Napatech A/S (NAPA) Earnings Call Transcript & Summary
May 16, 2024
Earnings Call Speaker Segments
Lars Boilesen
executiveGood morning. I'm Lars Boilesen, CEO of Napatech. I'm pleased to welcome you all to the Napatech Interim Management Statement presentation for the first quarter of 2024. Joining me today is our Chief Financial Officer, Heine Thorsgaard. Our interim management statement was released earlier this morning on the Oslo Stock Exchange and is also available on the Napatech website. For your information, a recording of this webcast will be available later today. [Operator Instructions] Please note that this presentation contains forward-looking statements that are subject to risks and uncertainties. Our actual results may differ from those discussed in forward-looking statements. Today's agenda will cover 4 main areas: First, we'll provide a status on the first quarter of 2024 with focus on the major news items that are showing promising signs of progress by Napatech towards achieving our goals to transform our business. Next, we'll provide a business outlook that shares information about the opportunity ahead of Napatech. We will then provide information on the first quarter financial results. We'll finally conclude with the question-and-answer session open to the attendees of today's call. We have begun 2024 in 2 ways. On the financial metrics, we were up to a slow start of our existing capture business. However, on our new strategic data center cloud, Programmable NIC business, we were off to a very exciting start in reaching a commercial partnership with one of the biggest global server OEM manufacturers in the world in cooperation with our partner, Intel/Altera. Our focus throughout the year is executing our strategic plan as it underpins our long-term growth aspirations. To achieve those goals, we have 3 areas of focus. We must service our valuable installed base of customers who are the early adopters of Programmable NIC solutions. They are a proving ground for the technology and the value that our solution brings to their networks and businesses. These are mainly customers in the capture niche business. We are accelerating our engineering software and hardware technologies that extend our products into new areas that align with the highest projected growth segment for Napatech solutions, mainly in the data center, cloud and 5G business. And we must continue to develop an ecosystem through partnerships that expand our go-to-market reach via marketing, business development and sales, enabling us to reach places we could not achieve on our own. I'm pleased to report progress in each of these areas, and I would like to share a few highlights from the start of 2024. To begin, on April 18, Napatech announced a collaboration with Altera and Intel company on the detailed specification of Hot Springs Canyon architecture for 400-gig IPUs. The collaboration develops new Programmable NIC hardware and software solutions based on an architecture that promise to enable an innovative approach to powering modern public clouds, private cloud, edge and enterprise data centers built for an era of artificial intelligence. The new solution is purpose-built for the highest growth in largeest volume applications, services and use cases for programmable NICs. These include modern cloud and edge computing applications and services, artificial intelligence infrastructure, desecrated storage, big data analytics, cybersecurity, 5G mobile infrastructure and micro services. Next, we disclosed that the product [ designed ] with Intel/Altera will develop in partnership with a Tier 1 global server OEM that continues to lead in enterprise and cloud data center deployments. The product is packed with the latest technology from Intel/Altera and is designed with optimizations to meet the exact requirements of the server manufacturer. The collaboration is anchored by commercial agreements, associated statements of work that define the product specifications and professional service fees from both companies to Napatech. The collaboration includes these significant contributions, reflecting the shared commitment required to expedite the underlying engineering to deliver products in the second half of 2024. As a result of these partnerships, the full potential is that the largest end of the spectrum for transformational growth that Napatech envisioned. At peak annual volumes, this single opportunity alone has the potential to transform Napatech's long-term revenue and profits. Consequently, we are maintaining our target of 4 to 6 design wins while increasing our guidance from 30,000 to 40,000 units to 50,000 to 60,000 units in peak annual production years. We also disclosed important partnership with companies, including Amantya, Hyperblox, A5G, Saviah and Supermicro, who are providing solutions for 5G infrastructure and artificial intelligence and machine learning. These valuable partnerships enhance our go-to-market strategy, taking Napatech products to customers beyond our reach. In support of these significant steps forward with our current customers and expanding design win pipeline, Napatech announced and successfully completed a private placement capital raise on May 7, 2024. Napatech offered 9 million shares at a subscription price of NOK 25, approximately 10% above the market price at the time. The private placement was significantly oversubscribed and attracted interest from high-quality accounts as well as receiving strong support from existing shareholders. The net proceeds will be used to finance the growth opportunity arising from the announced partnerships, resulting in the increased guidance, including hiring of more developers and working capital requirements in connection with increased productions. And finally, we have reaffirmed our financial outlook for 2024. Our financial results for the first quarter of 2024 were off to a disappointing start. However, the start of the year was expected to be slower than normal, primarily because one of our largest customer is continuing an inventory correction to deplete a COVID-driven safety stock that is expected to be completed through the first half of 2024. In addition, March order flow was unexpectedly slow, also driven by one large customer purchase that was deferred into Q2 2024. Our financial results included revenues of DKK 21.5 million, or USD 3.1 million, and gross margins of 72%. More details on our financial results will be provided in the final section of today's presentations. Given the progress in 2024 moving towards achieving our strategic goals, in the next few slides, I would like to share more information about Napatech, our products and the opportunity opening to us. Napatech's products are key for server performance. The image to the right labeled F is a standard server. Servers are the primary system in modern data centers and are used to deliver the applications, services and revenue to business in cloud, enterprise, telecom and government networks. Historically, the most important part of the server were the central processing units or CPUs. The CPUs from companies like Intel act as the brains of the server and are labeled E in this image. CPUs are surrounded by memory, labeled D. Memory stores data, including network traffic for the CPUs to process. A new and well-published area inside of servers are the expansion slots shown as C in the figure. These expansion slots are most often used today for acceleration cards based on GPUs, or FPGAs, for artificial intelligence, machine learning, training, large language models and more. And finally, the slots label A and B show the network interface cards or NICs. In this image, 2 of Napatech's NICs are shown. Together with Intel, Napatech is developing 2 types of Programmable NICs, including SmartNICs shown in B, and infrastructure processing units, or IPUs, shown in A. Napatech's Programmable NICs play a crucial role in modern data center and sever designs. The Napatech NIC cards are the sole link between the outside cloud and enterprise networks, providing access to critical CPU and GPU resources that deliver artificial intelligence, security and many other services. The underlying problem Napatech Programmable NIC helped to solve is a server performance gap. The problem has originated from a spectrum, megatrends, creating a huge demand for computing resources. These include cloud and edge computing and storage, artificial intelligence infrastructure, 5G mobile infrastructure, financial services and other fintech uses, cybersecurity, big data, analytics and others. This explosive growth is increasing network traffic as well as the demand for servers to more than 20 million servers a year by 2027. Unfortunately, these servers cannot keep up with the demands of the new applications and the services required by the megatrends. The CPUs within the servers end up spending significant processing power handling noncore infrastructure tasks for networking and security rather than running the applications and services. This leads to decreased revenue and lower overall performance of the systems. In small-scale examples, this is resolved by adding servers. But in most networks built by cloud, enterprise, telecom and government operators to support these megatrends, servers can only scale to a point. They reach break-in point limitation in physical space, cooling cost, power consumption and the raw compute capabilities. Napatech Programmable NICs helped to bridge the server performance gap. As shown in the first image to the right, historically, the way operators solve this problem was by adding servers. But stacking servers quickly runs into scalability problems as noted in terms of space, power, cooling and not to mention costs. The second image to the right illustrates the problem and how Napatech helps. As a number of locations, services and tenants continues to increase, so too does the network throughput. In addition, the workloads that these applications are required to perform on the increase in packet data rate is exponentially increasing. In an ideal scenario, the server would scale with the demand of the applications, services, tenants and network. Unfortunately, they do not as they run into limitations in space, power, cooling, computation and costs. Given these constraints of adding new servers to the network, Programmable NIC have proved to be a solution for bridging the server performance gap. Napatech, in conjunction with our partners Intel/Altera, have created a unique and compelling solution to this problem. Programmable NICs, including SmartNICs and IPUs, are used to improve server capacity and accelerate the performance of applications by handling infrastructure process that the server CPUs would normally handle. In a traditional server approach, the server is required to run applications and services for their users and tenants, but also noncore tasks such as storage, security, network and management function. Top network operators such that Google and Facebook estimate that these noncore workloads can consume up to 80% of the most valuable and most expensive server resources, meaning the CPUs. This leads to the low server performance. In the new approach to server design, a programmable NIC, such as SmartNIC or IPU, replaces a standard NIC. The new programmable NIC offloads and accelerates the infrastructure process for network, security, storage and management. This returns the compute resources to the [ CPI ] for the applications and services that were intended, resulting in higher server performance, lower server costs and significantly reduced power space and cooling demands. Napatech is pleased to partner with Intel/Altera in creating these solutions for the largest and highest growth segments for programmable NICs. Napatech is building a family of programmable NICs, including SmartNICs and IPUs that are packed with the leading technology from Intel/Altera, including the latest generation of CPUs and FPGAs. Napatech designs, develops and delivers a complete solution for the end users. It includes the hardware as well as Napatech's main competitive differentiator, the software that powers the system. Our partnership with Intel continues to exceed our expectations. We continue to benefit from joint engineering and product management collaboration, ensuring that our products meet the most demanding requirement identified by market makers like Intel, Altera and the global scale customers. We gained access to tip of spear technology, including CPUs, FPGA and servers. We all benefit from joint pipeline building and business development, improving our go-to-market by leveraging on Intel's global reach, including sales and marketing for demand generations. Each of these benefits results in Napatech entering large and high-growth markets that increased demand for Napatech products and transformational volumes and scale. Napatech finds itself uniquely positioned. The industry megatrends discussed earlier in our presentation are creating more applications and services, which in turn create more servers to be deployed. Consequently, every server requires one or more network interface cards to provide connectivity to critical CPU and GPU resources. The explosive demand for NICs is accelerated by an underlying trend away from legacy basic and offload NICs to Napatech's style of programmable NICs. Napatech is exclusively focused on programmable NIC market with our SmartNIC and IPUs. Intel and Altera's market-making position cannot be understated. FPGA-based SmartNICs and IPUs have already been opted by the largest hyperscale network operators globally. They have proven the technical and commercial benefits of this architecture and design, and the majority of these operators are dominated by Intel/Altera. While server growth continues among the hyperscale operators, the same cost, power, space, compute and cost benefits are demanded by a wide range of other network operators. Together with Intel and Altera, Napatech is targeting this next wave of cloud, enterprise and telecom uses. Although still early in our journey, we are seeing indicators of success. As noted earlier, we disclosed an agreement with Intel/Altera to develop the next-generation 400 gigabit IPU. This happened because of the early success in both joint engineering and joint pipeline development for the 200-gig solution we launched in the fourth quarter of 2023. That success made us moving on to the next-generation IPU from Intel/Altera. It is equally important to note that this next generation was defined in collaboration with a Tier 1 global server manufacturer with proven success in high-volume cloud and enterprise networks. These and other early customer successes have substantially increased our customer engagement and their associated annual unit demand. This, in turn, has increased the value of our design win pipeline by more than 40x while delivering a complete solution of hardware and software with attractive margins. The pipeline includes the top cloud enterprise, telecom and government networks, spanning all global areas, including North America, EMEA and APAC, the top use cases of network, security and storage offload for all of the top megatrends that I mentioned before. While there is much work to do to secure new design wins that drive large volumes in their peak production years, we are making solid strides. In success, we aim to transform our business as a niche supplier to a high-volume partner to customers in the largest and highest growth segments. Our historical annual volumes in our existing business has been ranging from 5,000 to 6,000 units per year. The 4 to 6 design wins we expect to close this year will result in annual unit demand of 50,000 to 60,000 units per year when they reach peak production. This increased volume guidance is due to a new OEM client that is in the plus 25,000 unit annual peak volume category. We maintain our guidance in terms of number of design wins despite we only had one big win so far this year. We still expect to land another 3 to 5 design wins in the rest of the year. It will require hard work from the team and some tailwind in terms of timing, but we believe it can be done. In our capture niche segments, we have had good design win progress in the first 5 months of 2024. Several new customer wins for solutions in fintech, cybersecurity, monitoring and recording that give us reasonable hopes are still achieving our 2024 goals despite the slow start in Q1 in our existing capture niche, but very valuable business. The last few slides of our update provide more details on financial results for our first quarter of 2024. I would like to ask our CFO, Heine Thorsgaard, to provide this update.
Heine Thorsgaard
executiveThank you, Lars. Revenue in Q1 amounted to DKK 21.5 million, a decline of 42% compared to Q1 2023. As we communicated in our Q4 presentation towards the end of 2023, we started to feel the impact of the slowdown in the U.S. SmartNIC market, especially within the finance and telco operator segments. Q1, therefore, was expected to be a slower than normal start to the year, with March being particularly slow. We expect this slowdown in the U.S. market to continue at least for the next couple of quarters. The gross margin in Q1 was 71.6%, an increase of 16 percentage points compared to Q1 last year. The gross margin of our product revenue in Q1 was 70%. We are now past the period of extraordinarily high component costs related to components purchased in 2022, and our margins have normalized to their historical levels. Our staff costs and other external costs in Q1 amounted to DKK 40.6 million compared to DKK 38.4 million in Q1 of 2023. The increase is a result of our strategic R&D investments and follows the plans we've communicated. EBITDAC in Q1 amounted to a negative DKK 25.2 million compared to a negative DKK 17.9 million in Q1 last year. Staff costs transferred to capitalized development costs in Q1 amounted to DKK 0.9 million compared to DKK 5.1 million in Q1 of 2023. EBITDA in Q1 amounted to a negative DKK 24.3 million and EBIT amounted to a negative DKK 32.3 million. Net cash flows from operating activities in Q1 amounted to negative DKK 5.4 million. And net cash used in investing activities in Q1 amounted to DKK 1 million compared to DKK 6.1 million in Q1 of 2023. Free cash flows in Q1 amounted to negative DKK 6.4 million compared to DKK 17.2 million in Q1 last year. At the end of Q1, our cash and cash equivalents amounted to DKK 27.5 million compared to DKK 12.3 million at the end of Q1 2023. With this, I'll turn the presentation back to you, Lars.
Lars Boilesen
executiveWe maintain our 2024 financial guidance unchanged. So that means revenue of DKK 170 million to DKK 180 million; gross margin, 69% to 71%; staff expenses and other external costs, DKK 145 to DKK 155; and capitalized development costs in the range of DKK 10 million to DKK 15 million. The change here is the increased volume guidance from 30,000 to 40,000 to 50,000 to 60,000 in annual volume production. This is due to the announced new OEM client that we have categorized in the plus 25,000 annual peak volume category. We maintain our guidance in terms of numbers of design wins despite we only have one big win so far this year. We still expect to land another 3 to 5 design volume wins in the rest of the year. Our current pipeline proves it's possible. But of course, it will require hard work from the team and some luck in terms of timing, but we clearly believe it can be done. This concludes today's prepared materials. We will now open the call for any questions that may exist from the attendees. Operator, do we have any questions?
Operator
operator[Operator Instructions]
Lars Boilesen
executiveI see the -- this is Lars Boilesen, CEO. I see the question come in. Would you read the questions or should I do that?
Operator
operatorI'm sorry, Lars. Yes, if you'd like to, go ahead, that would be great. Thank you.
Lars Boilesen
executiveOkay. So first question I see, could you please give us the number of actual product sales for Intel/Altera product design so far? So as we said, most of our products or more or less all of them are big projects for cloud, data center 5G. So the sales so far has basically been like test units for the R&D teams who are basically preparing and planning new designs for the data center and cloud. So this is in the range that some of these pipelines. We have a pipeline of 20, 30 clients coming from Intel and they are buying anything from 1 unit to 50 units. So, so far, it's just test units. Next question. Thanks for the update. You mentioned Supermicro as client, I think. Can you elaborate a bit further about how you help Supermicro. So in this specific case with Supermicro, they are delivering a solution, hardware, software for 5G. And they are using our cards with the -- and I think they actually are getting the software stack from a third-party vendor. So they're integrating all this and they're offering this for the 5G market basically as a UPF acceleration solution. Next question, what is the number of design wins you have won so far? So we had -- we're mentioning some design wins we have in the presentation. They're all for our profitable niche business, but very stable business. So we kind of divide between our existing old niche business. And when we talk about design wins, when we guide for 4 to 6 design wins, we're talking about the IPU, more cloud data center business where we are not sitting off-line, where we're sitting in the data stream. And we have so far announced one win. This was with the big new server OEM manufacturer. So it just happens that one of the first design win we got was one of the biggest we could get. But so far, one design win. Next question, what is the technical difference between your current SmartNICs and the ones you are developing in [ collaboration ] with Intel? So that's a good question because I think what has been -- what is really the advantage for us is we have been doing network cards for 20 years. And we basically are doing everything ourselves, hardware, software, the whole thing, and we have like 20 years experience in that. But up until now, we have basically just done this for one use case, which is packet capture, basically where we sit off-line in the data center. And this -- and we're not sitting in the data stream like we are with the IPUs we're developing together with Intel. So basically, what we've been doing for 20 years is very much the same as we're doing for Intel on the IPU. It's just that the use case is completely different. In the capture business, we're sitting off-line. You don't need so many units. But for the SmartNICs we are building for Intel, this is -- there, we are sitting basically in the data stream. And obviously, we're sitting in each server. So that is the main difference that we will be -- the volume are much higher. Then the other difference is that you -- on the IPUs, you have an FPGA, but you also have a CPU, and that makes the device more powerful for many use cases. In most -- in our existing product lines, we just have an FPGA, where we -- it's very good for programmable. But when you add a CPU, like you do in IPU, then you have disadvantage that CPU is really good, particularly for determined task. And then you have the FPGA as well, where you can reprogram the entire device and server according to what kind of traffic you be in, which you want to offload or you want to accelerate, et cetera. So that's the main difference. Next question, why will or can you not name drop the OEM partner for your first major IPU design when recently announced? The reason is simply that this is a big server OEM who's planning a complete upgrade of the cloud and data center going forward. And they want to control the communication around that big expansion they're doing and they don't want us to be the ones announcing this because we're delivering a very important component to this new server and data center. So we expect that will be announced later this year. Next question. How do you expect the volume to ramp? When do you expect to reach the 50,000, 60,000 units? Which year? So what we're doing here is that for each design win we have, we guide on what is the expected volume in peak production years. So for this first one, which is a very big OEM, and this is the next data center cloud solution, which will last for many years. So here, we probably expect the high volume, and this particularly one, we put in the category of more than 25,000 in peak production year. So that is expected in most likely '26. The other -- the remaining 3 to 5 design wins we have guided as well, there, we have basically said that they can be anything from -- in the category of 1,000, 2,000, 5,000, 10,000. And then we have these big whales, which is more than 25,000, which happens to be the first one we announced. So we expect this -- these design wins, we will hopefully sign and announce this year to have peak production volumes basically in 2025 and 2026. For the really big ones, which requires a bit more planning like the ones we have designed, we have announced so far, there is '26. Some of the ones we will announce later this year will be peak production volume in '25 already. Next question, is the 50,000 units expected from one design win you have announced? So in that one, we're basically saying we categorize our wins in 1,000, 2,000, 5,000, 10,000 or 25,000. Basically, there are some few monster deals that's more than 25,000. The ones we have announced is in that category, more than 25,000. Have you signed a contract with the first customer for specific volume? If not, when do you expect the customer to sign contract for firm volume? The way it works is that now we are delivering the hardware, which we have to deliver this year. As soon as -- we're also now in the process of agreeing on who is delivering the software. We will deliver many of the components there. We will have the responsibility to basically make the final product, meaning we need to productize the entire hardware, software stack. So that will be done on the software side. That will be done this year as well, but also next year. Then they will start building pilot in '25. And so I expect that next year, on this specific one, there will most likely be in the 4-digit numbers units. And then you will move into basically the rollout globally, their data centers in '26 and '27. And there, you will see the more than 25,000 units. So peak production will be in '26 and '27 for this monster big whale deal. Next question, that is the same as we just answered. When will you begin to deliver on your first customer for Intel partnership? So basically, we will deliver hardware this year. We will deploy the software also this year and next year. And we will start delivering the first units, full -- completed will be in '25 for this announced big OEM. Next question, does Napatech produce inventory products based on actual orders? Or do you need to hold stock for a [indiscernible] order? So we have manufactured or produced enough units so we can basically give this big pipeline we have, so they can buy enough testing units for building their cloud and the data center, et cetera. So that is in the range of 100 to 500 we are manufacturing already. Are gross margins lower from Intel/Altera partnership sales? Obviously, for this big OEM we have signed, which will be -- which is expected to be in production and sales in the thousands in '25 and then in more than 25,000 per year from '26 and '27. There, we -- it's very hard to say exactly what gross margin. It will most likely be lower than what we have on capture, but we still have an ambition to have very attractive gross margins on these high-volume deals as well. But it's too early to say exactly where we will end up, but the ambition is, of course, to be more than 50% on gross margin. What do you expect in volumes from the first customer in 2025? I would say like this, if this will be a 4-digit thousand number, if it's closer to -- if it's close to 5,000 -- sorry, if it's closer to a 5-digit numbers in '25, then we know that we will be in peak production in '26. So somewhere from 1,000 to 10,000 in '25. The closer we are to 10,000, the more the secure. We are much more than 25,000 units in '26. It all depends on the rollout of this OEM client, which is a bit different. Some remaining design wins, we expect to get this year because it is one of the biggest OEMs out there. You said the first design win with this top OEM was among the biggest you were working with. Can you ballpark -- predicate how much of the 50,000, 60,000 cards will come from one client? The way we're guiding on this, because it's very difficult, is that we basically say that these design wins, we will announce of the 4 to 6 we have guided. There will be in the categories 1,000, 2,000, 5,000, 10,000 and 25,000. So basically, we are predicting that they are more or less around 5,000 units in average per design win. So that's why we came up with the 30,000 to 40,000 units that is approximately like 5,000 in average. Then we got one of the biggest OEM we could get, and that is in the plus 25,000 unit category. So basically, we just added 25,000 units to our guiding. So that's basically how we -- that's the structure we're working with. So that's why it was added another 25,000 to it, approximately. In these last design wins, would gross margin be lower? Or do you expect them to stay in the same range? We expect them -- if this is getting into volume above more than 25,000 units, we don't expect to maintain 70% in gross margin. But we still expect it to be very attractive gross margins, above 50%. But we'll have to come back to that when we start really manufacturing in 4-digit numbers. What type of hardware are you delivering to the first customer already this year? So we are delivering full hardware this year so they can run full software on. So it's basically the real product, full product type, and we call it prototype because obviously, there might be testing where we need to improve it. But it's hardware, fully in line with the full specification from the client here. I think that was all the questions. Do we have anyone for the operator as well? I'm just reading the text board here.
Operator
operator[Operator Instructions] Okay. I confirm we have no further questions on the line at this time. Thank you. With no further questions, this will conclude the Napatech Q1 2024 IMS Conference Call and Webcast. Thank you all for joining. You may now disconnect your lines.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Napatech A/S transcript — plus 251,000+ transcripts from 12,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.
Get the API View API docs →This call discussed
For developers and AI pipelines
Programmatic access to Napatech A/S earnings transcripts and 251,000+ others is available through the
EarningsCalls.dev REST API. Plans from $24.99/month — full transcripts, speaker segments,
full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.