. Obviously, we've got good growth, but not the 30% you want in B2. And then we were flat here in Q3 on the CBU as we got through the last of the price increases. But to put it in broad strokes, we're looking at a year in 2025, where combined, we're at about 14% growth. I know I'm looking for that 14% growth to persist into 2026. But what are you thinking about for longer-term growth rates for B2 and CBU?

Yes. Eric, this is Mark. So 2025, right, for B2 is on track to be in that mid-20s year-over-year growth rate. 2026, we don't want to give guidance yet, but we feel pretty comfortable it's going to be in that range given what we're seeing now as well as that second phase of that go-to-market transformation. And those larger deals Gleb spoke about, if those come in, I mean, it could bring that number higher when they come in, in those specific quarters. On the computer backup, just like we said last time, it would be low to single, mid-single digits, the contraction of that business. We're obviously taking action to try to stabilize that. We probably need more time to come back and report a different outlook there. But for now, those are the right assumptions for B2 and computer backup.

Okay. And then the restructuring that you announced this morning as well, just curious to know where are we - you talked about some facilities, but it also talked about severance. Where are we cutting back on the headcount that we had versus the end of September?

Yes. Listen, that is how we're funding that next phase of our go-to-market transformation. Gleb talked about needing that mid-market, high-velocity deal volume to improve. It's done really well over the past year, but we think it should be way better than where it is. And that's the key formula to get to that 30%. So a lot of that restructuring cost is to transform our go-to-market practices. It's to bring on new talent on board. We have started that. So it's a lot more about getting talent that's a lot more operational in their go-to-market skills, know how to use the technology, the data science behind it. So I would say that it's more of a reinvestment versus a cost-cutting exercise.

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

This is Nolan Genvine on for Ittai. First, I just want to sort of double-click on these 7-figure customers. You're clearly showing traction in selling to enterprises now, and you've added yet another 7-figure customer this quarter. How many 7-figure ARR customers do you have now? And then how has their expansion dynamics evolved in recent quarters versus sort of the smaller cohort? And then I have a follow-up.

No, thanks for joining us, and thanks for the question. One clarification I would say is we have these 7-figure deals. The interesting thing is they're not all enterprises. They're heavy users of data, right? And so one of the things that we've seen is that customers are generating large volumes of data and needing high performance for those data sets. The AI-powered video surveillance customer that I mentioned, they're not an enterprise company, but they're a 7-figure deal. So, and that actually, I think, bodes well for us for the future because these smaller mid-market companies that have high data needs are generally faster moving than the traditional enterprises, and there's more of them out there. So that's one thing I would just say in terms of outlook and opportunity for us. In terms of the number of them, we've generally announced 1 or 2 per quarter over the last year. So it's not a large number of these 7-figure deals, but it's considering that in our first roughly 15 years of the company, we signed one. In the last year, we've highlighted at least one a quarter, I think is a great proof point to say that when we said we were moving upmarket, we moved up market.

Yes. And what I'll add, Nolan, is last quarter, we shared the number of customers with $50,000-plus ARR. It was up in Q2, 30% year-over-year. In Q3, it was up 41%. So we're maintaining good momentum on that front as well.

Got it. That's very helpful. And then I know you're not giving '26 guidance. But how should we think about the potential catalysts heading into the next fiscal year, things like overdrive ramping, this sort of Phase 2 of the go-to-market transformation? Could you just maybe put a finer point on what you think could maybe drive upside or downside, just to the sort of current growth rate?

Yes, it's a good question. So the go-to-market transformation Phase 2 is certainly something that we're leaning into heavily, right? We're putting a lot of focus on them in terms of talent system upgrades, leveraging advisers, and the consulting team. We believe that there's a lot of opportunity in driving the ability and execution for us to get to market faster with the products that we have. In terms of B2 Overdrive, we announced that product line fairly recently. We've already closed multiple 6-figure deals on there, which is, I think, a great sign that it has product-market fit. We believe it's the highest throughput per dollar offering on the market, and that's a great fit for customers that really need that performance. One of the other things that we've seen on Overdrive, which has been interesting, is that we've seen customers come to us because of Overdrive. They heard about Overdrive, they heard about the performance availability, and they came to us for that. And then when they showed up, they realized that our B2 standard offering is actually quite a high-performance offering, and it was sufficient for them. So I think it's changed some of the perception around what we offer. And so even customers that aren't signing up for Overdrive are often coming to us with these higher performance needs, which is a great direction for us as we become a more strategic partner for them. Overdrive provides up to 1 terabit per second of throughput. So it's a blisteringly fast offering for them. I do think that the GTM transformation is the thing that we look at as key to the overdrive offering. The other thing that I would just mention is I don't want to undersell the Phase 1 transformation that we did. The Phase 1 transformation was all about moving upmarket, and that is continuing to be an important catalyst for us, including for 2026. Mark mentioned that the big deals are something that can accelerate us even further, and that's something that we're excited about. The last thing I'll just mention is that the GTM Phase 1 transformation, I mean, it doubled pipelines, it doubled bookings, and doubled channel. So it did help quite a bit. And so we're excited for Phase 2 of that.

Your next question comes from the line of Simon Leopold with Raymond James.

I wanted to ask about this Phase 2 initiative, in that I don't imagine there are free lunches and have to assume that there's some investment involved. I know earlier, you talked about sort of reallocating. But if we think about it, your sales and marketing have been running between roughly 21% and 23% of revenue. What are you budgeting for next year? How should we think about either a dollar basis or a percent of revenue basis, but some metrics to try to get an understanding of what investment you're making for this Phase 2 initiative?

Simon, it's Mark. The percentage of revenue for sales and marketing should stay stable as a percentage of revenue. The funding is going twofold. The restructuring allows us to fund the one-time cost that's not in that percentage of revenue. That allows us to drive the transformation work, which is really to rejigger all our go-to-market systems to make them work better together, cleanse the underlying data, and drive that data science capability I was talking about. The talent refresh we're doing is within existing OpEx budget, right? As one of you highlighted earlier, yes, there are severance costs that are one-time charges that fit into the restructuring. But on a recurring basis, when we finish this as a percentage of revenue, it should be pretty stable.

Great. And then in terms of the outlook for B2, slight downtick, nothing to be embarrassed about, high 20% growth. But what changed in your mind versus your expectations when you set the goal? What's different?

Well, we set the goal all the way back when we launched our first phase of that transformation, which was exactly a year ago. That was a while back. That is our aim. Frankly, it remains our aim. Our Phase 1, as Gleb said, doubled our pipelines, doubled our bookings, and doubled our channel business. What we noticed candidly is our inbound motion is really strong, right? Like we addressed a lot of our technical marketing content, our blogging. We started doing more events and webinars, and all that really improved our inbound. I mean, our inbounds are up substantially. The inbound pipeline is up 100% year-over-year. Our outbound motion is newer to the company. So that muscle is newer. That's the muscle we're working on fixing now in this next phase of this transformation. And it's unfortunately one of these things where until you get into it, you don't know how much you don't know. And we're realizing it's a muscle that needs to be a lot stronger to make it successful, and we're hard-charging at it now. Our aim is that when we do our Q4 earnings release in February to give a lot more details around what makes up that Phase 2 go-to-market transformation. When should you expect a change in the outcome from it, an improvement, and how that would address the year-over-year B2 revenue growth.

Your next question comes from the line of Zach Cummins with B. Riley Securities.

Gleb or Mark, either one. I was wondering if you could give a little more context around some of the larger deals in the pipeline. It sounds like maybe those were pushing a little to the right when it came to the B2 side of the business. So any additional context around just the pipeline that you're seeing with some of these larger deals and maybe the reason that some of them pushed to the right?

Yes. Thanks, Zach, for the question. So one thing I'll say is when I looked at some of the deals and dug into it with Jason on the sales side, frankly, there wasn't any pattern in them. They were kind of random reasons. signer got sick, a different project internally came up that has to take priority, etc. So there wasn't any kind of clear pattern for why, but it just spoke to us about that in the small deals, a lot of times what happens is it's one person who is the person that's interested, the decider, the buyer, the purchaser, the user, and they go in, they make a decision and they go. In these larger deals, it's just more complex where you have a buying committee, you have the various different oftentimes security compliance reviews. There are sometimes different departments that need to be involved for how it's going to get used, how the migration is going to happen, etc. So it's just one where as we're moving upmarket, we're closing these bigger deals, and that's fantastic. And at the same time, we want to be realistic about that some of them take longer than I think we were seeing because we also have seen some large deals close very quickly. The BT Overdrive deal at the beginning, the first one that we announced last quarter closed incredibly quickly. Some of the customers that I talked about in my prepared remarks, when I was looking at them, the 6-figure deals that from start to finish were three months. And that's quite quick for a 6-figure deal from the first conversation to fully onboarded. But we're also realizing that not all deals move that quickly.

Understood. And in terms of the Phase 2 of the transformation, I'm sure we'll get much more detail in your Q4 earnings call. But can you give us a sense of some of the things you're looking to improve within the self-serve motion? Is this largely targeted towards some of these faster-moving higher data usage customers and reducing the friction there? Or anything you can provide there would be helpful.

Yes, it's interesting because we built the company primarily on a self-serve model. When we went public, almost our entire business was driven by this approach, with only a very limited sales effort in place at that time. A significant advantage was our blog, which attracted millions of readers each year and generated considerable inbound interest. Customers could easily sign up by providing their email and creating a password, allowing them to try our services before deciding to enter their payment information. As we targeted larger markets, we stated at the beginning of the year that our primary focus would be on this strategy, with minimal changes to our self-serve model. However, our team recognized that the SEO landscape was shifting, particularly as companies and individuals began utilizing AI tools like ChatGPT for their information searches. Our team proactively updated and optimized our content to appeal to these new applications. Consequently, we observed a 56% increase in account creation for our self-serve model. While many other companies reliant on self-serve or inbound content experienced declines due to their failure to adapt to new AI-driven search algorithms, our team effectively supported this transition. Moving forward, in Phase 2, our focus will be on how new AI-native startups and developers integrate into our platform. We aim to facilitate seamless adoption and learning experiences throughout their workflows, ensuring that our content, guides, processes, and integrations address their data-intensive needs. For instance, a customer I mentioned earlier began using our self-serve option and over time, as their demands for performance and larger deals increased, they reached out to our sales team for assistance and ultimately purchased our B2 Overdrive product. That kind of progression is what we aspire to see.

Your next question comes from the line of Mike Cikos with Needham.

This is Jeff Hopson on for Mike Cikos. I just wanted to see if there is any more info on how the Powered by white label solution is going. Obviously, NeoClouds is coming more popular trend, and it seems like that could be a good place for that or any sort of partnership to offer their customers more flexibility. So maybe just any info on that opportunity.

Thanks, Jeff. Good to have you on. So we're actually pretty excited by the Powered by. As you saw in my prepared remarks, one of our largest deals this quarter was a Powered-by deal. We also have seen some of our channel partners actually adopting Powered by where instead of trying to resell and integrate at the customer level, they actually build it into their own offering and offer it directly. And then, same like you said with the NeoHubs, we have various discussions that are in progress around that front. It's certainly an exciting opportunity. There's about 200 of them out there. So they all have GPUs, and they all have storage needs. So we engage with them today in a variety of ways where we service our customers using the NeoCloud by providing this open platform, free egress, high throughput, but also are in conversations with some of them to help them with their storage needs directly.

And maybe hardware storage has been on the top of investors' minds recently as AI video generation comes into the forefront. And I know you guys called out some AI video wins with customers. Just curious if you're seeing an actual uptick in AI video companies just in the past 6 to 9 months, as those models have kind of have become more popular.

I'm sorry, can you repeat the very first part of the question? Did you say hardware storage? So initially, AI was all about text, right? It was generative AI for text. Then it became generative AI for images, then audio, and now video. Obviously, each of those is in order of magnitude bigger in terms of the data sizes. So, video is a very data-heavy cloud format. So we absolutely are seeing customers signing up for that. I was looking, we had one of the interesting AI, Gen AI, video companies was a recent customer this quarter. There was another large one in the prior quarter. We have hundreds and hundreds of AI companies. So obviously, I don't know all of them, but just as they come up, I see ones that catch my attention. So it's something where they create models for the video that requires a lot of data on the front end, then it requires the data to generally get sent to one of the Neo cloud providers for the model creation. And then the inferencing itself, where they're generating the video, that video needs to go somewhere. And so we're supporting customers on the various fronts. One thing I'll tell you is that AI in general is a space that we're seeing significant adoption. Today, about 1/4 of all of our new business is coming from AI companies. Over time, it's going to become harder to say what is an AI company and what is just a company using AI. But today, about 1/4 of that new business is actually coming from companies that are specifically in the AI space. So it's an area that we're excited to lean into.

Your next question comes from the line of Rustam Kanga with Citizens.

Great to see the outperformance and new high watermark on the net income and adjusted EBITDA there. Mark, given your prepared remarks, it sounds like there's going to be a lot more honing in on the operational go-to-market skills in Phase 2. And given the success from Phase 1 moving upmarket, perhaps you have some better picture on the talent most equipped for what you're looking for in Phase 2. Is that reps with a prior focus on cloud computing and storage or more well versed, for lack of a better term, the language of AI? Ultimately, any insights you can share on what kind of reps have been the most successful on the direct sales front and that you might be looking to replicate in Phase 1?

This is actually Gleb. I'll start, and then Mark can join in if he wants to add as well. So first of all, what I'll say is in terms of talent, a lot of the talent is actually on the systems and operational side of it. So we're looking for a top-tier Rev ops person, the consulting firm that we're working with, we're doing a large project around both the systems transformation as well as the sales enablement and execution side. So Mark mentioned that we have a success with inbound. At the same time, what we see is there's a lot of stuff that comes at the very top of the funnel. And then there's a lot of opportunity to be more efficient and effective in having it go through the funnel. And then on the outbound side of it, one of the things that we realized is that we can do better with targeting the right types of customers who are ideal customer profiles, both in terms of identifying them and also in terms of how we reach out to them. So a lot of the work in the transformation isn't even specifically about the different reps. It's about the infrastructure of people and systems around the reps to help them.

This is Mark. Our reps have done well. Our win rate is 30% from opportunity to close, which is a very healthy win rate. So we just want to be flowing more opportunities through that team. And then as that opportunity flows through that team, you start expanding capacity and optimizing capacity within. So as Gleb said, there's a lot more around making things available to every stage of the process in a way where people are following up at the right time. They got the right content, the right messaging, the multimodal approach to the customer, when you place advertising versus e-mail versus text message. So there's just an incredibly scientific way of doing that these days. And the people helping us are people who help companies like Snowflake and Databricks. So they bring the best-in-class on that process. And we want to adopt it and drive a lot more volume through there.

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

Thank you, everyone, for joining us. With the double beat this quarter and how well we're positioned to help companies with AI workloads, we're enthusiastic about our opportunity ahead. I want to thank our employees, our customers, our partners, and our investors for being on this journey with us as we build this core storage backbone of the Internet. For our investors, we look forward to seeing you at the Needham Conference and at the Craig-Hallum conferences this month and chatting with all of you next quarter. Thank you.

Ladies and gentlemen, this concludes today's call. Thank you all for joining. You may now disconnect.