Prepared remarks
Good afternoon, ladies and gentlemen, and welcome to RUM Group Second Quarter 2026 Earnings Conference Call. This call is being recorded on Monday, August 10, 2026. I would now like to turn the conference over to Shannon Devine, Investor Relations for RUM Group. Please go ahead.
Thank you, operator. I'm here today with Chris Pavlovski, Founder, Chairman and CEO of RUM Group; and Mike Masci, CFO. A press release detailing our second quarter 2026 results was released today and is available on our Investor Relations website. Before we begin the formal presentation, I'd like to remind everyone that statements made on this call may include predictions, estimates, or other information that may be considered forward-looking. All forward-looking statements are made only as of the date of this call and should be considered in conjunction with the company's cautionary statements in our earnings release and the factors included in our filings with the SEC. Future company updates will be available via press release and the company's identified social media channels. I will now turn the call over to RUM Group's Founder, Chairman and CEO, Chris Pavlovski.
Good afternoon, everyone, and thank you for joining us. Last quarter, I told you this would be the last call before Rumble meaningfully entered the cloud and agentic AI era. Today, I'm glad to say that transformation is complete. On June 17, we closed our acquisition of Northern Data, securing approximately 85.2% of its outstanding shares. And with the deal closed, we renamed our parent company, RUM Group, Inc. We now operate two distinct, synergistic business units: Rumble, our video platform; and Quake AI, our new cloud and AI infrastructure business, combining Rumble Cloud with Northern Data's estate of roughly 22,000 NVIDIA H100 and H200 GPUs. To kick off our first earnings call as a combined company, I'm thrilled to announce that our revenue for the second quarter was $40.4 million, up 61% from $25.1 million in the second quarter of 2025. I'm proud to say it's been nearly five years since we became a public company, and we're still posting all-time records for our company, and we anticipate that we will post another all-time record in the next upcoming quarter.
I want to spend a minute on the strategic logic here because I think it's important for everyone on this call to understand where we are headed. Rumble spent years building its own rails as a streaming platform, our own bare metal compute, our own CDN, and the network to deliver low-latency streaming at scale. Combine that now with an AI compute-as-a-service business like Northern Data, and you get a compelling end-to-end AI infrastructure company. That's Quake AI, and it's going to be the financial engine of RUM Group going forward. On execution, Quake AI's existing GPU estate is running at more than 85% utilization today, up sharply from where it stood not long ago. That improvement reflects a deliberate focus on customer support, software, and Infrastructure-as-a-Service execution, and it's what gave us the credibility to win the next stage of growth. In June, we signed a multiyear agreement with Together AI to deploy NVIDIA HGX B300 GPU capacity, establishing RUM Group as a credible independent provider of large-scale AI infrastructure outside the traditional hyperscaler ecosystem.
It validates that our customers trust our execution enough to partner with us and that NVIDIA is willing to support us with the supply allocation to keep growing. The next stage for this team is monetizing our 250 megawatts of power targeted for 2027, the grid connections and agreements that put us in position for real deployment next year. That capacity spans our marquee 180-megawatt powered site near Atlanta, Georgia, one of the most important internet points of presence in the country. We continue to see strong customer momentum as well as meaningful progress on the site development, where the substation is already built and the transformers are on site. In addition, we also have a smaller site in Pittsburgh and two European sites; roughly 50 megawatts in Sweden and roughly 20 megawatts in Norway. We continue to see encouraging and unprecedented growth in demand for AI compute as a service, and our customer and demand pipeline remains strong.
Simply monetizing 250 megawatts of currently unmonetized capacity represents what we believe is a $3 billion-plus annual run-rate opportunity for RUM Group. Turning to our video business. Average global monthly active users were 57 million in the second quarter, and ARPU was $0.48, up 20% quarter-over-quarter. Our management team's focus is to continue to increase revenue by bringing in brand advertising. Additionally, with the recent formation of Quake AI, we are quickly learning from our AI clientele that Rumble may have a much more compelling monetization opportunity. As the AI industry moves into the robotic era in the coming years, Rumble's spatiotemporal data, otherwise known as video data, becomes increasingly valuable and very important to robotic learning. For example, contextual data, like what Reddit has, is a prime example of the value in today's world of AI, but we believe in one to two years, this will shift heavily towards video data.
And only a week ago, it was reported in the news that Amazon was exploring ways to tap into Twitch's video data. In fact, Quake AI clientele have already expressed interest in our video data. Our team is looking at various ways to capitalize on this opportunity and add another bucket of potential revenue for the creator community. With significant spatiotemporal data on Rumble, combined with the AI compute rails we are building and deploying at Quake, RUM Group sits in a very unique position compared to today's neoclouds. Like them, we can offer scaled AI compute, but unlike them, we have a trove of video data and a creator community that can help power the robotics era. In short, we have data, we have the rails, and we have the community to power the future of AI, which we believe is the robotic and agentic AI era. With that, let me turn the call over to our CFO, Mike Masci, who will walk you through the quarter in more detail, along with some important updates on our guidance process and reporting on our businesses going forward.
Thanks, Chris, and good afternoon, everyone. This was a landmark quarter, both operationally and financially. Let me start by walking you through the high-level financials. First, revenue for the second quarter was $40.4 million, an increase of $15.3 million or 61% compared to $25.1 million in the second quarter of 2025. Taking a turn to our expenses. We continue to make strategic investments to best position ourselves for high growth in AI. Cost of services were $30.6 million, up from $26.5 million a year ago, driven by higher programming and content costs as well as the incremental data center expenses from the Northern Data acquisition. General and administrative expenses were $16.3 million, up from $11.7 million, again primarily driven by Northern Data, which contributed $5 million of payroll and other administrative costs. Excluding Northern Data, the remaining increase reflects higher payroll and other administrative costs, partially offset by lower professional fees.
Research and development expenses were $6.8 million, up from $4.8 million. Sales and marketing expenses were $10.4 million, up from $7.9 million, attributable to higher marketing and public relations spend, increased payroll, and other sales and marketing-related expenditures. Adjusted EBITDA loss for the quarter was $16.6 million, an improvement from a loss of $20.5 million in the second quarter of 2025. Net loss for the quarter was $80.3 million, or $79.1 million attributable to RUM Group, Inc., compared to a net loss of $30.2 million in the second quarter of 2025. The year-over-year increase in net loss was primarily driven by $28.3 million of acquisition-related transaction costs associated with the Northern Data close, along with higher noncash depreciation and amortization following the acquisition. We ended the quarter with total liquidity of $220.5 million, consisting of $203.3 million in cash and cash equivalents and 293.14 Bitcoin valued at roughly $17.2 million.
Before I get into guidance, I want to talk you through an important change in how we will report going forward. This will be the last quarter we present MAU and ARPU as headline metrics for the company. Those numbers described a single video audience business. Following the Northern Data acquisition, that's no longer the full picture of what RUM Group is. Beginning with our third quarter report, we plan to move to segment reporting: Rumble Video and Quake AI, each with its own revenue and profitability metrics. I think that's a far more useful way for all of you to understand where we're allocating capital and how each part of our business is actually performing. And it's the reporting structure I'll be building around as I get further into the seat. So that brings me to guidance. For the first time, we are issuing formal revenue guidance. For the third quarter of 2026, our first full quarter reflecting Quake AI, we expect revenue between $87 million and $93 million.
We've heard directly from many of you that formal guidance would help you better understand this business, and now that Northern Data is closed with strong contracted revenue and high utilization, we felt this was the right moment to give you that visibility. Over time, as our forecasting visibility continues to improve, we plan to build this out further, including longer-term views. Building a disciplined, credible guidance process is something I'm personally focused on as we establish our track record with all of you. So in summary, this was a transformational quarter: record revenue, both including and excluding Northern Data. We closed Northern Data and established Quake AI as a leader in AI compute as a service. We signed a marquee agreement with Together AI, and we are now focused as a team on converting our 250 megawatts of unmonetized power targeted for 2027 into what we believe is a $3 billion-plus annual revenue run-rate opportunity.
I was excited when I came on board at the end of March. Today, I'm more excited about where RUM Group is headed, and I look forward to updating you on our progress. That concludes our prepared remarks. Operator, we are now ready to open the line for questions.
Questions and answers
And your first question comes from the line of Tom Forte with Maxim Group.
This is Henry Dare filling in for Tom Forte. I have two questions. Firstly, Chris, there's some debate among investors as to what happens to pricing for AI when capacity catches up with demand. I would appreciate your thoughts on that matter.
Henry, this is Chris. There has been a lot of debate about that. The way I look at it right now and what we see in our window is that we're still in the early stages of AI. In particular, I think that with inferencing and agentic AI exploding in the coming years, the demand for AI compute is going to continue to increase. I don't see at this point, in the near future, any point where capacity is going to catch up to demand on the AI compute side. I think it's scarce. It's very scarce. I believe at this point in time, I don't see any point in which capacity will meet that demand in the next one to two years.
Okay. And this may be too early as you just finished the Northern Data transaction, but I would appreciate your current thoughts on your strategic M&A strategy and the types of companies you consider adding moving forward.
Yes. Like you said, it's early in our process in terms of how we're thinking about moving forward. Right now, as a management and leadership team, we're really focused on that 250 megawatts of unmonetized capacity. We'll always look at opportunities that create value for customers and shareholders. But at the moment, we are focused on monetizing that 250 megawatts.
And your next question comes from the line of Jason Helfstein with Oppenheimer.
Congrats on giving guidance; I don't often congratulate on calls, so glad we hit this milestone. To dig in on some questions: when you think about getting what you need to get the compute going — you've gotten the cash now — help us understand the challenges around getting the chips, rack cooling, and all of the parts you need for the data center. How much has already been contracted and signed? Help us understand timing of when that comes online. Also, given how challenging it is to get all of the equipment to get data centers open, have you thought about potentially doing the powered shell model where you run the infrastructure and customers bring the chips and the racks? And I have a follow-up on advertising.
Jason, this is a record-setting quarter for us, and it's great to hit the milestones. On your question about CapEx, when we look at agreements like the one with Together AI, we don't sign the contract and the agreement until we have strong line of sight to secure the right amount of equipment and capacity to put that online. You can see from our cash flow statements that this past quarter we had a large uptick in our investing activities, upwards of almost $47 million, which is unusual for us. That really went towards purchases of IT CapEx necessary to execute that AI compute-as-a-service deal. Think of that as the start of the process, and we feel really good about our ability to get supply to ultimately fulfill our obligations for key deals, including the one with Together AI. Regarding the powered shell model — sometimes called the powered shell model in the industry — the idea is you build the data center and then lease it to someone on a long-term basis to do AI compute as a service.
We addressed some of this in our investor deck. It's a valid business model, but what we found is that the unique ability to deliver AI compute as a service is differentiated in the industry. It's one of the things that really interested us and is a huge value add from Northern Data. They've been doing AI compute as a service for a number of years. To make that real, in our investor deck we show that monetizing power and shell is roughly $1.5 million to $2 million per megawatt per year, whereas delivering AI compute as a service on the Blackwell generation delivers almost $11 million per megawatt per year. That's on the order of 5x in terms of value creation when you deliver AI compute as a service. It is not without cost — securing the AI compute hardware and operating complex machinery is challenging — but Northern Data and RUM Group have built the capability and partnerships to execute. We feel like AI compute as a service is our business model moving forward.
Jason, I'll add that with the Rubin chipset coming in the next year or two, that could even increase monetization further. We heard from other industry participants that quoted numbers much higher for the Rubin.
Okay. So the main point is whenever you make a public announcement, you've already secured essentially the hardware to do that. That's how investors should think about modeling this — building as you go with those announcements. Will that be the policy? Should we assume every major customer win you announce will have the supply and that timing will be reflected?
First, what I described is how we're thinking about the operating model. I would be careful labeling it a strict policy. From an operating perspective, our job is prudent risk management, so we wouldn't enter contracts we didn't feel we could fulfill. Supply allocation and execution are part of our diligence before signing a deal. We will remain disciplined in that process.
That covers that. On the ad business, there's a nice acceleration year-over-year. Can you tell us how much of the quarter was related to Tether ad commissions, if you're willing to break that out?
Yes. We did include this in the quarter. We had roughly $4.8 million related to Tether in Q2.
And I'm showing no further questions at this time. Ladies and gentlemen, this now concludes today's conference call. Thank you all for joining. You may now disconnect.