Blaize Holdings, Inc. Common Stock (BZAI) Q4 2024 Earnings Call Transcript

Good day and thank you for standing by. Welcome to the Blaize Fiscal Year 2024 Earnings Conference Call. At this time, all participants are in a listen-only mode. Please be advised that today's conference is being recorded. After the speaker's presentation, there will be a question-and-answer session. [Operator Instructions]. I would now like to hand the conference over to your speaker today, Lana Adair, Investor Relations.

Thank you and good afternoon. I am Lana Adair, Investor Relations for Blaize. Before we begin the prepared remarks, we would like to remind you that earlier today, Blaize issued a press release announcing its full year 2024 results. An investor presentation was published and is available on the investor relations section of Blaize's website. Today's earnings call and press release reflect management's views as of today only and will include statements related to our competitive position, anticipated industry trends, business and strategic priorities, or financial outlook for the first quarter of 2025 and full year fiscal year 2025, all of which constitute forward-looking statements under the federal securities law. Actual results may differ materially from those contained in or implied by these forward-looking statements due to risks and uncertainties associated with the business. For a discussion of the material risks and other important factors that could impact actual results, please refer to the company's SEC filings and today's press release, both of which are found on the Investor Relations website. Any forward-looking statements that we make on this call are based on assumptions as of today, and other than those that may be required by law, we undertake no obligation to update these statements as a result of new information or future events. Information discussed on this call concerning Blaize's industry, competitive position and the markets in which it operates is based on information from independent industry and research organizations, other third-party sources, and management estimates. These estimates are derived from publicly available information released by independent industry analysts and other third-party sources, as well as data from Blaize's internal research. These estimates are based on reasonable assumptions and computations made upon reviewing such data and Blaize's experience in and knowledge of such industry and markets. By definition, assumptions are subject to uncertainty and risk, which could cause results to differ materially from those expressed in estimates. During this call, we will discuss certain non-GAAP financial measures. These non-GAAP financial measures should be considered as a supplement to, and not as a substitute for, measures prepared in accordance with GAAP. For a reconciliation of non-GAAP financial measures discussed during this call to the most directly comparable GAAP measures, please refer to today's press release. Now I'd like to turn the call over to Blaize's chairman of the board, Lane Bess, to say a few opening remarks.

Hi. My name is Lane Bess and I'm the Chairman of Blaize. I've always enjoyed being involved with exciting technology. And I've been fortunate enough to have some great success working with several high growth technology companies that successfully built enormous moats in critical technology areas. I feel Blaize has all the necessary pieces to build a valuable and sustainable company in the enormous and rapidly growing AI edge space. The AI that I saw when I first met Dinakar and the founding team was something that I believe was going to be equally as disruptive as other companies I've been involved with. Particularly what intrigued me about Blaize was not only the advancement they had made in terms of the chip and silicon architecture, but also the software they developed that runs on top of the chip. To me, this was taking AI to the edge in a compelling way. If you explore what we've seen, particularly in the data center industry, these are very heavy LLM processing data center solutions. And Nvidia will probably dominate for a long time in this space. Others, like Cerberus, will probably take pieces of the market. But the area of edge AI computing, this is where we believe we're going to see a lot of innovation. The team at Blaize understands that the market for AI at the edge is expected to be every bit as large, if not larger, than what we see in the data center industry. And the solutions are going to take a unique design, which we believe Blaize can deliver. We believe there are going to be many use cases for Blaize's technology offerings as we go forward, including, but not limited to, smart city and defense systems. As was in the case with Nvidia, their growth began with gaming and later surged with OpenAI and generative AI, expanding their market share and adoption. We believe there is a similar and perhaps even more rapid demand trajectory here due to AI demands and semiconductors. Being in the public markets allows us to capitalize on this momentum and unlock the full potential of our expansion opportunities. And now I'll turn it over to Dinakar, who will get deeper into the technology and Blaize's opportunities as a public company.

Good afternoon, everyone. And thank you for joining us for Blaize's fiscal year 2024 earnings call. I'm Dinakar Munagala, CEO of Blaize. After my remarks, our CFO, Harminder Sehmi, will walk you through our fiscal year 2024 financial results and provide guidance for 2025. Then we will open it up for questions. Before we get into the numbers, I want to highlight several important developments from our fourth quarter and leading up to the closing of our business combination on January 13th. During this period, we finalized the necessary filings to enter the public markets while continuing to grow our commercial pipeline, advance our software roadmap, and expand our team with key hires to support long-term execution. We saw continued growth across our pipeline with several high priority opportunities anticipated to deliver revenue in the second half of 2025. We also secured new designments, advanced testing on core solutions, and received early customer validation across key verticals. During 2024 and continuing into the early part of this year, we formalized partnerships with KAIST, alwaysAI, VSBLTY, and Turbo Federal. These relationships represent meaningful opportunities in both commercial and public sector markets, supporting near-term revenue visibility and laying the groundwork for scale deployment in the future years. As we enter 2025, we have a strong pipeline coverage, aligned execution across our sales and product development teams, and increasing demand in markets such as manufacturing, public safety, healthcare, defense, and smart infrastructure. Our go-to-market strategy remains focused. Our roadmap is progressing on schedule, and engagement with customers and partners continues to deepen. Blaize is entering the year with confidence, operational focus, and increasing demand across the market that is rapidly shifting towards physical AI solutions. Technology moves in waves and AI is now entering its next evolution. Just as Intel and Apple helped shift computing from mainframes to personal devices, AI is now moving from the cloud to the edge, where data is created and decisions need to happen in real time. This is where Blaize excels. We deliver low power, real time intelligence at the edge where decisions happen, complementing cloud and GPU based AI in environments where efficiency and responsiveness matter most. We call it physical AI, AI that works in the real world, not just the data center. We believe Blaize is helping to define the next era of infrastructure for AI. In recent weeks, innovations like DeepSeek and Qwen have reinforced that disruption remains a constant in AI. These advancements are lowering the compute requirements for deployment, enabling AI to run on smaller, more power efficient devices. As models continue to shrink, the reach of AI is expanding into broader markets. Blaize's programmable architecture is well positioned to take advantage of this trend. While cloud AI has seen remarkable adoption, there's still a critical gap in the market, the need for AI at the physical edge. In mission critical settings on factory floors inside hospitals and across defense operations, latency, bandwidth constraints, and reliance on the cloud are simply not acceptable. These environments demand real-time decision-making, on-site processing, and solutions that are both scalable and cost-effective. Looking ahead, we are focused on a serviceable, addressable market that we estimate will grow to $71.3 billion globally by 2028. This figure reflects increasing demand for edge deployed AI across industries where latency, power efficiency, and real-time processing are essential. We're seeing strong demand across industries for AI that operates closer to the point of action. In manufacturing, companies need real-time defect detection and automation right on the production line, not delayed by cloud processing. In healthcare, hospitals are deploying AI-assisted diagnostics on-site to ensure both patient privacy and faster clinical decisions. In defense and security, AI must run locally, where the cloud access is often limited or entirely unavailable. Further, in retail and smart cities, physical AI can enable license plate recognition, logistics automation and surveillance without the latency or cost of sending data to the cloud. The message from our customers is clear. They need AI where they operate, not where the cloud providers dictate. We believe Blaize can enable and address this modern shift today. For years, the AI infrastructure stack has been too rigid, too centralized and too expensive. The market is asking for something different. Across multiple sectors, we hear the same message that customers and partners want more choice, more flexibility, and more tailored solutions. They want AI that fits their needs and not a one size fits all approach. They need technology that integrates into their existing infrastructure without the cost and complexity of a full replacement. And they're looking for alternatives that support AI at the edge, not just in the cloud. Blaize is not here to replace GPUs or the cloud AI. We're here to extend and complement what exists. GPUs are a great fit for training and data center workloads. But in physical environments where power, cost, and responsiveness matter, Blaize offers a more efficient solution. And in many deployments, Blaize and GPUs work side by side. We're collaborating with manufacturers, system integrators, ISVs, and AI developers to build a more modular, adaptable AI ecosystem. The market is clearly signaling the need for more options and Blaize is well positioned to meet that demand. Our research and development is focused on three priorities, low latency, power efficiency, and real-world deployment. Our Graph Streaming Processor enables real-time inference at the edge with significantly lower power than traditional GPUs without relying on constant cloud access. We believe we offer strong ROI by developing more AI per watt and per dollar, while reducing energy use and cloud costs. Our end-to-end platform is built for practical deployment, not just lab benchmarks. Blaize complements existing infrastructure by bringing efficient, scalable AI to the edge. Blaize is a pure-play AI company, purpose-built for the intersection of AI, edge computing, and real-world infrastructure. As AI demand moves beyond the cloud, we're leading the shift to physical AI where intelligence is deployed onsite and where latency, power, and cost matter. Our platform delivers scalable, power-efficient performance at the edge, complementing cloud and GPU solutions by extending AI into environments where real-time local processing is essential. This is where Blaize stands apart and where we're building long-term differentiation. AI is moving beyond the cloud, and Blaize is prepared to help lead the transition by bringing intelligence closer to where data is generated and decisions are made. We're focused on a clear opportunity, delivering efficient, flexible AI at the edge to meet real world customer needs. Our platform stands out by providing performance where it matters most, close to the data in real time and under real world constraints. We're executing against a roadmap, scaling with discipline and making steady progress. Thank you to our customers, partners and stakeholders for your continued support. We are early in our journey and I'm hopeful that the narrative that I have shared with you helps you understand the opportunity in this market and the capabilities Blaize can offer in the edge AI market. We look forward to sharing our growth as a public company and momentum in the quarters ahead. Now I'd like to hand over the call to our CFO, Harminder Sehmi.

Thank you, Dinakar. And good afternoon, everyone. My name is Harminder Sehmi, and I'm the Chief Financial Officer of Blaize. Before we begin, I'd like to reiterate that today's discussion will include forward-looking statements. Please refer to our earnings release and SEC filings for additional information. Our business combination with BurTech Acquisition Corp. closed on January 13 of this year. So my focus in this earnings call will be on our fiscal year 2024 results ending December 31. I will also provide guidance on the first quarter of 2025 and full fiscal year 2025. 2024 was a pivotal year for Blaize as we developed partnerships with selected hardware manufacturers and independent software vendors and established a revenue pipeline which is expected to start converting at scale in the second half of this fiscal year. Let's start with revenue for fiscal 2024. Revenue was $1.6 million compared to $3.9 million in the prior year. In both years, the revenues primarily reflected the recognition of strategic consulting fees. These were received from a major European automotive OEM as part of a multi-year and multi-vendor program aimed at developing an advanced L4 autonomous driving platform for the OEM. Phase 1 was largely completed during 2024. Revenues from automotive are several years away, given the design and manufacturing lead times associated with autonomous vehicles. Adjusted gross margin for 2023 and 2024 includes an allocation of direct labor costs in connection with the strategic consulting program and adjustments for inventory cost realignment. I will discuss our future adjusted gross margin trends shortly. Now to our operating expenses. Selling, general and administrative expenses for fiscal 2024 excluding depreciation and amortization were $22.4 million, an increase of $5.1 million over fiscal 2023. This change reflects the investment in our preparation to transition into a public company and to scale our go-to-market capability globally. This came on top of exercising firm controls on holding down our expenditures in 2023. Our plans to grow costs in the future will be very closely tied to delivering revenue growth. Research and development costs grew by just under $7 million to $25.1 million in 2024 as we invested in technical resources, software tools, and third-party IP for our next generation chip. We remain committed to driving long-term value through innovation while ensuring an efficient allocation of capital. Our net loss for 2024 was $61.2 million, down 30 % from the net loss of $87.6 million in 2023. The resultant adjusted EBITDA loss, excluding stock-based compensation and other non-cash items for fiscal 2024, was $43.3 million compared to $30.3 million for fiscal 2023. A brief comment on our cash position. The company held $50.2 million on the balance sheet at 31st December 2024. And following the conclusion of the business combination in January this year, when additional capital was raised through a private investment in public equity, we believe that Blaize is on a path to execute on its operating plan over the next 12 months. In summary, 2024 was a year of transition for the company with capital raised before the transaction closed. We developed a strong revenue pipeline, which is continually growing and invested in key partnerships in our chosen markets. We will continue to evaluate capital allocation strategies that align with our long-term growth objectives of focusing on the edge AI opportunity. Let's now turn to financial guidance. For fiscal 2025, we are reaffirming our revenue guidance range provided in February this year. We expect revenue to be in the range of $19 million to $50 million. The adjusted EBITDA loss is projected in the range of $70 million to $75 million. The stock-based compensation charge is expected to be around $20 million. As noted at the time we issued our 2025 guidance, we expect to see customer solutions being deployed in the second half of the year through a combination of hardware systems, software licenses, and professional services. We are still assessing the impact of recent tariffs on our adjusted gross margin percentage and plan to provide additional color during the year. Our chips are manufactured on US soil and we maintain very close relationships with our contract manufacturer to tailor supply to demand. Now I'd like to share our guidance for the quarter ended March 31st, 2025. We expect revenue to be around $900,000, mainly from product shipments. We anticipate an adjusted EBITDA loss in the range of $15 million to $16 million. And finally, our stock-based compensation charge is estimated at $7.7 million. As our strategic partnerships continue to grow with more announcements planned this year, we expect that the revenue pipeline will reflect the large underlying customer pipelines that these partnerships are expected to bring to place. Solutions based on the current generation of silicon are expected to support revenue growth over the coming three plus years. These will coexist with products and solutions based on our next generation chip that opens additional revenue opportunities in edge AI for Blaize. In closing, our merger enabled us to be funded and we have an accelerating revenue pipeline across a variety of customers and partners. Blaize has been actively preparing to operate efficiently as a public company. We feel confident in our technology, long-term strategy, and adding shareholder value by delivering effective solutions into the large and rapidly growing edge AI market. With that, I'll turn it back over to the operator for questions and answers.

[Operator Instructions]. Our first question comes from Alex Platt with D.A. Davidson.

I just have one question, maybe a more high level one, but sort of beyond hardware improvements, I'm sort of interested to know how you see the improvement of small models, I guess, parameter wise, contributing to long term adoption of Edge AI solutions. And then maybe within that question, we know some of the use cases in verticals you can serve now or sort of in the near future. But are there any other areas that really get unlocked for you with drastically improved levels of intelligence in models?

It's regarding the small language models specifically. And DeepSeek has shown that models will continue to shrink and there are several efforts in the space. And small language models, what essentially it does is, it truly democratizes AI and the work will be shifted more and more towards the edge. So we're talking about not 1x, 2x, 1000x kind of shrinking of AI models. In fact, the initial shipments that have been announced are true multimodal and in the space of these small language models. So this is actually quite a good thing for edge computing and particularly Blaize architecture which is programmable to be able to address this. In addition, your question was also about other areas. Definitely this is something that will, although we're focused on initially computer vision, public safety, smart city, that area, this is definitely going to impact all the major edge computing use cases. I'll let Harminder add any more.

No, I think that's fine. You've covered it.

[Operator Instructions]. Our next question comes from Kevin Cassidy with Rosenblatt Securities.

Just as you're looking at this huge market that you're looking at, how do you filter out which programs or which partners you choose just to make sure you're getting the best ROI?

Our go-to-market has two ways we go to market. One is high value customers, high contract value customers that we engage directly. But also we engage with the ecosystem. The ecosystem comprises ISP partners as well as system integrators targeting end customers. So the idea is that the high value customers we go direct and as we engage with the ecosystem and land a particular use case with a particular end customer, their pipeline, the ISV's pipeline also becomes ours. So there's a multiplier effect there. So by going direct, we also focus on a land and expand within the individual customer. And with the ecosystem as we engage, we have a multiplier effect also that grows the pipeline. So that's the combination.

I'll amplify a couple of things. We are very focused on specific areas just from the allocation of resources. So one thing to note is that the AI algorithms that are behind each of these use cases are largely similar. So when we've selected our ecosystem partners, what we're doing is making sure that the beachhead customer, i.e. the first customer that we're going to go deploy the solution with, we test, we pilot, and we make sure that the solutions are relevant for them. And as Dinakar mentioned, the land and expand where the second customer, the 10th customer, the 100th customer in that vertical, the path to revenue is much, much faster with those customers. As our roadmap expands, we will open up other verticals, but for now, our focus, our pipeline is based on the focused smart cities, some defense type of customers.

Just some of your new system partners, have they tested your chips using DeepSeek?

Each customer, of course, has different flavors of language models as well as computer vision algorithms. DeepSeek so happens to be a certain flavor of a language model. But as you know, there's Phi, there's Gemma, there's others which also fall in a similar category of multimodal AI. So the key is multimodal AI. We happen to be very unique in that our processor architecture is completely programmable. And therefore, as the world is making this huge transition to incorporate true multimodal AI, we're here and ready. The combination of our programmable processor and our software stack is coming in very handy to our partners and then customers. So different flavors, but the commonality is how we actually uniquely activate it.

Just one other question, more on the balance sheet, the inventory you're holding, is that products, those in die form, or is that something else? I guess what is being held in inventory, about $8 million?

Yeah, that's all finished good. We have a contract manufacturer. There'll be certain chips in there too, just the SOCs, but all PCI cards, various form factors, that's what we have.

Our next question comes from Richard Shannon with Craig-Hallum Capital Group.

Maybe I'll ask the first one on the pipeline since [Technical Difficulty]. If you could quantify the pipeline progress since the last time you published in your new filings very early this year, [Technical Difficulty] hundreds of millions of dollars, anybody that can characterize that.

Richard, your line is breaking up. It's very difficult to follow you.

I'll turn the call back over to Dinakar Munagala for any closing remarks.

Okay. Thank you all for joining and we look forward to updating you as we make progress for the business in the near term and the coming quarters.

Thank you. This concludes the conference. Thank you for your participation. You may now disconnect.