Anne Mushow spent 15 years in the water utility industry watching the same problem repeat itself — utilities couldn't get granular data because they were stuck reading meters once a month, by hand. When she was inside Amazon Web Services and watched them quietly launch Amazon Sidewalk, she recognized instantly that it would make the old way obsolete. She left to build Subeca: a retrofit device that clips onto any existing water meter in under a minute and turns it into a connected device — no new network required.

She identifies three things that make Subeca's timing right now:

Subeca has only seven technical staff — yet has grown 10x in revenue two years running. Anne attributes this to two specific AI-driven workflows that replace what would have required 20–25 people just 15 years ago.

Anne was inside Amazon Web Services when Sidewalk launched. She watched the incumbent OEM players respond with skepticism and slow, cautious evaluation — giving Subeca roughly a two-year head start in development. That head start has hardened into an unexpected competitive moat.

Anne Mushow (00:00) I find data to be the great equalizer. When it comes to conversations, it doesn't have emotions. It really helps you get to the root of the problem.

Roland Siebelink — Host (00:41) Hello, everyone. Welcome back to Breakthrough AI Operators, the show where we talk with founders who've made AI structurally native to how their company works. Not tool adoption, not productivity wins, but actual operating model breakthroughs. Today's guest ran a 90-person team deploying smart water meters for one of the biggest names in the industry, and she hated how hard it was.

Then she watched a hyperscaler quietly release a piece of infrastructure that most of the industry completely ignored, and she recognized instantly it would make her old job obsolete. She left to build the company that proves it — a device that clips onto a water meter already in the ground in under a minute and turns it into a connected device. No rip and replace, no network to build or maintain. Global water giant Suez backed her, and so did Amazon, as an investor, a customer, and the network her whole company runs on. With that, everyone, meet my guest today, Anne Mushow, the CEO of Subeca. Welcome to the show, Anne.

Anne Mushow (02:16) Absolutely. Thanks so much for having me.

Roland Siebelink — Host (02:19) Of course, it's an honor. Anne, let's dive right in. You call Subeca the data layer that AI models need, but then you also said you're not quite an AI company. Walk me through that split.

Anne Mushow (02:31) Yeah, absolutely. I think what becomes really important when you're looking at the abilities of AI is that the fundamentals — what is the data that is available to make AI actually effective and usable — are everything. There are lots of software solutions out there that folks are building for the water industry, but what they keep banging their head against is that the fundamental water data — how much usage, and the granularity of that data — is highly limited. And that's because the water utility industry remains trapped in 100-year-old technology where you have to put eyes on a meter to collect a reading, which means you probably only do that once a month or maybe once a quarter. So identifying leaks, understanding usage patterns, understanding non-revenue water just becomes that much harder when the data granularity isn't there.

(03:52) What I wanted to focus on with Subeca is making it easy for utilities to gather that data, which then enables other tools within the water utility ecosystem. We're really trying to drive digitization forward because the data acquisition part has always been a bit of a barrier for utilities. And we're at an interesting inflection point in the industry where increasing demand and increasing constraints on supply — whether that be from climate change or just an increase in demand — create a deeper need for an understanding of usage patterns, because there are new sectors with higher levels of demand. We've all heard about AI data centers and their usage of water.

(04:14) So you've got all of these pieces coming together that says: I really need to have a much better understanding of what my system is capable of supplying, where my loss is, and how am I going to be able to support where we're going in this digital age over the next 20 to 30 years? Because of that, AMI technology — advanced metering infrastructure technology — has been around for 20 years. I've been working in it for 15. But in the water space there just hasn't been as much demand, particularly for the market segment of smaller systems that have 10,000 service connections or under. It was too expensive a capital investment to go after the more complex technologies that are available. We're really trying to create that simple solution — an Amazon-like experience where you take a device out of a box, plug it into your meter, get connected, and you're able to get hourly data from that.

Roland Siebelink — Host (05:15) Your company, Subeca, is still a relatively small startup, I understand. How much impact are you creating with how small of a team these days?

Anne Mushow (05:25) For us, we have to be creative. We are very lucky that we have big partners — Suez brings decades of deep institutional knowledge — and then we're able to leverage our big brother of Amazon, who is on the cutting edge of technology. Marrying those two things together has been a lot of fun for us, frankly.

(06:00) Many of the folks on our team came out of AWS. They have that very innovative big-tech mindset and the ability to move forward quickly and use the tools that are available from Amazon to innovate. My team is able to utilize AI very efficiently to help us move much faster. We can create test cases and create lots of automation to ensure the solutions are going to meet the demand of a utility-grade solution. But ultimately, for us, it's about being lean, very focused on what the customer needs without a lot of bloat, and leveraging the partners with deep institutional knowledge.

Roland Siebelink — Host (06:27) Okay, great. What does that mean in practice? How many people do you have on your team? What are their roles? If we had to compare this to a company started five years ago, how big would the company have been? What are the roles that they're able to replace with AI agents today?

Anne Mushow (06:47) It's a fascinating difference when you look at where startups have been and where we've been able to innovate quickly. From a technical perspective, we are only a team of seven technical experts, and we've got a few operational folks and a manufacturing support team because we do make hardware. But I think if you look at where we would need to scale if we had existed 15 years ago, there's no way we'd be able to do what we're doing with a team of less than 20 to 25.

(07:20) Ultimately, a lot of it is about the ability to iterate on code with the use of AI. We know what the use cases are deeply, and so we have that contextual product focus, which allows us to create small pieces of our code that are then deeply validated and tested through a system test process. Having the ability to use those types of tools — as opposed to just generating all of that code manually or, as many others are doing, finding similar pieces of code and modifying them — we're able to leverage AI to do that, and that has been fantastic.

(08:10) The other place that we've been able to very effectively utilize AI is in identifying the right marketing qualified leads. In the utility industry, when you've got 50,000 water utilities all over the United States that you're trying to reach at the right time, it's really, really difficult, and you typically need a fleet of people. We've covered that in two ways. We've found distribution partners, which are critically important to our ability to scale. But we also use AI-driven lead generation from our side to make sure that we understand which customers are showing buying signals in public forums. Many of these conversations happen during municipal board meetings or HOA meetings where they say: we have this problem in water. A lot of those documents are published on the internet. So we can use AI agents to help us identify the right buying flags with the right environmental pressures — "Hey, I think they're ready." That way, we can appropriately size our sales team and point our distribution partners in the right direction much more efficiently than conventional utility sales approaches would allow.

Roland Siebelink — Host (09:14) Yeah, and that's been a theme in AI-based lead generation in many different industries — it's not just about defining your ICP and the right messaging, but also the right timing based on the signals that customers actually give. I've seen that in many different cases. What has your growth been like the last 12 to 18 months?

Anne Mushow (09:39) Yeah, it's been a really exciting time for us. I've talked a lot about our municipal market, but we've also had a lot of traction in the commercial and industrial space. One of the interesting things that we found is that hyperscalers also need water data. They often operate as a disparate, geographically dispersed water utility because they have all of this commercial real estate they need to understand. And so we've been working with hyperscalers like Amazon to help them understand their water data, which has really helped us grow rapidly on the revenue side. From '24 to '25, we were roughly 10x in revenue. And this year we're on target to at minimum be another 10x. We're scaling incredibly, incredibly fast.

(10:24) The big thing for us is that water utilities have the reputation of making decisions very slowly. I would say the increase in pressures that we highlighted early in the conversation have helped them adopt more rapidly, because the need is there to justify the spend. Also, targeting the right beachhead market has been super important. Smaller utilities have less cumbersome procurement processes — they're still highly regulated as public institutions, but they can go faster. And they have the highest need because they've been underserved historically, but they can also make their decisions faster.

Roland Siebelink — Host (11:09) That's amazing. And maybe to some degree this also points to my other question — when you're so dependent on gigantic partners such as Amazon and Suez, that can often impact your own roadmap and your ability to grow and innovate very fast. How's been your experience in working with those partners? Because I believe you had the original idea already in 2016, right?

Anne Mushow (11:36) Yeah, and it has been a long journey. And I know this will surprise everybody who's ever worked in a startup — we pivoted several times over the course of that life. We started thinking more about a utopian concept of water utilities and end consumers sharing data and having visibility to the same data. And the reality was that the water utility didn't even have much of the data we were hoping to provide to customers.

(12:01) That evolution of learning — for the original founders, learning where they started and where they needed to go — involved adopting technologies. Originally, we started with LoRaWAN as the only communications technology built into our devices, and we realized what we really needed was flexibility. LoRaWAN is a fantastic solution, but it doesn't work in all cases. It's still a fixed network that requires some level of maintenance by the customers. And Sidewalk became the ultimate game changer for that simplicity piece — that was the true unlock for this particular segment. Because we offer Amazon Sidewalk, which is completely infrastructureless as an option, most of our customers can leverage at least some portion of their system connecting over Sidewalk, and then we can augment that with the LoRaWAN connectivity we still have built in. They have this belt-and-suspenders approach to connectivity, which is unique in the water industry — most technologies only have one type of communications built in.

Roland Siebelink — Host (13:13) Okay. And for those of us not that well versed in Sidewalk and the other technologies, can you explain to us like we're five — what are the differences between those technologies and what made Sidewalk such a game changer?

Anne Mushow (13:25) Yeah, absolutely. When you talk about AMI in the utility space, there are different RF communications protocols that can be leveraged. One that everybody's going to be familiar with is cellular. Cellular is fantastic because it has really great coverage throughout most of the United States, but it does come with a higher operating expenditure need because you're going to pay for every transmission and connection that device needs to send. What that means is you've got this ongoing operational cost, which sometimes can be difficult for utilities to make work in conventional procurement vehicles.

(14:06) On the other side of that coin are technologies based on fixed networks where a network is deployed specifically for that particular utility. Those can be either LoRaWAN-based — which is an open protocol communications technology — or some can be an FCC licensed-based structure, which is highly secure but very tightly managed. And often what that equates to is very expensive, both to deploy and to support from an ongoing perspective. This is a higher burden on the front end and then a lower ongoing maintenance operational cost.

(14:41) Where Sidewalk comes in is it's a combination of the two in many ways — in that there is in fact no operational cost, because the gateways are actually already deployed inside consumer homes through Ring cameras, Echo devices, Ring doorbells that are existing in neighborhoods throughout the United States, Canada, and Mexico. Those are the locations Sidewalk is actually turned on. And so what we do is we utilize that as our infrastructure backbone. So it's a cellular-like experience in the way that you don't have to maintain it, without the ongoing costs associated with it. Because Amazon has already paid for it in the devices that consumers have deployed to help with their smart home solutions, we simply leverage that technology. We can leverage this communications technology at this really low barrier point for entry and create a unique experience that makes it possible for utilities who've historically not been able to adopt this type of technology to get there very quickly.

Roland Siebelink — Host (15:49) Excellent. We talked a little bit about your go-to-market and using signal-based lead generation. How have you managed your actual sales process? Do you manage distribution partners, do you completely outsource sales to them?

Anne Mushow (16:04) Not completely. We have multiple arms for go-to-market. The joy of the utility industry is there are lots of ways to meet our customers where they are. We do a lot of direct sales and we do have sales folks on staff. We also leverage distribution channels, which are highly regionally based — you've got distributors that cover the northeast, the southeast, the west coast, those kinds of things. And so we've set up those relationships, which is fantastic.

(16:36) They have long-standing relationships with these utilities because they don't supply just AMI technologies but across the gamut — if they need pipes for their distribution system or valves and pumps for their wastewater treatment, they supply all of those. So that long-existing relationship helps us move fast. We also partner very heavily with engineering consulting firms, who are often the folks who come in and advise utilities when they've decided to make this large capital investment. They help utilities across the United States set up the types of RFP processes that help them identify the right technology for them. And so for us, it's super important for them to know that a new game-changing technology is on the market with Sidewalk, because it is such a different total cost of ownership business model — it really helps them understand the value that it could drive for the utilities themselves in the long run.

Roland Siebelink — Host (17:30) And you've been growing 10 times, two years in a row. That reminds me of one of my own most successful hyper-scaling stories in the '90s. How long can you maintain this? How big is Subeca gonna be in a few years' time?

Anne Mushow (17:50) It's a really interesting question, and one we think about frequently. Sidewalk has only been turned on, as I mentioned, in North America — and that's where we've been focused. But Amazon announced in February of this year that they will be expanding to the EU, Japan, and New Zealand and Australia by the end of this year, with more to come on the roadmap later. And so for us, we're at this interesting inflection point of what is next and where do we want to expand — whether it's geographically to meet the needs of other underserved regions, because we really can help in places that have not historically had attention, or whether we want to look at new use cases like more information about pressure data, wastewater, level sensing, those types of things. And so because we have the core module technology figured out and everything else is just an input, it's really about where the next batch of revenue is going to come from.

(19:01) We'll be focused on a combination of geographic expansion and non-revenue water.

Roland Siebelink — Host (19:01) Okay. In general, how do you look at this constant conundrum when a startup has found some product market fit — of just exploiting that product market fit or exploring further and finding additional areas of product market fit?

Anne Mushow (19:16) That's a great question. I think that may be the most fun portion of our job. We've very much pulled on the leadership principle from Amazon that customer obsession is our number one core value. And in particular, it's the most rewarding part of our conversations with our utility customers day to day. Once we've gained their trust and the connectivity, they're telling us about the next problem that they have.

(19:44) And what we've been able to do — in particular with AI — is iterate on the base layer that we have in some small way to then drive and unlock more value. We really treat it as an incremental growth process that is unlocking more recurring revenue very rapidly. Just within this year, we've gotten a 10% increase in our split between hardware and ARR simply by listening to what our customers are asking for. We just focus on what is going to unlock the next round of revenue based purely on the constant feedback of our customer partners.

Roland Siebelink — Host (20:27) Okay. You're really saying as long as you stick with the same customers and build on that trust and the relationship that you've already built, then new product opportunities will just open up for you. I love that — that's actually really good guidance.

Anne Mushow (20:41) Correct. We rely on that stickiness for sure. Fortunately for us, utilities are incredibly loyal customers.

Roland Siebelink — Host (20:54) Yes, I could see that exactly — not exactly the early adopters that switch off to the new provider very soon, right?

Anne Mushow (21:02) Correct, exactly.

Roland Siebelink — Host (21:04) How is competition looking in your market? You don't need to name any names, of course, but in a growth market like this, there must be some other players.

Anne Mushow (21:13) There are. And what is interesting is we sit in a bit of an intersection. We provide RF comms and there are several other large OEMs that provide this type of communication. But typically they don't just provide communication — they often provide the meter plus communication. So we're a little bit ancillary. What has been very interesting for me is I was inside Amazon Web Services when they launched Sidewalk and felt very deeply that it was going to be disruptive. And it was very fascinating to watch the OEM technologies' response to this new communication and the skepticism around it.

(21:55) What we've seen is very slow, cautious adoption from those OEMs — in fact, they're still just evaluating from afar. What that means for us is that there's been tinkering in the space, but we've got about a two-year head start from a development standpoint, which is fantastic. There are other smaller players who have looked at the success of deploying Amazon Sidewalk and have found it a bit challenging because it is difficult to navigate inside the larger organization of Amazon if you don't have experience. And so an unexpected moat that we have around our competition is our deep knowledge and partnership with Amazon. It has really helped us unlock with speed.

Roland Siebelink — Host (22:48) Okay. Let me summarize what I think I heard — the jujitsu move of building on a new technology that's competitive to what some of your competitors have already built up, so they have no incentive to move away from their older technology, and then second, that you were actually to some degree the one launching Sidewalk and know everything inside Amazon around it. That's excellent.

If you're maybe more of an ancillary or point-to-point player as you said, then is there a risk that you would lose track of what's the overall customer journey — that basically you're only solving one tiny point of their problem and that they still have to figure out what other partners to put into the package? Or have you guys got a good solution for that?

Anne Mushow (23:42) For us, we have very much focused our roadmap on creating more of a turnkey solution through an ecosystem of partners. We focus on making sure the data is available, we want them to manage their network successfully, we want them to have visibility to performance, and we want to make sure meters — the cash registers for utilities — have the ability to unlock that.

(24:12) But we are also very willing to work with other players in the ecosystem. I think this is a unique position that we've taken — we don't have to be the whole solution. So we work with partners frequently. Maybe they already have an AMI solution deployed but need the last mile of connectivity, the last 5%. We're happy to work in that environment. We won't disrupt their entire software tech stack, but with very creative connectivity options, we can pipe that data into the software that they already have. Similarly, I mentioned our startup ecosystem — there are other partners that we work with that have billing solutions or environmental requirements reporting solutions that we integrate with as well. And so if a customer mentions a problem that we don't solve, it's very often that we have someone in our Rolodex that we can connect them to who's used to working with these types of partners and can help them move quickly.

(25:20) That's been a great piece of being part of the water ecosystem overall — choosing to lean in with partners rather than trying to create a wall around our solution. We want that data to get to the places where the utility needs to use it, and we're happy to connect them to others that can provide it.

Roland Siebelink — Host (25:27) Excellent. Anne, this has all been fascinating, but of course I don't want to just hear about the business. I also want to hear a little bit about the person behind the business. Where did you grow up? What was your youth like? Were you always predestined to be a successful founder?

Anne Mushow (25:51) Funny questions. Born and raised in North Carolina — I live in the Research Triangle area, which I somewhat feel like an oddball about now because there are lots of transplants in this area, thanks to the great universities we have in the triangle.

(26:06) I was destined to be some type of engineering nerd. My mom taught science in high school, and my father was a computer science instructor at a local college. I didn't stand a chance. I was always going to be a woman in STEM and loved engineering. I actually found the place that was most interesting — and the career that really drew me into water utilities — was failure analysis and understanding what was wrong with products that were failing in the field, which was my first role at Xylem. That grew into a love of quality, really deeply understanding how products need to perform when you've got a use case as tough as a water pit in South Florida with a water table where the pit is always flooded, and the high standards that are required to service a utility. I have stayed in water for the last 16 years for that reason.

Roland Siebelink — Host (27:05) Okay. What especially is it about quality that draws you so much?

Anne Mushow (27:10) For me, it's really about that commitment to constantly improving. I was fortunate during my time at Xylem that I was a Lean Six Sigma Black Belt in training. And it was a very data-driven approach to understanding what is happening in a problem that you're trying to solve in a really deep way based on the data, and then driving improvement using that information. I find data to be the great equalizer. When it comes to conversations, it doesn't have emotions. It really helps you get to the root of the problem.

(27:52) I think that was why I was so fascinated with my time at Amazon — because of the volumes of data that they deal with.

Roland Siebelink — Host (27:52) Who in your childhood or teenage years was the first to see that special spark in you — maybe an inspiration or a good mentor?

Anne Mushow (28:05) Certainly in my teenage years and in high school, I was very fortunate to have an AP chemistry and physics instructor — just a teacher at the local high school — who also ran the Science Olympiad team. Competitive science is the best thing possible for a nerd like myself. And so we would go do bridge building or robotics competitions, and those kinds of things. It was just such a rewarding experience to see the application of the things that we were doing in the classroom come to life. Definitely Mr. Creighton was one who was an inspiration — a demanding teacher but also one who was able to see the spark was there.

Roland Siebelink — Host (28:57) Was there also something in your childhood, teenage years, maybe early college years, that set you up to become more of an entrepreneur?

Anne Mushow (29:07) Yeah, I think always interesting is being a female — particularly a female in electrical engineering. I'm a biomedical and electrical engineering double major. I was always going my own way. I was almost always one of maybe two females in class, really loved the hands-on portion, and really just wanted to be elbow deep in everything myself. I was used to plowing my own field and getting into the details, going really deep. And I think that was just always a trend of: if I'm gonna do something, I'm not gonna let anything stop me. And if I see something that needs to happen, I'm gonna push forward with it. I'm gonna know all of the reasons why. Whatever decisions we're making, very database-driven structure to those decisions, but we're just gonna keep going. I think it was in my nature from very early on.

Roland Siebelink — Host (30:10) Okay. On the one hand, of course, as an entrepreneur, being persistent is probably the most important asset that you can have. On the other hand, sometimes you've spent years building things the hard way and then you start seeing things can get easier. In your case, when you moved from Xylem to over time at Subeca — was there a specific moment that you just had enough of the hard way?

Anne Mushow (30:39) I think there've been many of them for me. Change is not something to be feared — so be disruptive. I honestly believe, and I go back to this data example, being unapologetically accountable to what the reality is, making sure you're not putting a layer of emotion on top of the data and misinterpreting it, makes it really easy to be disruptive, even in your own career path. I've been really fortunate that every experience — every single place I've worked, even as an adjunct instructor at a community college — has helped me throughout my career. Whether that foundation helped me explain technical concepts to customers who don't know anything about Amazon Sidewalk, or as a Lean Six Sigma Black Belt, which I obviously end up using every day.

(31:47) Fundamentally, it's about moving those pieces forward in a way that's true to what the data is telling you. It is good to fail fast. It is good to recognize that something isn't working and pivot. If you're able to have that level of accountability, you'll be able to successfully push your way through on the entrepreneur side, I think.

Roland Siebelink — Host (31:59) Okay. Last question — for those founders and entrepreneurs being a few years behind you, what would you add as wise advice?

Anne Mushow (32:12) I'm gonna speak in particular for a second to those interested in water and climate. Many folks think water is slow — it has a reputation of being slow. And I would actually say: don't believe them, push forward. The belief is there, but the reality is the need is growing rapidly for innovation in the water space in a way we've never seen it before.

(32:45) And I'll give a shout out to Tom Ferguson, the head of Burnt Island Ventures, who's also on my cap table. He believes we're on the very lowest part of that S-curve for innovation and growth in the water space. If you have a passion for conservation and environmental stewardship, turn that passion into innovation and you'll serve yourself, but also an industry that desperately needs it. So go for it.

Roland Siebelink — Host (33:06) Excellent. Just love the power move of inviting hundreds of competitors into your space. Why not, right? It's always better for the customer.

Anne Mushow (33:13) Do it. Yes.

Roland Siebelink — Host (33:17) Anne, where can people reach you and what should they download or investigate about you if they want to hear more?

Anne Mushow (33:25) Yeah, absolutely. Obviously, we have all kinds of information about Subeca at Subeca.com. I'm easiest to find via LinkedIn — other podcasts are constantly publishing new case studies about what we're doing and how Sidewalk is changing how the water industry is getting data. That's the best place to find me. Just look for Anne Mushow from Subeca on LinkedIn.

Roland Siebelink — Host (33:48) Excellent. And if somebody knows me and doesn't know Anne yet, of course I'm also happy to provide an introduction through LinkedIn or otherwise. Anne Mushow, thank you very much — the founder-CEO of Subeca. This has been a great episode.

Anne Mushow (34:03) Absolutely. Thanks for the time.