Earnings Call Transcript - AIP Q3 2025

Operator, Operator

Good afternoon, everyone, and welcome to the Arteris Third Quarter 2025 Earnings Call. Please note, this call is being recorded and simultaneously broadcast. All materials contained in the webcast is sole property and copyright of Arteris, Inc., with all rights reserved. For opening remarks and introductions, I will now turn the call over to Erica Mannion of Sapphire Investor Relations. Please go ahead.

Erica Mannion, Investor Relations

Thank you, and good afternoon. With me today from Arteris are Charlie Janac, Chief Executive Officer; and Nick Hawkins, Chief Financial Officer. Charlie will begin with a brief review of the business results for the third quarter ended September 30, 2025. Nick will review the financial results for the third quarter followed by the company's outlook for the fourth quarter and the full year of 2025. We will then open the call for questions. Before we begin, I'd like to remind you that management will make statements during this call that are forward-looking statements within the meaning of federal securities laws. These statements are based on management's current expectations and assumptions and involve material risks and uncertainties that could cause actual results or events to materially differ from those anticipated, and you should not place undue reliance on forward-looking statements. Additional information regarding these risks, uncertainties and factors that could cause results to differ appear in the press release that Arteris issued today and in the documents and reports filed by Arteris from time to time with the Securities and Exchange Commission. Please note, during this call, we will cite certain non-GAAP measures, including, among others, non-GAAP net loss, non-GAAP net loss per share and free cash flow, which are not measures prepared in accordance with U.S. GAAP. The non-GAAP measures are presented as we believe they provide investors with the means of evaluating and understanding how the company's management evaluates the company's operating performance. These non-GAAP measures should not be considered in isolation from as substitutes for or superior to financial measures prepared in accordance with U.S. GAAP. A reconciliation of these non-GAAP measures to the nearest GAAP measure can be found in the press release for the quarter ended September 30, 2025. In addition, for a definition of the key performance indicators used in this presentation such as annual contract value, confirmed design starts and remaining performance obligations, please see the press release for the quarter ended September 30, 2025. These key performance indicators are presented for supplemental informational purposes only should not be considered as a substitute for financial information presented in accordance with GAAP and may differ from similarly titled metrics or measures used by other companies, security analysts or investors. Listeners who do not have a copy of the press release for the quarter ended September 30, 2025, may obtain one by visiting the Investor Relations section of the company's website at ir.arteris.com. In addition, management will be referring to the third quarter 2025 earnings presentation, which can be found in the Investor Relations section of the company's website under the Events and Presentations tab. Now I will turn the call over to CEO, Charlie Janac.

Charlie Janac, CEO

Thank you, Erica, and thanks to everyone for joining us on our call today. In the third quarter of 2025, we achieved yet another record annual contract value plus royalties of $74.9 million, resulting in 24% year-over-year growth. We saw increased product adoption in chiplets and SoCs across multiple vertical markets. AI applications accounted for over half of our licensing dollars in the third quarter reflecting the growing adoption of Arteris system IP technology from data centers to the smart edge. We continue to see growing adoption of our product portfolio by top technology companies. An example of this is Altera, which selected Arteris technology portfolio to streamline design workflows, optimize data movement and enable intelligent computing across data center, communications, vision, industrial applications, robotics, aerospace and defense applications. This includes our network on-chip IP products, including Ncore and FlexGen and the management platform for IP block integration and hardware, software integration automation, which Altera plans to use in designing their next generation of FPGA and SoC FPGA solutions. Speaking of FlexGen, last quarter, we announced that AMD licensed the Smart NoC IP to provide high-performance data transport in AI chiplets across AMD's broad portfolio from data centers to edge devices. I'm happy to note that in the third quarter, AMD has ordered additional incremental licenses. In addition to the Altera and AMD relationships, we added 4 other new FlexGen customers in the third quarter. Within the automotive sector, FlexGen was deployed by Dream Chip, a custom SoC design house for high-end automotive semiconductor design. Additionally, a leading automotive OEM adopted FlexGen for next-generation EVs. Within the industrial sector, NanoXplore, a provider of radiation-hardened silicon technology serving the aerospace, defense, avionics and industrial markets, licensed FlexGen's Smart NoC IP to address the demanding mission-critical computing requirements in space while supporting their product performance, team productivity, device reliability and meeting the underlying area and cost targets. This represents another example of our products being used not only for applications on Earth, but increasingly in terrestrial orbit, where performance, safety, reliability and security are essential. These examples illustrate the broad applicability of our new FlexGen Smart NoC IP, helping design teams deliver on expanded needs of chiplets and SoCs. Additionally, we expect demand to scale with rising design complexity and the move to advanced foundry nodes, particularly 5-nanometer, 3-nanometer, 2-nanometer and as we head into the Angstrom era of silicon. As the semiconductor industry accelerates efforts to increase performance and efficiency, especially driven by AI workloads and data centers and the edge, we are continuing to see a growing shift from traditional monolithic chips toward chiplets for multi-die SoC architectures, particularly for AI infrastructure and data center applications. One of the key chiplets is the IO Hub chiplet, which controls data movement across heterogeneous multi-die SoCs. 2V Systems licensed our Ncore and FlexNoC interconnect IPs to develop just such an IO Hub chiplet where Arteris technology serves to control multi-die data traffic meeting the high bandwidth, low latency energy efficiency and total cost of ownership objectives while meeting the needs of enterprise computing in data centers and cloud infrastructure. In the quarter, we also saw increased adoption of chiplets for high-end automotive applications, including our recently expanded multi-die solution. For example, one of our advanced automotive semiconductor customers shifted from a single chip to multi-die SoC architectures for their next-generation ADAS design, leveraging Ncore FlexNoC IPs for underlying data movement. Aside from various automotive semiconductor companies, we also saw expanded adoption of Arteris technology by automotive OEMs. Two of the top 5 EV automotive OEM companies expanding their use of silicon proving Arteris technology with functional safety for their next generation of vehicles, which increasingly include a wider array of advanced electronic functionality. Given the accelerating demand for increasingly advanced chiplets and chips from the AI surge in the high end to the growing needs of advanced microcontrollers, the needs for more specialized computing is becoming increasingly evident. This trend drives a broad range of specialized processors or XPUs, for a growing number of applications by providers who increasingly rely on Arteris technology for their underlying connectivity and data movement. With our growing ecosystem, we recently announced an expanded collaboration with Alibaba Damo Academy, enabling better integration and optimized performance between the risk 5 CPU cores and our data movement system IPs. This collaboration is intended to further enable mutual customers to more efficiently design AI server communications and automotive chips. Such ecosystem collaborations help enhance support for end customers, enabling them to accelerate their pace of innovation, with recent examples being Axelera AI, a provider of purpose-built hardware acceleration technology for AI inference. They recently expanded the use of Arteris to help accelerate computer vision for edge devices using our technology to help achieve high bandwidth, low latency and scalability required to optimize their next-generation inference products. The need for ecosystem collaboration is also evident as industry standards continue to evolve. In particular, AI data center infrastructure needs are rapidly evolving, driving demand for purpose-built solutions that can better support rapidly expanding AI workloads. To better meet the associated demand from customers, Arteris joined the Ultra Accelerator Link Consortium, or UALink. The goal of this organization is to establish an optimized, scale-up ecosystem across multiple AI accelerators with Arteris NoC IP serving as data movement transport in chiplets and SoCs. We joined with other companies in the consortium, such as AMD, Astera Labs, AWS, Cisco, Google, HP Enterprise, Intel, Meta and Microsoft, all of whom deal with high-end computing and some of whom are requesting the related support in our products. Lastly, I'm proud that Arteris continuous innovation was recognized with yet another award this time as the winner of the most innovative technology company of the year by the 22nd Annual International Business Awards, while also being recognized for new FlexGen Smart NoC IP and Magillem registers integration automation software product, both announced earlier this year. We believe the scale and scope of our opportunity remain robust, supported by our current products, and strong pipeline of new data movement system IP technologies as well as growing relationships with the largest and most advanced electronics companies in the world in collaboration with a broader ecosystem. Our customers continue to innovate in exciting high-growth areas across multiple applications from AI data centers to the edge, autonomous driving, advanced communications, consumer and industrial use cases. Many of these customers are increasingly turning to our products and solutions to support their innovative designs. With that, I'll turn it over to Nick to discuss our financial results in more detail.

Nicholas Hawkins, CFO

Thank you, Charlie, and good afternoon, everyone. As I review our third quarter results today, please note that I'll be referring to GAAP as well as non-GAAP metrics, reconciliation of GAAP to non-GAAP financials is included in today's earnings release, which is available on our website. Also, as a reminder, I will be referring to the 3Q 2025 earnings presentation which can be found in the Investor Relations section of the company's website under the Events and Presentations tab. We had a strong third quarter meeting or beating our guidance on all financial measures. Turning to Slide 5 of the presentation. Total revenue for the third quarter was $17.4 million, up 5% sequentially and 18% year-over-year and above the top end of our guidance range. Notably, trailing 12-month variable royalties were 36% higher year-over-year. At the end of our third quarter, annual contract value plus royalties was $74.9 million, up 24% year-over-year, above the top end of our guidance range and at a new record high. Remaining performance obligations, which is our contracted future revenue at the end of the third quarter was $104.7 million, representing a 34% year-over-year increase, a new high and exceeding the $100 million milestone for the first time. Non-GAAP gross profit for the quarter was $15.9 million, representing a gross margin of 91%. GAAP gross profit for the quarter was $15.6 million, representing a gross margin of 90%. Now turning to Slide 6. Non-GAAP operating expense for the quarter was $19.5 million. We continue to reinvest a portion of our top line growth into technology innovations, solution support and our global sales team. Total GAAP operating expense for the third quarter was $24.4 million. We believe that our ongoing investments will help accelerate our top line growth in the coming years. At the same time, we are delivering operating leverage by controlling G&A spending, which has now remained broadly flat on a non-GAAP basis for over 3 years. This has resulted in a 15% improvement of non-GAAP operating expense as a percentage of revenue for the year-to-date compared to the same period in 2023. Non-GAAP operating loss in the quarter was $3.5 million, in line with our guidance. GAAP operating loss for the third quarter was $8.7 million compared to a loss of $7.9 million in the prior year period. Non-GAAP net loss for the quarter was $3.8 million or diluted net loss per share of $0.09 based on approximately 42.7 million weighted average diluted shares outstanding. GAAP net loss in the quarter was $9 million or diluted net loss per share of $0.21. Moving to Slide 7 and turning to the balance sheet and cash flow. We ended the quarter with $56.2 million in cash, cash equivalents and investments, and we have no financial debt. Free cash flow, which includes capital expenditure, was positive $2.5 million for the third quarter, above the midpoint of our guidance range. I would now like to turn to our outlook for the fourth quarter and the full year 2025 and refer now to Slide 8. For the fourth quarter 2025, we expect ACV plus royalties of $74 million to $78 million, revenue of $18.4 million to $18.8 million with non-GAAP operating loss of $2.3 million to $3.3 million and non-GAAP free cash flow of $0.2 million to $3.2 million. For the full year, 2025, our guidance is as follows: ACV plus royalties to exit 2025 at $74 million to $78 million, an increase of $1 million compared to our prior guidance. Revenue of $68.8 million to $69.2 million, also an increase of $1 million compared to our prior guidance. Non-GAAP operating loss of between $12.5 million to $13.5 million and non-GAAP free cash flow of $2.5 million to $5.5 million. We remain encouraged by our strong deal execution witnessed by the 34% year-over-year growth in RPO at the end of the third quarter. We are seeing promising signs of accelerated interest by some major customers to increase their outsourcing of system IP products to Arteris, which we believe will help accelerate growth in our license and royalty revenue, ACV plus royalties and positive free cash flow.

Operator, Operator

Your first question is from Kevin Garrigan from Jefferies.

Kevin Garrigan, Analyst

Charlie and Nick, congrats on the results in the Altera announcement. Can you just talk a little bit more about Altera? Are they fully away from using internal interconnect teams? Or is there still more opportunities for you guys to expand there?

Charlie Janac, CEO

I think there's more opportunities. Basically, the application is for FPGAs and FPGA SoCs. So Altera is using their own interconnect in the FPGA matrix, and then we are used essentially in the SoC part. But Altera business is going to continue to evolve and grow and we believe that there's future opportunities, but this is a major milestone because Altera, as they spun out of Intel chose to go with Arteris for their primary system IP requirements. But yes, there is more potential going down the road.

Kevin Garrigan, Analyst

Okay. Perfect. And then since the initial discussions with AMD and the initial order announcement, it seems like it took about 1 quarter, maybe a little bit longer for them to expand the use of your products. So what were they kind of most impressed with that led to increasing usage in such a short time frame?

Charlie Janac, CEO

Yes, AMD is a large company. The deal in the second quarter involved their central engineering group, and the deal in the third quarter was for another group. AMD has many teams for us to collaborate with, and there are additional opportunities available. We are very enthusiastic about helping them speed up their chip deliveries.

Kevin Garrigan, Analyst

Got it. Okay. And just one more question if I can. Can you elaborate on the importance of reliability and safety when it comes to interconnects, especially in end markets like space, as you mentioned? I believe you have done a very good job in this area, but do you see this focus as a real competitive advantage for you?

Charlie Janac, CEO

Absolutely. I mean, basically, all the important data goes through our network on chips. And basically, if that has problems or doesn't work, the chip doesn't work. So customers are very risk averse in choosing system IP solutions because any problems there can cause major delays in tape-outs and field problems. So we're being recognized as very much a silicon-proven company. I think now our installed base has shipped something like 3.9 billion SoCs and they all work. And probably some of the stuff used daily probably has our Arteris Interconnect in it. So yes, we are very much focused on reliability. We're very much focused on quality because if the system IP doesn't work, the chip doesn't work.

Operator, Operator

The next question is from Kevin Cassidy from Rosenblatt Securities.

Kevin Cassidy, Analyst

Congratulations on the great momentum. Just on the UALink consortium, what kind of timing can we expect for licenses to come out of that consortium and some of the players there?

Charlie Janac, CEO

Well, some of the players are already customers. But basically, the objective of the UALink consortium is to essentially scale up data center solutions. And so we're basically developing technology to support that, and we're already involved in some of those designs, but we're basically following that consortium's protocol in order to support the data center scale up efforts that they are pioneered by the companies that we mentioned.

Kevin Cassidy, Analyst

Okay. Great. And with the penetration you're getting within AMD and combining it with the Altera announcement, is there opportunities for Xilinx? Or is that already included in your AMD discussion?

Charlie Janac, CEO

Well, Xilinx is an important part of AMD. And in fact, Xilinx was the first customer that was involved with us prior to the AMD acquisition. So Xilinx has been a long-time user of Arteris.

Operator, Operator

Your next question is from Gus Richard from Northland.

Gus Richard, Analyst

Real quick, you've had a number of design wins for a while. And just wondering the royalty relative to most mature IT companies is relatively low. Just wondering when do you expect that to start to accelerate?

Charlie Janac, CEO

Nick, do you want to take that one?

Nicholas Hawkins, CFO

Sure. It's a great question. As we've discussed before, an increasing number of customer design starts is a strong indicator of future royalty growth. Typically, there is a lag of about 3 to 6 years between the beginning of a design and when it reaches mass production and scale, and it might take a couple more years to achieve full scale after mass production starts. There is definitely a strong link between these two factors. We are already seeing this trend and the onset of the increase in royalties. In our Q3 Investor Day presentation, there will be new information on royalties. The growth of variable royalties is particularly impressive; in fact, year-over-year growth in variable royalties over the last 12 months, ending in September, was up 36% compared to the previous 12 months, which aligns with our statements about royalties growing at about double the rate of licenses. An interesting point in the investor deck is when you look back to 2020, which is significant because that’s when HiSilicon dominated our royalties, which have now dropped to zero. We now have a higher rate of variable royalties than we did during the HiSilicon days, and instead of relying heavily on one customer for 90% of our variable royalties, we now have five customers that together generate a greater royalty stream than HiSilicon did. This diversification with more major clients and an additional 50 smaller players is leading to positive growth. So we are starting to see this trend, and I believe that the acceleration will continue over the next couple of years, with an even faster rate of growth expected by 2028.

Gus Richard, Analyst

Got it. That was super helpful. And then, Charlie, for you, you guys talk about the top tech companies that you've penetrated, I was wondering if that's just for everybody to find what those companies are and then how many you've, at this point, penetrated? And then specifically in the AI ASIC crowd, are you starting to penetrate those, both U.S. and Taiwan?

Charlie Janac, CEO

We define the large companies as the top 20 semiconductor companies and another 20 of the largest system electronics companies. We jokingly refer to this as the Arteris index. More than 50% of these companies are our customers, though not all are major clients. There is still significant potential for expanding our business. With the announcements from AMD and Altera, along with a few others we couldn't disclose, we've made good progress among the top 40 technology companies. However, there is much more to achieve. The market is about $1 billion to $1.2 billion, and we are at approximately $68 million this year, indicating that we have a long way to go.

Gus Richard, Analyst

Okay. Got it. And then the Lord Baltimore questions. When I go through cash flow on balance sheet, blah, blah, and it looks like bookings were in the ZIP code of $32 million in the quarter, book-to-bill about $1.8 billion. So Nick, am I in the right ZIP code?

Nicholas Hawkins, CFO

I prefer not to discuss bookings as it could lead to future disclosure issues. Bookings can be quite variable due to our large customer base, which might create misleading expectations if we start revealing those numbers. Therefore, I’ll leave it to you to calculate bookings on your own, Gus.

Operator, Operator

Your next question is from Joshua Buchalter from TD Cowen.

Joshua Buchalter, Analyst

Charlie, I found your remarks in the prepared statement about gaining traction from AI applications, particularly in the data center, to be intriguing. There's been significant development in the AI field recently. Could you clarify how much of your future opportunities you anticipate will come from the data center compared to edge devices and embedded edge devices, which have traditionally been your main focus?

Charlie Janac, CEO

Yes. Our thesis is that, over time, nearly all electronic endpoints or edge devices will connect to the data center. For each endpoint or edge device, there is a corresponding ratio of blades in the data center. As everything becomes connected, the number of chips in these data centers will increase significantly. We are responding to customer demand, and there is considerable attention on AI workloads within the data center, along with numerous project initiations. NVIDIA remains a major player and will continue to be important, but some system houses are also developing their own chips for specific data acceleration related to their AI workloads. We see this as a significant opportunity and are collaborating with those customers, increasingly adapting our engineering efforts to meet the needs of data center companies and hyperscalers that manage high-end AI workloads. Over time, I believe the data center segment will account for around 25% to 30%, possibly 35% of our business. Currently, AI makes up about 50% of the design starts we are involved with, resulting in a surge of design activity. However, I expect the long-term contribution from this segment to stabilize around 35%.

Joshua Buchalter, Analyst

Maybe, Nick, could you share any insights on the strong progress you're making with FlexGen, which appears to be associated with higher average selling prices for both royalties and licensing? When can we expect this to significantly impact the model?

Nicholas Hawkins, CFO

FlexGen is beneficial for both average selling price and licensing, and it's also advantageous for royalties since it outperforms the more basic FlexNoC 5 that lacks automation. For instance, we have begun several FlexGen cycles and deals since mid-year, with four additional ones in the third quarter alongside Altera and AMD. The impact varies with use cases, and currently, most applications are in the server and FPGA environments, which don't produce high volumes. However, we anticipate greater penetration of FlexGen into higher volume sectors. Notably, the automotive market contributes about half of our total royalties, and if FlexGen is used in automotive design now, we might not see royalties from it until around 2030 to 2031. Therefore, we expect significant future royalties from this.

Operator, Operator

There are no further questions at this time. Mr. Janac, please proceed with closing remarks.

Charlie Janac, CEO

Thank you all for your interest in Arteris. We're very excited about the current quarter and look forward to meeting with you in the upcoming non-deal road shows and investor conferences in the following quarters to update you on our business progress. Thank you very much.

Operator, Operator

Ladies and gentlemen, this concludes your conference call for today. We thank you for participating and ask that you please disconnect your lines.