SoundHound at Cantor Conference: Voice AI Leadership and Growth

On Tuesday, 11 March 2025, SoundHound AI (NASDAQ: SOUN) presented at the 2025 Cantor Fitzgerald Global Technology Conference. The company highlighted its leadership in conversational and voice AI, emphasizing strategic growth and competitive positioning. While the outlook remains positive, challenges in competition and market expansion were also discussed.

Key Takeaways

• SoundHound is a leader in conversational AI, with a strong focus on proprietary technology.
• The company operates a three-pillar revenue model: royalty licensing, subscriptions, and voice commerce.
• SoundHound aims to achieve adjusted EBITDA positivity by year-end, supported by a strong cash position.
• Strategic acquisitions, such as Amelia, are expanding SoundHound’s enterprise capabilities.
• The company is gaining traction in automotive, IoT, and enterprise sectors, including healthcare and financial services.

Financial Results

• Voice Enabling Products: The royalty licensing business, which was 85% of the company until 2023, targets a $900 million to $1 billion annual recurring revenue opportunity. This is based on 90 million global light vehicles produced annually, with an average royalty of $10 per vehicle.
• Subscription Services: Comprising two-thirds of the business in 2024, subscriptions include monthly fees for restaurant applications, varying by customer and location.
• Voice Commerce: Although not included in the 2025 guidance, voice commerce has significant revenue potential, connecting voice-enabled services with products.
• Balance Sheet: SoundHound holds $200 million in cash and has no debt, with expectations to be adjusted EBITDA positive by year-end.
• Future Margins: The company anticipates 70% gross margins and 30% EBIT margins, while aiming for breakeven as it reinvests for growth.

Operational Updates

• Automotive Sector: SoundHound is displacing competitors like Nuance/Serence, with over 20 automotive brands as customers and integration with OpenAI in Stellantis vehicles.
• IoT Sector: The company sees a vast opportunity with 75 billion IoT devices projected, leveraging low-cost microphone technology.
• Services Sector (Restaurants): SoundHound’s AI surpasses human order accuracy, working with 7 of the top 20 quick-service restaurants.
• Enterprise Sector (Healthcare, Financial Services): The Amelia acquisition enhances SoundHound’s enterprise stack, replacing third-party agents with in-house solutions.

Future Outlook

• Voice Commerce: While not factored into the 2025 guidance, voice commerce is poised for growth, with a flywheel effect benefiting car manufacturers and restaurants.
• Investments: The company focuses on reinvesting in opportunities with high returns.
• Industry Expansion: SoundHound plans to enter new sectors, including large electric utilities.

Q&A Highlights

• Competition: SoundHound’s innovation and flexibility have helped it outpace competitors like Nuance in the automotive sector.
• Balance Sheet and EBITDA: With a robust cash position, the company expects to be EBITDA positive by year-end, targeting substantial future margins.
• Electric Utility Market: SoundHound aims to be a first mover in this sector, capturing market share aggressively.
• SEC Filing Delay: The recent filing aligns with previous earnings announcements, with no material differences.

For a detailed understanding, readers are encouraged to refer to the full transcript below.

Full transcript - 2025 Cantor Fitzgerald Global Technology Conference:

Tom, AI at Kantor, Kantor: We’re an AI here at Kantor. We’re we’re very happy to have Nitesh here, the CFO of SoundHound AI. I would make a bet that a good chunk of you here in the room are absolutely familiar with SoundHound. You might know a couple of things they do conversationally. And but maybe I’ll just give maybe a one minute intro for you and attention to that would help us kind of level set before we jump into all the questions about what SoundHound AI is doing for the market.

Nitesh, CFO, SoundHound AI: Sure. Well, thanks for having me, Tom. Glad to be here. So SoundHound is a conversational AI and voice AI leader. We’ve been a pioneer in this space for over twenty years.

Our founders built this company with the vision of voice enabling the world with conversational intelligence, did the PhDs at Stanford, pre iPhone days, and speech recognition and machine learning, and have built a proprietary stack. So we’re one of only a handful of companies in the world that own all the pieces parts of the voice AI tech stack from a software perspective. And we think that gives us competitive differentiation because we can innovate quickly, we can modernize, and we’ve been growing rapidly. We announced our earnings just 10 ago. And we’re really trying to disrupt the world and bring a new horizon of how humans interact with technology.

And I know we’ll talk a little bit further, but every fifteen years, twenty years, there’s a new major inflection in technology. We think we’re in the middle of them. We think this is the Gen AI LLM era. And within that, we think this is gonna change how humans interact with technology rather than just a keyboard input or a touch type swipe on your phone. Humans will be able to do a lot more, interacting with technology through natural conversations and just conversing how we’re doing now, how humans have done it for their history, tens of thousands of years.

And within that world, we think voice AI is the killer app, and that’s what we’ve built over time. So we have many applications across products, automotive, in the services space with restaurants and financial services and health care, and we’re building an ecosystem of monetization, voice commerce to connect voice enabled product with voice enabled services.

Tom, AI at Kantor, Kantor: That’s a great overview for a relatively complex company in the sense that you do cover a lot of areas. And I like the analogy that you talk about with conversational search, even comparing your company sometimes to Google, which I think is interesting. And you mentioned competition. We get a lot of questions from investors around how crowded is this market, who do you compete with. And again, there’s probably a number of different competitors because you do have this diverse model that we’ll get into.

But maybe kind of help us understand who you’re competing with and trying to win against.

Nitesh, CFO, SoundHound AI: Sure. Yes. Depending on the industry application and sort of pillar, we see different competitors. So first, like I said, our differentiation is generally in tech. When we go head to head with competition, we generally win on the basis of better technology.

We also have different go to market motions and different ways we partner. So I’ll kind of unpack a couple of these. First, where we got our foray, we launched unveiled after ten years in stealth, we lost our we launched our voice AI engine in about 2015, ’20 ’16. And the platform first got a lot of attention from the automotive makers. Naturally, voice makes sense while driving.

It’s a safety issue. You want to keep your hands on the wheel. And there was one major player. It was Nuance for many decades who was a leader in the voice in the AS speech recognition space. They spun off a public company, Serence.

And in automotive, that’s who we’ve been stealing share from, frankly, over the last several years. They came in with edge solutions, meaning they did it in car without Internet connectivity, so you could turn on the turn up the volume, turn on the AC, open the windows

Tom, AI at Kantor, Kantor: type of thing. We got

Nitesh, CFO, SoundHound AI: traction with cloud capability. So you can connect to maps and navigation and connect to what happened in the game last night or what’s going on with the stock. And and we got entry and ultimately have been displacing them. So we now have over 20 automotive brands in our part of our work that are customers of ours. And we’re continuing to scale across new players.

We’ve gained a lot of traction with EV companies, which tend to operate at a greater speed, more software centric, and so we’ve been able to gain a lot of traction there. So a lot of displacement with Serence. There’s also the big tech that play there. Our differentiation, number one, where we go head to head, we have customer benchmarks that we do head to head better technology and better performance from sentence accuracy, how does the voice AI operate on them. Yeah.

And we’ve the other difference is we do it in brand centricity. So we’ll do white labeled. You can say, hey, Hyundai. You don’t have to say, hey, Alexa. And and we have a broader suite of products.

So we again do edge, cloud, hybrid engines, and so that’s where we’ve been playing. So that’s the automotive space. When you get into IoT, it’s really much more nascent. There aren’t many players, and and so we, it’s a lot more greenfield. When you get to services and restaurants, also pretty greenfield.

There’s some legacy players, private companies may not have heard of, some call center players. But really, technologically, when you see, one of the ways that people measure success customers is when I order something in a drive through or phone ordering, do a how successful am I at completing that order? So order accuracy. Right. Humans, we know post pandemic, actually, there’s been deterioration even human quality.

You might drive into a drive through and say, I want a cheeseburger hold of pickles, you’ll still get pickles. Or you’ll say, I want French fries to get onion rings. That happens when we say humans are about 85% accurate. Well, our technology, purely AI, in many cases is exceeding human performance. So we go 85% out of the gate.

We work with them for several months. We’re 95% accurate. And on an AI to AI basis, we really don’t see many equivalent players. There’s some people who put a human, you know, it’s kind of Wizard of Oz and have a human there, and they’ll say they’re 95% accurate. But really, ours is fully AI.

We will have a human backdrop if needed, or a fallback. But, but generally, it’s greenfield. Now there are there are one thing our cofounder CEO, Kevan Maharajar, says, I’ve I’ve always looked this line, it’s voice AI is hard AI. It is tough to build. We, again, 10 in stealth have been building on this for for years and years, and it takes ten years to develop.

It takes you three years to realize. It takes ten years to develop. And we’ve seen that over and over again. We have actually investors in the company, major tech players who, tried to do it by themselves for a while and then came in and partnered with us. Probably one of the most notable examples recently is McDonald’s who, tried to build it in house.

They actually acquired some companies. They ended up selling, that capability because it’s too difficult to do it in house to IBM Watson. IBM, had a partnership and scaled to, like, a hundred locations as a pilot. And then recently, within the last six, nine months, all public information, you know, said, hey, we’re gonna stop this, relationship. We’re gonna look for voice AI solutions elsewhere.

So you see this happen over and over again. People are like, well, how hard can it be? Let me build it. But it truly is hard, and that’s why many customers come to us solutions. I know I’m going long on the competitor answer, but one last play I’ll say, you know, we’re getting a lot of traction in the enterprise side, financial services and health care.

And again, there’s some conversational AI players. We acquired a company last year, Amelia. It gave us footprint deeper into the enterprise stack into these new verticals for us. And one of the big things when we were diligencing the company, many of the we have seven of the top 10 money center banks. And, the biggest thing they wanted to see was how do you move conversational text based to voice.

And it’s just a hard horizon where they were using third party agents from the likes of Microsoft to Google. And we came in and we’ve been able to displace those third parties with our own solutions. For us, obviously, it’s moving into our own stack. It’s cost efficiency. It’s speed benefit.

But, frankly, performance wise, we’re able to do a lot more and innovate. And so a lot of different players in the different verticals we play in. It’s an early days in emerging space. We know more and more players will come in, but we think we’ve got a nice running start.

Tom, AI at Kantor, Kantor: It’s, I would implore the audience to go to the SoundHound YouTube videos where I a few years ago when I met this company was impressed. It is truly better technology, at least from a usability perspective. I’d be just double clicking on the competition before we leave it is we’re reading very recently about Alexa and even I think Apple made an announcement this week about maybe double clicking and doing some whole software uplifts in terms of conversational technologies. Do you think this could expand and kind of naturally acknowledge the market type of thing? Or do you think that this could change the competitive landscape in these big guys?

Well, let me make

Nitesh, CFO, SoundHound AI: a more broader point and then I’ll hit the specific question. So every, again, generation, there’s some new major inflection in technology. As I mentioned at the beginning, we think we’re in the new conversational, natural conversation era, led the capabilities that LLMs are providing in terms of flexibility and capacity to hold conversations, very complicated conversations, just like how humans converse. This is a new era. This is where technology is shifting.

So anytime you end these sort of brand new areas, there’s going to be all sorts of new competition. And incumbent players in one space are going to try to get their footprint into it. And I think there that’s a sign of a very healthy attractive market. And we welcome it and we compete head to head with many different players. I do think there’s validation absolutely.

We see it in the restaurant space, we see it in the enterprise space, we see players who have traction in one part of the market saying, hey, this is where conversations are shifting. We want to make sure we’re following track. So I think that’s a sign. And yes, there are going to be major players. Obviously, the companies with unlimited resources could spend unlimited amounts and could do a lot of things.

But I also will say, and I spent a large part of my career at mega cap companies, that innovator’s dilemma is a real thing. It is hard to reinvent yourself. When you have a big profit pool to be able to risk it and try to enter new areas is a very challenging dynamics. And that’s why we see new breakthrough customers all the or new breakthrough companies all the time. We hope to be that in this new era.

We believe, again, our right to win is our proprietary voice AI stack. And I think the difference also is in this world, as said many, many times, data is gold. And companies, our whole model is brand centric. I mentioned this in the auto space, but generally speaking, we are very flexible and agile with how data is managed. We work with customers.

We know it’s their customer relationships. So we work in partnership for how to manage the data. We want to use it to refine and improve our models and to scale. But we understand, like I said in the automotive space, that Hyundai, it’s their customer and we want to work with them. We also think of that with privacy.

We think of that with security domains. We think of that with all the capabilities we’re bringing. So our go to market motion is different. I raised that in the point of some of the names you just threw out. I spent many years prior to coming to SoundHound at Nike.

And one of the biggest challenges that Nike had was, it’s their customer footprint. And when they would start to partner with the big tech, how is the data and the customer conversation holding because there’s always risk of disintermediation. And again, our model is we’re brand centric and we so we have our own unique way of winning in this market.

Tom, AI at Kantor, Kantor: That’s a whole other vector of wanting to control the brand, some of these large OEMs. And I think that’s a loose segue to the diversity of your business in terms of different types of products, but also different types of rev rec, right? You have a your percentage of revenue from recurring sources has increased through the Amelia acquisition. Maybe talk about where you expect that model to go to in the future and if there’s anything you want to maybe say about the product royalty business as well, that’s about, I think, we model about 20% of your revenue.

Nitesh, CFO, SoundHound AI: Yes. So three pillars and I’ll unpack each of them. So the pillar one is voice enabling products, cars, TVs, IoT and that’s a royalty licensing business. And up until 2023 was the vast majority of our business, almost 85% of our business came from that. And these are if you think of the application in vehicles as a car is shipped, we get a royalty.

And the royalty rate depends on the capability we provide. So with Edge Solutions, you can take roughly kind of single digits per vehicle. When you get cloud capabilities, it could be higher price points. And we are the first company, who went live in partnership with Stellantis, went live with integration with OpenAI. And so that was early twenty twenty four with their premium line DS premium brand, and then they’ve scaled into many other brands.

And that GenAI capability is a price point above it. So if you think of like there’s 90,000,000 global light vehicles produced per year, take whatever average price you want it, but just for simple math, say 10 per vehicle, that’s $900,000,000 or nearly $1,000,000,000 of opportunity annually, call it reoccurring revenue, because as new vehicles are shipped, that’s when we get the collections. And there’s a it’s a lot more complicated because you do upfront work. You might have to build custom environment, product feature sets, new languages, etcetera. And then for if it’s a five year contract, for example, revenue recognition for that workload will spread over five years even if you’ve collect the cash.

But that’s kind of pillar one. And to your point, it has definitely shifted as we’ve grown pillar two aggressively this year, which is pillar two before you

Tom, AI at Kantor, Kantor: get to pillar two. On pillar one, because there’s a lot more IoT devices, there’s a lot more televisions, there’s a lot more thermostats.

Nitesh, CFO, SoundHound AI: Absolutely. Besides

Tom, AI at Kantor, Kantor: the 90,000,000 cars from New Jersey.

Nitesh, CFO, SoundHound AI: It’s a huge opportunity and How would the royalty business work in terms of

Tom, AI at Kantor, Kantor: because you can’t take $10 from

Nitesh, CFO, SoundHound AI: No, the price point will change depending on the application, but there’s a royalty rate. You could also put consumption economics on it. I’ll talk about the third pillar in a second. Second. But you’re absolutely right on the there’s projected to be 75,000,000,000 IoT devices in the next several years.

So and then you’re right, it’s smart appliances, but it’s all sorts of infrastructure. And one of the real differentiators with voice AI as compared to traditional technology architectures is that to for many of these IoT devices to have, a keyboard or a big GUI interface or have a mouse connection is really difficult. They don’t have the economics. They don’t have the space. But to unlock the power of voice AI, all you need is a very inexpensive microphone and, you know, $5 microphone and you can unlock the power of voice.

So so for us, we see that as a huge horizon. You’re right that depending on the what it is, I mean, a larger refrigerator or TV can can actually afford a decent royalty. But if you get into smaller footprint, devices, maybe not. I’ll I’ll come back to that point because it’s really important point that even a light bulb can unlock the power of VoiceAI and to have like a capability to be able to say time to replace my light bulb or those things can be connected through our monetization framework in Pillar three. But just before I go there, I’ll go to Pillar two, which has become the biggest growth engine.

It is now comprised of two thirds of our business, in 2024 that we just reported earnings on a short while ago. And pillar two is subscription typically. And think of the restaurant application. So in a drive thru setting or in phone ordering for pizza, per location or per lane, we generally get a monthly subscription fee. And that range could depends on the customer, depends on the number of locations.

And generally, it’s a fixed rate, but there also can be per order or per query. And oftentimes, like, these contracts are generally based on, like, up to a certain number of queries, there’s a fixed rate, and then to go above, there’s an incremental rate. We have the same applications with health care and financial institutions, which could be based on interactions. So if you call and you need to reset your password, you need to do some banking service like transferring money, or you need to book an appointment for a follow-up dental exam, these things are per interaction. And again, they’re subscription SaaS like typically.

And the vision was always and what we built in which we unveiled earlier this year at CES was our voice commerce engine, what we call pillar three. And to give you an example, this is connecting basically voice enabled services with voice enabled products. So just one simple, you know, many people who drove to this conference or maybe you’re driving into work and you wanna access coffee in the morning, you could just now instead of fumbling around with your phone, and looking for that app, you could just talk to your car and say, I’d like coffee. Well, the car knows where you’re going, and it could say, great. There’s a, you know, next exit five minute detour.

Use one coffee shop or there’s one closer to your work. Would you like me to order your cappuccino for pickup? That simple transaction created value to the user because, again, you’re not stumbling around. You’re focused eyes on the road, and you got access to the coffee you needed to pick you up in the morning. And that restaurant would would happily pay for the lead generation of that transaction.

The new lead is new business. And then if you consummate the transaction, the transactional economic or kind of think of the commission rate of it, they’d be happy to pay. And our model is that we share part of that commission with the product creator, or in this case, the car manufacturer. So if you’re a car manufacturer, even you’re incented because you’re like, wow. Not only did I sell the car and I got users engaged, but I actually can get monetization through this.

You can think of that with TV applications and watching football on a Sunday and you order the pizza at halftime. You can think of so many applications of booking for flight reservations or your next vacation or, you know, again, medical services because we just launched it at CES in January. The number of OEMs who are engaged, the number of restaurant partners engaged, beyond restaurants that are starting to get engaged, It’s early days. This is ultimately like we think in terms of meaningful revenue contribution is still a little ways out. But really early traction and very excited.

Yes. I think you mentioned that in your 2025 guidance. You’re not factoring in, if any, voice commerce. We’ll call this pillar

Tom, AI at Kantor, Kantor: three. There was a couple of mentions about proof of concepts, a big uptick or in the number of proof of concepts. And it seemed like there was a flywheel effect to that to those wins. Talk to us about how that’s going to impact the model from a flywheel perspective like voice commerce maybe impacting royalties or other way around subscription impacting voice commerce. It seems like there’s an interplay to the division.

Nitesh, CFO, SoundHound AI: Yes, exactly. Going back to the example of the person driving into work and picking up their coffee. Again, the user is excited because they’re getting exactly what they want when they want a very seamless transaction. The restaurant gets new leads and the car manufacturer is getting new revenue streams. So that is actually the ecosystem that we’re building, that flywheel effect.

So to your point, we’ve gotten POCs not only with the, by the way, our existing automotive partners, but we’re getting new automotive partners who are saying, how do I get into this? Even if it might, you know, maybe the royalty business will come down the road. I just want to get into this ecosystem because, again, from a Pillar one perspective, they’re seeing new revenue economics, connectivity for new services for their drivers or passengers. Restaurants are super enticed and they’re leaning in because they’re like, wow, now I got people just driving and I wanna make sure when it shows up on the map that my restaurant is on there. If you’re gonna pick pizza, make sure I’m on that list.

They see a new way of getting leads. And then again from, from other ecosystem partners, it’s like how people, I don’t know the latest data, but I think it’s like people on average spend sixty minutes a day in the car, you know, through commute and so forth. So it’s a captive audience. And again, safety issues or, you know, and and it’s and it’s also platform providers that are engaged here because voice, again, it’s how humans have interacted for our history, with one another. It’s the most natural way we we converse.

We’ve learned to type really fast with our thumbs over the last, whatever, fifteen years, but we don’t have to learn how to speak. And so to be able to access services and things that you need in your daily life, we’re getting a lot

Tom, AI at Kantor, Kantor: of traction. It is building a flywheel. The, you mentioned the Amelia acquisition. This was a relatively large acquisition for SoundHound. Maybe just like refresh us maybe the why and why now.

It seems like with this large voice commerce opportunity, again, going back to the queries that we get from investors and we have them as well, why not just kind of take that path and not try to double down on healthcare and financial services subscription deals, which is it’s a big business as well, but it’s kind of maybe more of a we talked about Enterprise Connect earlier, it is more of an enterprise type of business as opposed to commerce. How do those fit?

Nitesh, CFO, SoundHound AI: Yes. So I’ll go back to our foundation is a tech platform, the voice AI platform, and we believe it’s, it will be ubiquitous. So voice will be the preferred way that people seek access to services and, and we’re just trying to, again, voice enable the world with conversational intelligence. So when we started down this path, we didn’t necessarily say we want to limit ourselves to certain industries or sectors, but we thought of our entry into restaurants, which is a big opportunity. On its own, it’s billions of dollars of annual opportunity.

And like I said, running start, more greenfield competitive landscape, we think we’ve got a lot of traction. But for us, we believe restaurants to us was like what Books was to Amazon. They were building a big e commerce ecosystem, but they weren’t stopping with Books, right? They wanted to start with that. And we started to get great traction with restaurants over 10,000 locations and great partnerships and continue to scale seven of the top 20 QSRs now and many more in pilot that we’re hope to unveil over time.

But I would say at this point, last year, we weren’t necessarily looking and screaming to say we need to diversify industries. But we do have in a burgeoning market and something that’s moving really fast, you need to keep a pulse on what’s going on. And like we talked about, a lot of competitors moving in, a lot of new players showing up. And we have a dedicated team that’s constantly on the pulse of what’s going on. Do we want to learn in, you know, who’s a new competitive threat, new partnerships we want to get into, and sometimes that manifests into an acquisition.

With respect to the Emilia transaction, we met the team. We got to know them a little bit. It we could have had a lot of runway with the automotive and restaurant sector in the IoT, but it was a time and a place that made a lot of sense and attractive valuation from our perspective and ability to jump into enterprise settings like healthcare and financial services that on their own because of the regulatory, the security, the complex integrations probably would have taken us several years to build on our own to build a go to market motion to accelerate that journey and build that capability, made a lot of sense for us. And again, an attractive economics. And a lot of it, I mean, there’s 15 reasons why they don’t happen.

You need to have a lot of things click and ultimately it comes down to team and culture and aligned vision and sellers who are motivated and that made sense. So we’re very excited and the early traction with it is great. The teams are integrating really well. We have dedicated enterprise team. We’re investing in it.

They went through their own journey. We’re kind of re inflecting in certain areas where we see outsized return opportunity. And there’s real complement. So since we are the voice AI tech stack, for example, our speech recognition engine, which we benchmarked, we talked about in earnings how comparable to other major players out there. We talk about performance of our new Polaris speech recognition engine compared to Google and how it’s 30 plus percent better in performance.

And we compare it to OpenAI Whisper, and it’s how it’s equivalently better, but also at one tenth the model size, meaning the cost efficiency there is. So we’re competing against all the best out there. We’re winning on that. And we know we can, in these major enterprise customers, displace those engines, put our own voice stack in. We don’t have to pay third parties.

That has multiple benefits, not only a performance, but also our ability to capture data, improve our own models. So for many reasons, it made sense for us. We’re not on the path of just acquiring for no reason, like we’ll be thoughtful about it. We’ve like you said, it’s a meaningful side. So we’re in digestion mode.

We’re integrating. Early days are very positive, but we don’t take things for granted. So we’ll be mindful about opportunities as they come. We’re not out there thirsting for the next one. But if something makes sense, then we’ll definitely take a look.

Tom, AI at Kantor, Kantor: Right. Let’s pause there for a second. Are there any questions from the audience by chance? Right here.

Nitesh, CFO, SoundHound AI: Yes. The first question to give you the history, so Nuance had the relationships and they definitely, you know, had the partnerships. In fact, when we first unveiled our technology and there is a nice, just a quick one minute demo on our investor page if you want to take a look, and this is from 2015, ’20 ’16. So even compare to state of the art today, but it was well differentiated then. When we unveiled that, we had OEMs come to us because they were looking for somebody who had better innovation, could move faster, was much more flexible to work with their existing relationships.

And that’s how we got entry and we had actually many of the OEMs who invested in us and still are investors in this company. And that’s what got us traction to where now we have 20 brands, that we work with. And there’s a couple of things on the technological footprint. First, again, we built native stack up, from the ground up and, you know, getting built by our engineers that have been together. But by the way, this is one side differentiator I think that doesn’t get on the headlines for SoundHound.

In Silicon Valley, we’re in the backyard. We’re in Santa Clara. You know, we actually have a founding team and a bunch of engineers who’ve been together for over fifteen, twenty years. And that is a rarity in the valley, and I think that gives us a lot of, you know, advantages that sometimes don’t show up on the surface of the P and L. But anyway, back to your question.

So we had differentiation, for example, on speed accuracy and low latency. We had a technology we called speech to meaning, which means where traditional stacks and the ones for Nuance and others were, like, speech to text, and then you go from text to meaning, two step process. We did those simultaneously inspired by the brain, and that again allowed for two two benefits. One, it was faster, lower latency, but it was more accurate. It was no error propagation.

If you had an error in step one that would carry forward to step two, we would be able to catch that early. There was other ones with the architecture on how we did natural language understanding. We had our own domains. We brought we brought in and got entry originally with cloud capabilities. So Nuance was traditionally on the edge stack.

We came in and integrated with cloud, and then we subsequently would go in with cloud and then displace on edge. So there were we have generally we in the OEM space because relationships matter, our disadvantage was those long term decade long relationships. But we always won on technology, and it was a constant innovation. We’re also the first, like I just mentioned, to bring integration with OpenAI stack and ChatGPT with Solentes vehicles. We have an arbitration engine and then sort of agnosticity to our framework that allows us to integrate with any LMM out there.

I mentioned OpenAI. We work with perplexity. We build models off of LAMA, and we have our own knowledge domain. So we have this sort of what we call the yin yin yin structure of, we could build our own and we can partner with the best LLMs out there and we can arbitrate. So we have and by the way, we have full stack solutions, edge, hybrid, cloud.

So all of that allowed us to bring new innovation and a speed, which as I mentioned, one of the best ways in the auto space to look at like how are EVs choosing. And we’ve won now several EV brands around the world, again, I think on the basis of tech. So hopefully, that’s one color commentary. I can answer it for other industries as well because we think on the restaurant side, it’s even more distinct. But maybe to jump to your second question.

So our balance sheet is very strong. At the end of I just reported, at the end of the year, we were $200,000,000 of cash, no debt. We have more than enough. I’ve also mentioned in the outlook that we expect to be adjusted EBITDA positive by the end of the year. For us, software company, adjusted EBITDA is a really good proxy for free cash flow.

So the burn we’re moving from losses for many years to now breakeven. And I’ve conveyed, and I’ll say it again, the expectation the investor should have from us is that at full like sort of down the road as a software company, we should have 70% EBIT margins, 30% sorry, 7% gross margins, 30% EBIT margins. Let me check that one back, 30% EBIT margins. We’ll scale to that. But frankly, for the next horizon, thinks a few years, we’re probably going to be in the breakeven zone because we want to keep investing.

So to your question on where do we see opportunity, there is outsized return opportunity. When you’re in these major inflections of new technological era, we want to reinvest capital because there’s outsized returns. There’s returns well in excess of the risk adjusted cost of capital. And so we will be reinvesting in the business, but we could do it from breakeven profile for, you know, the subsequent years. Just a little bit on because you have such a good EBITDA margin, right, you say you’re gonna try to invest more, right, to, you know, bring it back down, Yes.

I think organically, we want to voice enable the world. So we want to go further in new industries and grow deeper in the restaurant stack, for example. Just in The U. S. Alone, to give you a data point, there’s probably 250,000 quick service restaurant drive thrus.

Drive thrus on average, and I won’t give you specific names, but there are customers north of this. Average roughly is $1,000 per lane per month recurring revenue. And if you take the hundreds of thousands times that, you quickly get billions of dollars opportunity. I’m not even talking about phone ordering, which might be a smaller footprint, $100 per location per month. But from a recurring revenue basis, it’s massive from the scale we’re talking about.

We have a running start. That’s a huge TAM. Healthcare, financial services, millions of interactions a year that can be automated. And by getting customers, and one of the things we’ve done with the acquisitions, you’ve seen a little bit of a reduction in our gross margin, is we’re going in with some parts of our business that are call center oriented. Well, we’re not trying to be in the call center business, but what’s great about is you get access to production ready data that you can utilize into our engines and automate over time.

If we can get more and more of these interactions, more and more of these orders, we can just continue to build and leverage that and scale. So when and this last quarter, we talked about entry into new industry energy, like, large electric utility, you know, for people to call in, like, what my power’s out, you know, what’s going on with my billing. All these use cases, how the world is moving, we want to aggressively go capture that and be first movers in this business.

Tom, AI at Kantor, Kantor: Please.

Nitesh, CFO, SoundHound AI: Yes, we filed this morning, so it’s out there for everybody to see. So it was just a few day delay, but thank you for the question. We so first of all and everything was materially exactly what we talked about a couple weeks ago earning. So you can go through it. There’s obviously a lot more content with the notes.

Let me go to your question on complexity with the acquisition. So we bought companies we have, as a small company, we have three ERP systems. We have a lot of processes that were manual, that we’re automating, improving over time, fully disclosed upon where we are in a control environment. We’re early emerging growth company. We’re trying to aggressively scale.

So we’re going to continue to make meaningful investments in our infrastructure and processing controls. And obviously, we don’t want to delay. We got it done. And we’re going to this year integrate our ERP stacks. We’re integrating payroll systems.

All these things are just as a small company with limited resources, we’re trying to do a lot. So part of the team to kind of that we’re getting through this, again, we got it filed today and nothing really meaningfully different

Tom, AI at Kantor, Kantor: than

Nitesh, CFO, SoundHound AI: what we said a couple of weeks ago. Yes. Yes. First of all, I think any new breakthrough technological area, there’s going to be a lot of new competition, and we welcome it. It’s a sign of a good healthy market.

I think there are two

This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.