Hello all, and welcome to Northern Data Group Q2 2024 earnings call. I will now hand you over to your host, Aroosh Thillainathan, to begin. Aroosh, please go ahead.
Thank you, Azra. Good afternoon, everyone, and welcome to our Q2 2024 earnings call. My name is Aroosh Thillainathan, CEO of Northern Data Group, and today I will be presenting alongside our Group COO, Rosanne Kincaid-Smith, and Elliot Jordan, our Group CFO. Thank you all for joining. Today, we will give you an update on Northern Data Group's progress in the first half of 2024. Before we start, I must remind you that the presentation is conducted subject to this disclaimer. This is a regular disclaimer, which you well know, but as a reminder, included in this presentation will be forward-looking statements and underlying assumptions. We will not read the disclaimer, but assume it, as has been read for the purpose of this meeting.
In terms of our agenda today, we will start by providing an overview of Northern Data Group's vision and strategy, and update you on our operational and financial performance in the first half of 2024. At the end of the presentation, we look forward to answering any question you may have. I'd now like to hand over to Rosanne, who will take you through the first section of today's presentation before Elliot update on the financials. Over to you, Rosanne.
Thank you, Aroosh, and good afternoon, everyone. Before we look at our most recent set of results, I'd like to take the time to highlight Northern Data Group's vision and strategy. Northern Data Group sits at the nexus of three market-shifting technology trends. First, an exploding Generative AI market. The Generative AI market is expected to rise to more than $1.3 trillion by 2030. Generative AI tools for speed and automation, and their significant benefit to life sciences and education, ensure that artificial intelligence is poised to deliver true societal progress in the long term. Second, the rapidly expanding demand for data center capacity, driven by both the advancements in cloud technology and Generative AI, will not only underpin the AI revolution, but also provide a platform for uniquely sustainable solutions for energy use and supporting technologies like liquid cooling.
Of course, there is a growing global adoption of digital assets into the wider financial ecosystem. Building on our strong foundations, leveraging our expertise and commitment to operational excellence, we continue to grow and foster meaningful partnerships with customers and technology partners across our markets. We offer diversified, end-to-end, high-performance computing solutions, utilizing valuable industry partnerships with NVIDIA, HPE, Supermicro, Gigabyte, and VAST Data, to name just a few. With the ongoing support of our major shareholders, as demonstrated by our recent capital raise and our close supply partnerships, we are truly empowered to make investments and deliver on our technology-first strategy. Importantly, we have put in place the corporate structure to allow for sustainable growth across our group, deploying assets and expertise to capitalize on new opportunities. Our path to solidifying our position as a leader in HPC solutions is clear and is quickly being realized.
Seamless execution is our absolute focus in 2024. The delivery of our technology and product roadmap is paramount, and inherent in that is the rapid deployment of our H100 GPUs now and for the remainder of this quarter three. With clients onboarding as capacity comes available, our success is building, and we are on track to meet our targets for the year. Looking ahead to 2025, you will see us accelerate. We will scale. We will capitalize on growing market opportunities. We will deliver sustainable, predictable, repeatable growth. We will expand our technology and product estate, establishing our leadership position in the technology market, and we will continue to invest for future growth. In 2026, we will double our ambitions.
We will develop new cutting-edge products and solutions while forging new strategic partnerships, which will enable us to increase our market share, and we will realize scale and optimized capability while continuing to deliver value to shareholders. Now, turning to our performance in the first half of this year. We've had many highlights for the first half of this year. We have more than 12,000 GPUs live, with customers onboarded or in the process of being onboarded to all available GPUs. We have continued to plan for our next phase of growth, staying ahead of the market and initiating strategic partnerships that will further extend our capabilities. These have included being selected for NVIDIA's Grace Blackwell partnership and purchasing next-generation NVIDIA H200 GPUs. We also entered partnership with VAST Data to deliver scalable storage services, and with Supermicro, a known leader in the technology OEM space.
Our data center business has also expanded its capacity, breaking ground on the recently acquired Pittsburgh, Pennsylvania, site and commencing a retrofit to make it a cutting-edge, high-density data center, which will deliver 20 MW of capacity on completion. This means that our own portfolio now has 60 MW of capacity. Additionally, we have 150 MW of brownfield targets under LOI and more targets under strategic review. Our own capacity is nicely complemented by our ecosystem of third-party co-location partners. Mining has added a partnership with Penguin Infrastructure, delivering an extra 28 MW of mining capacity in Paraguay, and as a first, powered by 100% renewable energy. Mining remains on track to deliver 7.9 exahash by year-end.
A second 300 MW ERCOT-approved site in Corpus Christi, Texas, was acquired, providing optionality for the development of new HPC or cutting-edge data center location facilities. Finally, in the first half, we officially launched our AI Accelerator, an initiative which we are truly proud of, and a concept offering the world's most innovative startups access to Northern Data Group's generative AI platform, empowering them to amplify innovation bravery. Momentum in our cloud platform was driven by further progress of our GPU deployment and the onboarding of customers. Across the first half of the year, we deployed over 12,000 GPUs. We expect to have fully deployed our more than 20,000-strong NVIDIA H100 estate across 6 locations by the end of September 2024, and our recently acquired NVIDIA H200s will further expand our offering in late Q4.
Our deployment has been strong, and it's important to distinguish between the deployment of GPUs and the proportion of GPUs being billed in any quarter. Particularly when onboarding large customers across many thousands of GPUs, there is almost always significant testing and tailored configuration required before billing will commence. We are proud to be a provider that offers its customers bespoke solutions, which will in turn foster long-term commercial relationships. Our structured deployment schedule and our commitment to operational excellence in onboarding customers represents undeniable progress towards solidifying our position as a leading HPC solutions provider, and of course, further investments are in the pipeline. I will now hand to Elliot, who will talk us through our financial performance in the first half, as well as our financial outlook for the rest of the year.
Thank you, Rosanne. Good afternoon, everyone. I'm delighted to be sharing with you the Q2 and overall H1 financial position for the group. This is an exciting time for Northern Data as we successfully execute against our 2024 plans. Our Q2 financial performance and forward trajectory demonstrates the success, particularly evident in revenue delivered from cloud services, which increased 56% quarter-over-quarter and up 236% year-over-year. In conjunction with delivering revenue growth, we have been deploying capital into essential infrastructure, talent, and technology to ensure we can deliver against our longer-term strategy, and this means we are setting ourselves up to achieve sustainable, high-margin revenues moving forward. In Q2, overall, the group achieved EUR 26 million in revenue, up 22% year-over-year.
This was slightly down versus Q1, due in part to the halving of Bitcoin in April, as expected, offset by the significant increase in revenue quarter-over-quarter from the Taiga Cloud business. Overall, H1 revenue was EUR 55 million, up 49% year-over-year, with both cloud and mining increasing year-on-year over the half. The group achieved a break-even adjusted EBITDA position for the first half of the year, which was in line with expectations after the slightly negative position in Q1 was offset by a slightly positive position in Q2 on the back of the stronger cloud revenue. I expect Q3 will see a further increase in revenue overall, driven by both cloud revenue and mining capacity increases over the summer. This will in turn drive a positive adjusted EBITDA position for Q3.
Cloud increased revenue from EUR 7 million to EUR 12 million from Q1 to Q2. This followed the increase in NVIDIA H100 GPUs available for our customer base to use. We continue to see strong demand for this infrastructure, attracting a global customer base to the European-based infrastructure. H100 GPUs are currently deployed in our facilities in Sweden and Norway. Mining revenue stepped back from EUR 22 million in Q1 to EUR 14 million in Q2, driven by Bitcoin halving, but well controlled through maintaining our exahash capacity and driving 87% hardware uptime, which is consistent quarter-over-quarter. Mining revenue is down 21% year-on-year, principally due to the Bitcoin halving, but this was offset by an increase in Bitcoin price and higher hardware uptime at our facilities.
We expect to see mining revenue increase in Q3, with new infrastructure going live, increasing our share of the global hash rate. As we discussed at our Capital Markets Day in May, we are shifting our revenue base to predictable, recurring, and higher-margin cloud revenue. Q2 delivered another step in this direction, with cloud revenue increasing from 26% of overall share in Q1 to 46% of overall revenue in Q2. Our expectations is that circa 60% of 2024 full- year revenue will be cloud-driven, increasing to 75% of our overall revenue position in 2025. This is important, as over the medium term, we would expect to see our cloud EBITDA margin above 60%, driving the bulk of our EBITDA position and our historical reliance on mining EBITDA to drop back as a contribution to the group.
As a result, this means that, EBITDA position will be driven 10%-15% from mining, which is a positive step in the right direction as we drive our cloud business to deliver strong, high-margin revenues. This H1 position and our H2 GPU deployment plans, as well as our strong customer pipeline, means we remain on track to achieve our previously stated 2024 financial goals. We continue to expect EUR 200 million-EUR 240 million in revenue and EUR 50 million-EUR 80 million in adjusted EBITDA for the full- year of 2024. This includes a sequential increase in revenue from Q2 into Q3 on the back of additional cloud revenue from new clients using recently deployed GPU capacity and new deployments of GPUs going live across the quarter, plus an increase in mining revenue from new sites live in August.
As Rosanne explained, we make sure our infrastructure is tailored to the requirements of our customers, including testing and configuration. This means revenue increase Q3 from existing deployments will be higher than we achieved in Q2, with a significant increase in utilization and billing. Q4 is expected to be our strongest quarter of the year, with well over EUR 120 million in revenue, driven by the 20,000 NVIDIA H100 GPUs that will be deployed as we exit Q3 and the full 7.9 exahash capacity in mining. This Q4 exit rate puts us on a strong footing to see strong overall revenue growth in 2025. Turning to cash, the group remains with strong levels of liquidity. We had a June cash balance of EUR 350 million and a net debt position of EUR 225 million.
Our cash balance increased post-quarter end by EUR 214 million we've achieved because of the capital increase we announced in July. I'd like to outline the key cash movements in the first half. Total financing across H1 was EUR 685 million. This included drawing EUR 400 million from the five-year loan facility we obtained last year. This facility is now fully drawn at EUR 575 million. In addition, we closed the acquisition of Damoon, which had a cash balance of $285 million, which we've included under financing. These funds were included as other receivables as at the thirty-first of December, 2023. As expected, we have been very active in deploying these resources year to date, with a total of EUR 600 million in cash paid out for capital investments.
This includes EUR 495 million into Taiga Cloud Infrastructure, primarily GPU acquisition and deployment, EUR 15 million in Ardent as we start to retrofit our Pittsburgh, Pennsylvania, location, and EUR 90 million on mining infrastructure as we grow our mining capacity through the year. Note, we also had EUR 55 million in receipts from hardware sales, so net cash flow from investing was EUR 545 million. Finally, H1 operating cash flow was -EUR 35 million as we continue to ramp up revenue to cover the cost base. For the full- year, we continue to remain well-funded, with an estimated year-end closing cash position of EUR 100 million-EUR 200 million.
Cash flow into investments is now expected at EUR 1 billion, including a full-year total of EUR 800 million in investments within Taiga Cloud, including the purchase of NVIDIA H200 GPUs announced in July. We also expect to continue the build-out and retrofit of our Pittsburgh, Pennsylvania, data center location, as well as the development of a second data center investment in the U.S. Finally, we will complete the fit-out of our North Dakota and Texas mining locations. You will note we expect positive cash flow from operations across the full- year, driven from the strong cloud-based revenue in Q4. This is set to continue into 2025 as we are fully operational over our cloud technology, driving positive operating free cash flow. As we look ahead into next year, we are well set up to be delivering revenues of over EUR 500 million.
This is because we have built the infrastructure needed to support a business that can achieve this level of revenue with high-margin expectations flowing. Our total guidance for 2025 is revenues of EUR 520 million-EUR 570 million and an adjusted EBITDA of EUR 300 million-EUR 350 million. 2025 will be a high-margin phase of our GPU asset life cycle, achieving strong margins, positive free cash flow, and an ability to take our infrastructure and investment plans to the next level. I look forward to updating you on progress after Q3, but for now, I'll hand you back over to Rosanne.
Technology has kept me at bay, but thank you, Elliot. Moving on to our priorities for the remainder of the year and our roadmap for success in 2025. We will continue to break boundaries in H2, which holds exciting and large milestones, along with doubling down on our delivery. We will continue to rapidly deploy our H100 GPUs and onboard more customers as capacity goes live. Implementation of our investment in NVIDIA H200 technology, alongside further expansion of our technology estate, will be our primary focus. Other notable projects, including the substantial upgrades taking place to our data center in Pittsburgh, Pennsylvania, and in Boden, Sweden, along with co-location partnerships, will enable our accelerated growth. Our investment in our Paraguay sites and our first Corpus Christi site will maintain operations in mining.
We have generated strong momentum over the last six months, and we are poised for strong continued commercial growth into 2025 and beyond. As first announced at our Q1 Capital Markets Day in May, we are proud to be one of just 11 sovereign AI clouds globally to be selected as a partner for the latest Grace Blackwell technology by NVIDIA. The future's greatest opportunities will be driven by giant technological leaps, and Grace Blackwell is truly the embodiment of that giant technological leap, both in compute power and energy conservation. Our commitment to liquid cooling means our technology is uniquely suited to support this high-density, ultra-performant configuration of the GB200 Superchip. This shows that our advanced technology foresight, state-of-the-art configuration, and commitment to providing cutting-edge data center environments has primed us to be the perfect partner for the next generation of high-performance computing.
These foundations are the drivers for continued commercial success, not just today, but also to meet the growing future demands of the developing inferencing markets. We will grow our high-margin cloud offering, which will be supported by sizable investment in our data center footprint, software advancements, technology innovation, and we'll be delivering these solutions at scale to market. Our vision and our North Star are clear, and we are poised to harness the opportunities the future will bring. I will now hand back to Aroosh to summarize today's presentation and invite questions.
Thank you, Rosanne and Elliot, for today's update. To conclude, we are proud of the progress made so far this year. We are wholly focused on advancing the technology which creates leading high-performance computing solutions and are well positioned to continue to execute on our strategy in the second half of the year. We have built a strong foundation with diversified and clear core capabilities, commercial clarity, backed by industry expertise. As a result of this foundation, Northern Data Group has achieved consecutive successful capital raises in 2023 and 2024, securing nearly +EUR 1 billion in debt and equity funding. As such, we are in a strong position financially and primed for delivery. We have executed rigorously against our roadmap in 2024, and our first half results show that we have begun to see the expected financial results.
We are well positioned to accelerate into 2025, and we are very excited for the next 6-12 months. We hope this has been insightful and our strategy, roadmap, and positioning for future growth has come across clearly. We will now conduct a short Q&A to answer any question you may have.
Thank you very much. To ask a question, please press star followed by one on your telephone keypad now. When prepping to ask your question, please ensure your device is unmuted locally. If you change your mind, please press star followed by two. Our first question is from Lucas Pipes with B. Riley Securities. Lucas, your line is now open. Please go ahead.
Thank you very much, operator. Good morning, everyone. This is Nick Giles asking questions on behalf of Lucas. Guys, congrats on all the progress so far. In the release, you noted that multiple acquisition projects are in the due diligence phase. Should we think of these as greenfield opportunities or brownfield? And is there any details you could share from a geographic perspective? Thanks very much.
Hi, Nick. Absolutely. So our strategy to date for the Ardent business has been to have a mixture of retrofit, greenfield, and brownfield sites that we can capitalize on to meet the growing demands of that particular industry. The acquisition targets that we are focused on at the moment are almost exclusively retrofit or brownfield, and we are intending to close on some of those and announce some of those quite imminently.
Rosanne, that's very helpful. On the Pittsburgh site, you know, I believe it has immediate capacity of 5 MW, plans to upgrade to the 20 MW. How should we think about potential timing for those incremental 15?
So the first 5 MW, which we've actually been able to now increase to 6 MW, will be available by the end of the fourth quarter. The remaining megawatts are due to be completed by the end of 2025. And I would just add to-
Got it. Got it.
Just sort of as an additional point to that, to that timing, that entire site is being upgraded, not just for direct-to-chip liquid cooling, but to be Blackwell capable, which means rack capacities will be increased quite significantly and well up to 200 kilowatts power. So it is a significant undertaking, which is slightly more advanced than even just the basic direct-to-chip liquid cooling.
Well, Rosanne, I think you answered my next question, but I'll ask it anyway. Just wanted to get some more color around deployment plans for the Blackwells. I guess my question now would be, would these occur solely in Pittsburgh or, if not, some in Europe, where would these occur? Thanks very much.
So Blackwell is part of our broader opportunity. As I said, it's really a part that is very exciting for me and certainly particularly for me, because I'm a little bit of a super nerd. But those chips will certainly. We'll be looking to deploy them across our portfolio. So certainly in Europe, in Pittsburgh, for sure, we will be able to support the direct-to-chip liquid cooling, as I said, for Pittsburgh, certainly that will be Blackwell capable, and we will have a deployment there. And then we will have much wider deployments in Europe, but more on that to follow.
Got it. Got it. Well, guys, again, congrats on the progress so far, and continue. Best of luck.
Thank you.
Thank you very much. Our next question is from Gerard Orgonas with Berenberg. Gerard, your line is now open. Please go ahead.
Yeah. Hi, good afternoon. I have a question on your H100 deployment. Compared to your Q1 presentation, you're about 4,000 GPUs behind the schedule that you were expecting at Q1, but you expect to recover this by September. Could you tell us a little bit what is the bottleneck currently? Or, and how are you going to overcome this by September?
So I can add a little bit of color to that, Gerard. Like every other supplier in the H100 ecosystem, we are subject to supply chain constraints, which we have been experiencing, and have been remedied, and that is why we are confident that we will unplug that bottleneck. It's no secret in the market at the moment that there is a shortage of Mellanox and transceivers. Fortunately, as I say, our strong supplier relationships have meant that while there has been some shifting in our deployment plans, it is not, it has been quite marginal in comparison to what others have experienced.
This is not the H100s per se, this is other equipment or?
The actual server units are not the bottleneck, no. It's the other equipment-
Okay.
that relies on the architecture.
Okay. Will all the H100s be deployed in Sweden and Norway?
Sweden, Norway, Portugal.
The capacity for the 20,000 is secured, yeah?
Correct.
Okay, and then just one more question. You alluded to a time difference between deployment and revenue generation with large clients, and there could be some testing. So in the light of that, do you still expect to have a fully ramped Q4 run rate revenue?
Yes, we do.
Perfect. Thank you.
Thank you very much. This concludes our Q&A session. I will hand back over to Aroosh for any closing remarks.
Thank you, Azra. Thank you everyone for dialing in. Please reach out if you have further questions directly to the company, and we will get this addressed. Thank you very much again for dialing in, and hope to speak to you soon.