What if the most powerful climate intervention isn’t policy, but precision?
Deep Dhillon sits down with Ryan Sullivan of Bridger Photonics to unpack how AI, physics, and aerial sensing are being used to hunt down methane leaks that can cost operators dearly and accelerate climate impact. Ryan explains how methane detection has matured over the past decade, and why Bridger’s approach (laser-based lidar tuned to methane) can identify the specific valve, tank, or piece of equipment responsible.
They walk through the full pipeline: scanning swaths over infrastructure, reconstructing plumes from point-cloud data, applying supervised learning trained on ~15 years of labeled leak history, and then having expert analysts validate results with tooling that overlays plume density, wind conditions, and site geography. Then the two zoom out to the uncomfortable questions, like why Bridger refuses to play watchdog, how trust and data ownership shape the market, and what the “north star” looks like; near-real-time detection, autonomous flight patterns, and predictive maintenance that catches failures before they happen.
Learn more about Ryan here: https://www.linkedin.com/in/sullivar/
and Bridger Photonics here: https://www.bridgerphotonics.com/
Check out some of our related content here:
Get Your AI Injection On The Go:
xyonix partners
At Xyonix, we empower consultancies to deliver powerful AI solutions without the heavy lifting of building an in-house team, infusing your proposals with high-impact, transformative ideas. Learn more about our Partner Program, the ultimate way to ignite new client excitement and drive lasting growth.
[Automated Transcript]
Ryan: we are flying over facilities or near facilities and using a combination of lidar. And regular old fashioned photography. We also take infrared data not only does the lidar capture sort of the gas emissions is also looking for topographical data, so we can see some of the structures and understand more about the physical environment that the leak is coming from.
As we're flying over a facility, we're getting, both, longitudinal data as well as top-down data so we can plot out where a gas plume is, how it's emanating from a particular source, and that allows us to get a better view of measuring. quantity flow, and direction and, and things like that, which again, points back to the specific geolocation where the leak is coming from.
CHECK OUT SOME OF OUR POPULAR PODCAST EPISODES:
Deep: Hello, I'm Deep Dhillon, your host, and today on your AI injection, we'll be exploring how AI and advanced sensing are tackling methane emissions with Ryan Sullivan. Ryan is the CTO at Bridger Photonics.
He leads the company's software and data strategies, scaling platforms that combine ai, machine learning, and big data to deliver real time emissions intelligence and drive measurable climate action. Ryan, thanks so much for coming on the show.
Ryan: Hey, thanks. Deep. Nice to meet you. Nice to be here.
Deep: Awesome. So we'd like to get started by, maybe tell us what your typical customer looks like, what they did before your solution came on the market and how things are different with your solution.
Xyonix customers:
Ryan: Great. So, uh, you know, Bridger Photonics is a methane. Detection monitoring, and essentially we allow our customers to, you know, have actionable data to try to detect and repair methane leaks, typical customers in oil and gas business. We also expand into other things like, you know, municipal waste.
Anywhere that, methane is a byproduct is somewhere that, that, you know, will try to engage to see what kind of methane they're leaking and help them, Create sort of actionable data to repair monitor and manage it. you know, it's a relatively young science, right? probably the last 10 years it's really come of age.
So I think there's a wide range of technologies from satellites to handheld cameras, from on premise, fixed position cameras. bridger's value prop is that we take a top down look at your leaks. We can really detect. To a fine precision exactly where the leak is coming from and help you quantify the amount of gas leaking, which are critical for any operations business to repair and manage their facilities.
Deep: Got it. So maybe let's take a step back 'cause I'm not sure everybody's familiar with methane leaks. What is the kind of scenario like you're piping methane, you're producing methane somehow in your network, methane's leaking out of a pipe somewhere. Why do your customers care? how much of it's regulatory versus trying to save the gas that's leaking?
Maybe paint us a detailed picture of what a typical customer environment physically looks like.
Ryan: Sure. So I'll, start with the standard sort of oil and gas 'cause that is our bread and butter and really where we've cut our teeth, everything from the production.
So when you talk about, drilling for oil, as those. Pumps are pulling oil out of the ground. Methane is leaked and is essentially a byproduct of that process. those gases are typically captured. for some customers that's a product. For some customers, it's a byproduct. really at every step of the way.
Methane can potentially be released. everything from, distribution, piping, the gas between facilities to refinement, where they're turning crude oil into finished products. Methane can be a byproduct you know, it's a colorless, odorless gas.
if it floats off into the atmosphere, it's a pretty extreme, impact in terms of, climate science. for most of our customers, this is a matter of, regulatory management, trying to make sure they're producing their product as cleanly as possible. But for a lot of 'em, it is their product.
I think natural gas is. Something like 98% methane. realistically, for a lot of customers. This is actually their product. when it's leaking into the environment, that's not only bad for the environment, it's bad for their business as well. So both answers are true.
Deep: So everything from capturing methane to the distribution process. Now I'm imagining before your solution, people walking around with maybe handheld devices or something, picturing like the Star Trek team winding up with their little tri quarters and running around looking for some, emissions.
Is that kind of how they do it without your solution where they're physically having to traverse the network of piping and the environment.
Ryan: generally speaking, they're only gonna marshal somebody to go out with a handheld device if they think they need to find something typically a combination of on-premise monitoring, fixed point cameras looking at facilities to detect whether or not they see gas. And then satellite, which is gonna fly over every couple of days and give big picture like is there gas in the general area? Where Bridger really stands out is we're able to detect within a few feet of exactly where that, gas is emitting.
So we can give the operations teams for these businesses much more precise, data in terms of exactly where to go and what to fix. we can tell them it's this valve, whereas a satellite's gonna tell them it's this facility. The handheld cameras are obviously really precise, but you also have to get right up on the leak and you're generally coming horizontal to the problem.
sometimes it can be a lot more difficult to understand exactly where the gas is coming from unless you have more precise data.
Deep: Got it. Not to mention, it's expensive to send out humans to go one large facility. so your solution is, you guys fly cessnas or drones or something, and then you're using lidar to like detect the methane plumes, something like that.
Ryan: Yeah, so the standard W standard, most of what we do is, either fixed wing or rotary aircraft, and definitely drone as well. we are flying over facilities or near facilities and using a combination of lidar. And regular old fashioned photography. We also take infrared data not only does the lidar capture sort of the gas emissions is also looking for topographical data, so we can see some of the structures and understand more about the physical environment that the leak is coming from.
as we're flying over a facility, we're getting, both, longitudinal data as well as top-down data so we can plot out where a gas plume is, how it's emanating from a particular source, and that allows us to get a better view of measuring. quantity flow, and direction and, and things like that, which again, points back to the specific geolocation where the leak is coming from.
Deep: Got it. our customers typically like hiring you to fly a regular flight pattern on a fixed interval, like once a month or once a quarter or is it more episodic where there's suspicion and then they, based on that, you go out
Ryan: the typical flights right now are quarterly typically we're flying a series of facilities on a quarterly basis to, determine and make sure that there aren't any new significant problems within that timeframe.
Deep: Got it. And then maybe like, walk us through the data science a little bit. So you've got a lidar image. What is it about the lidar that's detecting the methane exact, like how, how exactly does the. Gas composition differ from the air such that Lidar picks it up and then what's the signal you actually get from that?
And then I imagine you have to do quite a bit of cleaning to characterize it, build a 3D topo map of it, and all of that. So maybe walk us through that.
Ryan: Yeah, so the founders of the company were all optical and laser physics majors, right? Uhhuh. So these guys are world class experts in terms of, seeing how Lidar is going to be reflecting.
we're shining a laser down on the ground and the methane gas absorbs the laser light in a specific frequency, a specific part of the spectrum. So as that light is refracted back up to our sensor and we collect it, we can detect changes in the frequency of the light pattern
To know if there is, methane available on the ground. So we're, shooting a specific frequency of a laser that's tuned to, interact with the gas molecules. specifically.
Deep: And then you're looking for a particular filtration characteristic, of whatever the guests absorb.
are you doing that in a grid pattern or something? you're striping up and down a grid hitting the laser, getting all that data and you're looking for that. And then. I mean, at that point you, have a big matrix of data.
are you trying to reconstruct the details of the gas cloud or the mere presence and quantity? Maybe.
Ryan: we are reconstructing the actual details of the gas cloud. it depends a little on the facility, right?
Some facilities are huge rectangles. Others like pipelines are long stretches, right? Swaths. Our scan swath is, I think about 20 meters wide, maybe 30 meters wide. It depends on the, mm-hmm. The elevation we're able to fly at, depending on the facility, we may take a single pass. and get enough data to detect what's going on.
Or we may have to take multiple passes across it. but for a typical installation of a typical job, we're gonna fly once, twice, maybe three times over the facility. And, typically we're gathering all the data from directly under the sensor as we fly over it.
We're capturing everything directly underneath it. We just want to get enough widths to cover any potential equipment that might be leaking.
Deep: Got it. And like when you reconstruct that plume, is there some ki like are you using machine learning or is it kind of a purely physics-based approach to.
transform the LIDAR data into that 3D representation?
Ryan: a little bit of both. the original was deep, heavy computational physics. So it's math, very sophisticated math. we need a lot of, edge computing to calculate and understand the data that we're dealing with.
But there's no machine learning initially on the edge. once the data is captured, we start doing, varying types of processing. both with respect to understanding the facilities we're flying over, and understanding the characteristics of the gas that we're, that we're taking in and trying to do more, sort of predictive analytics around, you know, the, the type, the size, the quantity, the, the characteristics of the gas cloud that we're flying over.
so that we can, accelerate some of the more manual processes that are done visually by data analysts at our headquarters.
Deep: And like, to characterize those attributes that you're describing, is this a supervised machine learning problem?
Like, are you defining a ground truth? and if so, like how are you doing that? Are you doing that based on the physics models, outputs, or? Do you have some technique for, manually, like inspecting some kind of image transformation and like characterizing it or
Ryan: we're getting back from the lidar as a point cloud, right?
it's a 3D point representation of the detected gas, there's a bunch of different supervised models that can help us with the comparison of almost 15 years worth of us manually. taking, this data in and doing all of the detections and labeling things, we have, probably, you know, more data than just about anybody in the industry.
because we've been doing this for almost 15 years, so, we have a vast trove of detected leaks that we can then recharacterize back into the image data, the plume data, the point data to try to, predict and accelerate detection of the quantity, the size, scope, release rate and, actual location, right? Uh, you need to understand where it is in a geo-referenced sense, but also how high off the ground it is, to really understand the volume and scope of the plume.
Deep: Got it. So it's, kind of like, a combination of more generically trained lidar interpretation models with presentation to humans from maybe more refined labeling that's unique to your place.
specific objects that matter to you guys. and then maybe taking that, those that transform data, maybe you do some fine tuning on top or something like that.
Ryan: I think that's right. there's everything from noise reduction, right? We're trying to filter out background noise and dirt on the window as the laser's shining it out.
So there's a lot of refinement we do with the data that we get back. it's things like object detection. So what are we flying over? Is it a tank a separator or is it a pipeline or is it a valve or whatever, right. So, so some of that object detection, we're using more, trained image, computer vision models.
to, to try to specifically understand what kinds of, facility we're flying over and what kinds of equipment they're running, which can then help, inform the operators in terms of what they need to work on.
Deep: And then in terms of the user, are they getting a 3D visual representation of the facility with the plume and where it's at along with the estimated metrics of what you think the volume and, discharge rate and stuff.
Ryan: Yeah, more, 2D for today. It's mostly top down, but there are some circumstances where we do show a 3D representation. a little hint of what's, coming in the future. we'd love to do more 3D presentation.
we do it internally for ourselves. but exposing more of that to our customers is definitely a big desire.
Deep: What would you say are some of your biggest challenges in terms of transforming the LIDAR data and getting it to the point where you can get to this crisp 2D representation plus the actual predictions on volume and discharge rate
Ryan: Yeah. I mean, there's a bunch of different ways I could take that question, right? There's, everything from weather and, you know, sensitivity of the equipment and things like dirty windows that can really, spoil the data coming back to, operational inefficiencies.
How do we. Manage it. How do we understand it? I think, our biggest goal is to help oil and gas operators have actionable data that they can really,
Deep: Make sure that the data
Ryan: is as close to a hundred percent accurate as possible. we still end up using a lot of human in the loop. visually confirming and verifying exactly what it is that the algorithms predicted, to make sure that, that it's exactly what our, operators want and need.
there's still enough error and noise in the system that we verify and validate we hold ourselves to that really high standard because, you know, some of our, customers are, use data for regulatory, obligations and, you know, it would be awful if we were to report false positives and show data that wasn't appropriate in those circumstances.
And some of 'em are, deploying resources to go out and fix.
Deep: Yeah. You don't wanna waste their time and money. Maybe walk us through like one of these humans in the loop. what's the environment? are they in a no center somewhere and looking at this stuff.
what are they presented with visually are they making their decisions solely based on the lidar? the original point cloud and the interpretation, or are they actually visually going in there and looking at something and if it's includes the ladder, like what can you see to help you figure it out?
Ryan: Yeah, we have a data center, a call center, essentially an HQ in Bozeman, Montana, right on the MSU campus, where we have operators trained in understanding, not only the science, but also the tooling that allows them to do the analysis. and they're looking at everything from.
the point clouds. the 3D representation. densities, understanding, where the gas is concentrated, where it's less concentrated. They're looking at that, superimposed over the geography. seeing where that cloud is, where that plume is relative to equipment.
And they're using all of that data and a lot of sophisticated imagery, a lot of sophisticated tooling that had been built to allow them to, confirm the detections. And essentially label them as positive or negative and quantify them in terms of how much gas is leaking. All of those are tools that are available to these operators to manually validate all of the data so that we give the most accurate data back to our
Deep: customers.
Can you maybe gimme a little bit of a flavor for like what some of these tools are? is it more concentrated analysis within a particular location? So you're putting a lot more, CPU and GPU cycles at that particular thing, or is it something specific to what they're looking at?
give us a little more flavor for that if you can.
Ryan: Imagine facility by facility. they'll probably start with a top-down view of a facility and the gas clouds detected over that. we'll give them essentially a visualization of the plume and they can click into the point data and rotate it on a 3D basis.
And see things like elevation above ground level. where is the gas concentrated, where is it diffuse, and what direction is it flowing? They're seeing prevailing wind, you know, at the time that we took the captures they have access to all of the tooling and modeling that our physicists have put together over the years to take all those data confirm the presence or absence of the leak, and then tie it back to the emitter, the actual piece of, equipment that might be leaking Some desktop tools, some web tools, some incredibly sophisticated, analysis products. Some, more intuitive just at a glance. Got it. Information as well as we're superimposing that with a lot of our predictive modeling so they can see, what we think it is and, usually jump to conclusions faster in terms of, confirming the labeling.
Deep: What are some examples of false positives where they see something that's obvious that for whatever reason your model's missed?
Ryan: I mean, it's more about noise, right? So imagine you've got a bunch of facilities that are all close together and you have a bunch of different emission sources.
Right? Mm-hmm. Sometimes the models aren't as good at picking out each individual emission source, and it still requires a human eye to look at it. You know, you may have multiple neighboring facilities emitting at the same time, maybe from multiple parties, right? And the gas cloud is collecting into one big plume, and you really need more precision to understand the individual leaks, the individual.
Problems and to, sort of, quantify the amount of gas that's leaking at a, at a point place because each individual emission is as important as the overall plume.
Deep: I got it. So it's it sounds like yes, there might be some false negatives, or positives, but it sounds like that's not so much what it is, you're trying to capture their.
investigative capability as a human combined with this powerful tooling. I'm gonna read between the lines and say that they're sort of like contributing to a more detailed reported view that goes to the customer. it's not just machine output that the customer's getting.
It's colored with their experience
Ryan: different customers have different things that they're concerned about, different, types of leaks. sometimes this equipment leaks intentionally, they're venting a tank that's gotten highly pressurized different types of events. may mean different things to different people and different equipment may contribute to that. So our analysts are using all of that data and a knowledge of the customer that they're working on to make sure they are culling it down to the specific types of data that, that, customer, that client may want to see,
Deep: that's a potential source of false positives right there is the fact that there are intentional leaks that have to happen for whatever operational reason.
Ryan: Yeah. I think there's also environmental leaks, right? There are times when perfectly well-functioning equipment is going to have some amount of leaking.
it's understanding the difference between intentional leaks that they know about and unintended leaks that they need to, marshal resources to try to address.
Deep: Got it. So do you guys construct like a temporal view that's more longitudinal that captures maybe not just the few seconds or that you're sampling the plume, but days or weeks or months or even years of the plant maybe walk us through what that kind of perspective looks like?
Ryan: Yeah, and, it starts with the short term, right? Some of our operators are very concerned about what we'd call persistence. the only way we can essentially detect persistence is if we fly twice. So there are certain facilities where we'll fly twice within a fairly short, maybe less than a week.
ideally to detect. If it's leaking twice, then we know there's a real problem. Another general source of false positives is, imagine a truck that's parked on a facility that's leaking, right? That truck is not part of the actual permanent equipment on that facility.
That truck is probably not something that they want to remediate as part of site operations, right? It's not even there.
Deep: What do you guys do in that case, by the way?
Ryan: you know, it depends on the amount of leak, but we'll typically detect it.
Label it and tell the customers that, this plume is likely, emanating from this truck. And something that you can either ignore or, go have the truck looked at. we can label it, we can definitely detect that it's coming from the back of a truck.
Deep: So on your injection, we kind of cover three areas. One is, what is it that you do? I feel like we've covered that pretty well. the other thing is how is it that you do it? I feel like we're getting there, and covering quite a bit of it.
and then the third area is I mean, we call it should for like simple summarization, but it's sort of like getting at what are some of the ethical issues that you guys face. It seems like there's probably, three different people that could buy your services.
three different types of organizations, one are the, emitters themselves who are trying to like, keep themselves from being, in violation of regulations or to capture their gas for whatever reason. I'm gonna guess that regulators probably hire somebody like you, if not you guys, to,
Audit organizations, you know, and companies for emissions. And then there's probably like watchdog groups or nonprofits. So are you guys solely in the camp of selling to the emitters? And if so, maybe why? And If you aren't, like why aren't you just a service for hire sort of regardless or divorced from the the who?
Ryan: Yeah. I think the most important thing since day one, we have positioned ourselves as picks and shovels for industry. we are here to help oil and gas businesses. Quantify, identify and action any leaks to try to fix the, the global methane problem. that discipline is critical to how we approach the market, and I think if we were to service those other customers we would start to run into skepticism.
And concern from our customers. in terms of, them not wanting to work with us, not wanting us to fly their facilities, for fear of, you know, us becoming, more of that watchdog persona.
Deep: Yeah. It's a, there's a key trust factor here. I imagine they, they trust you guys.
To take all of the data you gather on their dime and give it only to them. jumping back to machine learning for a second. So do you learn across customers in your models? Is there some kind of notion of that at the end of the day the leaks are very similar, probably I'm guessing, across customers.
So are you allowed to take that to like improve your models?
Ryan: Yeah, so we own our data, all of it. and we use that data to improve our models, to improve our data intelligence. we use our data in the abstract and in the macro sense to inform. regulatory and government and, and feedback.
Like we want to be a good player in terms of helping really the international bodies that are regulating these types of leaks and coming up with patterns and practices for monitoring and managing, overall leak detection. we're very involved in those conversations, but we're not going to expose any of our customer's data directly
Deep: We take great pains
Ryan: not to expose any of our customer data in even a sort of anonymized sense. But we would, for example, get involved in conversations about, the overall leak profile of say, a given basin in a given year. we have great data because we're flying.
a significant percentage of the assets that are in a given field, and we're flying with great precision and great accuracy. So we have an opinion on whether or not that basin has overall improved. Its, methane management or has. decreased in terms of its capacity and capability and maybe things have gotten worse.
we'll use that type of information to help inform the conversation about how we think, regulatory action can improve or move forward.
Deep: And is that, something that you guys do because those regulatory bodies are customers of your perspective? Is it something that you do because you're mandated to by law or something?
Or is it something that. You do? for Goodwill or to help the scenario out?
Ryan: simply put, we've got a some of the best methane detecting laser.
Deep: probably in one of the best positions in the world to comment on this problem that all of us and our children and grandchildren face.
Ryan: Yeah. And so I think it's incumbent on us to be a part of that overall mission. I think we have some of the best and most actionable data, both on the, you know, finite scale, but also on a global scale in terms of, you know, moving that conversation forward. And so we have a, deep commitment to being a part of the conversation in terms of how do we manage, how do we monitor, how does, you know, the industry hold itself to account in terms of what are, you know, the long-term, goals.
And in, solving this methane problem globally. But again, it's really critical that we're not part of the sort of monitoring state because we're not trying to shame our customers into better behavior. We're trying to help them, improve this position in the world.
it's something that was surprising to me, getting deeper into the. Into the industry is just how environmentally forward a lot of the oil and gas businesses actually are. I mean, they care deeply about making sure that this problem is solved, you know, globally, and they're just involved as involved in that conversation, largely as as anybody else.
And again, we want to play the role of helping facilitate them in solving things moving forward.
Deep: So maybe let's jump up a couple levels and I'll ask you to. Maybe speculate a bit, but what is the global perspective on, on methane emissions are things getting better globally and how much of a problem is it today relative to, 10 years ago or 20 years ago?
Ryan: I guess I don't have a precise answer for you there. without getting too far into politics, there's a lot of puts and takes. methane is a growing concern and was a growing concern, you know, really for the last decade or so, obviously domestically here, there's some.
There's some pushback against that now, but I think the international community is still, pretty focused on eradicating this problem. I, I couldn't tell you because I honestly don't know whether we've made good progress towards it or not over the last few years. the problem really has only begun to be attacked in earnest in, the last decade, So I think, we're in the first few innings of this fight and you know, hopefully with Bridger's help, we can wake up a decade from now and say we've really, put it behind us and we've managed to, wrestle it to the ground.
Deep: I mean, that said, at least within the.
fence of your customer space. I imagine you have a good sense that with respect to whatever they were admitting before that's been mitigated to some extent otherwise.
Ryan: No, we've got great data in terms of annual emissions that we've detected and helped mitigate.
So absolutely we're helping these businesses every day. Close the gap. And I think any one of our customers could provide testament to the fact that, using bridger data is the most actionable way to solve their leaks and fix their emissions. our biggest constraint as it comes to that is the periodicity.
Right. We're only flying once a quarter and that fact means there's still some pretty big time gaps. And I think I'm focused on trying to figure out how to fill in the picture between those, points and get to more sort of, you know, real-time actionable data and make sure there isn't a 90 day window between when we've, confirmed that things are working and when we check next
Deep: What are the constraints there? I mean, that's a interesting point. Is it the fact that you need some significant human in the loop resources to interpret that data? Or is it the cost of assessments and the flight patterns? what are some of the things that constrain you and keep you from doing, daily or weekly or monthly, um, flights.
Ryan: The simple answer is yes to everything you just said, cost is certainly prohibitive and, marshaling fixed wing resources on a regular basis can get expensive. marshaling helicopters or any other. Rotary is even more so even today, drone continues to be, You know, prohibitively expensive in terms of, you know, having a line of sight operator that has to, to watch the drone fly.
limited inventory. So our ability to actually have sensors in market in the various regions where we want to cover, is gonna be another factor, right? We can only have so many, so many planes in so many places. although, you know, continuing to manufacture as quickly as we can to address that problem.
And then, yeah, data processing, although less and less of a problem and I think we'll wake up in a year, and have that, much more under wraps as we. Are starting to roll out more and more data intelligence products and have higher confidence in our products. but the data management is still a big time suck and making sure that we're as accurate as possible.
You know, it takes a lot of time and a lot of humans, But is also very well appropriately built for machine learning. and frankly, most of what we can learn our way out of is tedious, time consuming and mind numbing work anyway. So, uh, hopefully we can reserve the parts of the job that are cerebral and interesting and take away all of the, null sets and not a problem, click through that people have to do.
Deep: I'm. curious, like, you know, if you fast forward out five or 10 years, what's the holy grail vision for your. company's set of services. you mentioned, for example, the, the drone operators kind of standing by and there's a number of companies now where the drones are just self-operating.
You walk a grid pattern with a cell phone, the drone wakes up every half hour flies, the grid goes back, recharges repeats forever. Unless the surface area is like really massive that you're covering. I imagine over time, humans in the loop are doing less and your, models can do a lot more reasoning and a lot of that higher stack thinking that they're doing to like formulate the narrative.
walk us through like five, 10 years out if everything goes well. Like how is your solution different? how's the world better off or worse off if you think that.
Ryan: I think, you know, you put your finger on a lot of the most important things, right?
Autonomous, drone management is gonna be a big part of the future, as we can get our sensors to be lighter. right now we require pretty heavy lift drones in order to fly the equipment that we fly. as we get it lighter, and as the drones become more capable and, flight duration improves, I think you'll see a lot of the type of solution that you just outlined becoming more and more feasible.
These flight platters don't change once we've plotted 'em out, They're always the same. So autonomous flight is certainly something that we will continue to lean into and develop for. Particularly in terms of data processing, particularly for the sort of more urgent, type of models, right?
So alerts, major leaks, bigger problems. I think, really deep integrations with our customers to where those things are. Realtime processed in the device, confirmed and detected and immediately shipped to customers is a, big focal point for me and something that I want everybody to be thinking about, right?
Our North Star really directly from the plane to delivery to the customer. and that's gonna take a long time to get there and a lot of work. but we think it's ultimately achievable. And the last is really closing that gap in terms of, predictive intelligence. Like how do we know when equipment is likely to fail?
How do we take other signals that businesses may already monitor, whether that's, the on-premise cameras or some kind of iot sensing. Within the industry, right? Flow rates in, pipes and, pressures in, in tanks, et cetera. How do we start to integrate some of that data as well as any of the other data that are available to come up with more predictive and more realtime monitoring.
to where we know when something is likely to fail and maybe we can then change our monitoring patterns for those types of equipment. anything along those lines, is interesting to me and something that I consider to be part of the long-term vision.
Deep: I imagine there's a wide variety though, of sensors that your different customers might have. So that might end up a pretty heterogeneous environment that you can't control as well as the one you have now. So I imagine there's a lot of challenges there. One question I have is, are you always only looking at methane?
it seems like there's a lot of other things you can figure out when you're flying over somebody's plant with, lidar, you can probably throw other sensors on that device too if you need to, are there other proximate problems that you would grow into over time?
Or is it really methane that kind of dominant and only focus?
Ryan: methane is definitely number one. And, and when we talk about what our customers are asking us to detect and manage, that's, one, two, and three. I think CO2 continues to be interesting. It's, it's a different problem to solve. there are other gases that are less prevalent There's other things, I mean, we're flying over these facilities. They have other providers that fly those facilities for other things, like, right of way encroachment and equipment degradations. there's other parts of that stack that we certainly could get more involved in and are constantly, looking at RD and trying to understand how we can better move into those spaces.
but methane continues to be. the big source of our problem set and frankly, there's a lot of additional opportunity for us to improve what we're doing, you know, staying within this specific space. we don't need to necessarily expand other services immediately, but always looking.
Deep: Yeah. And one question I have is how does your customers Drive level change with respect to the regulatory environment. Like right now we have a new administration. They're very, you know, I don't think it's any surprise that they're pretty anti-regulation and they're chopping a lot of these. clean climate initiatives and regulations, does that hit your business and hurt you guys? Or do you hop to different countries where, people still care about this how does that affect you?
Ryan: Yeah, a couple of things I'd say one is, what's affected us more is the dollars flowing out of the, federal administration,
There are a lot of big projects, a lot of. funded activity that is all of a sudden facing a lot of headwinds in terms of the dollars. So certainly that's where we've seen, I think the biggest headwind beyond that, these businesses, generally speaking are international businesses that want to operate in a standard way across all of their facilities.
And there are also, big picture, long-term CapEx businesses that are thinking about the operating theater, you know, 10 years from now, more than they are, what this current administration is doing. So I think for the most part. the current regulations, of course, they're sensitive to it. but I think it's less of a factor than what, I would've expected coming into the business.
that said, as anybody who's, focused on, reducing methane gas and improving our overall climate environment, you know, I wanna make sure that we're all focused on the right things and not being distracted by any sort of political shenanigans that are throwing us off the scent.
the answer is kind of yes and no. it does affect us, but I think in the big picture, the industry is moving in the right direction and cares deeply about solving this problem. And quite frankly, for a lot of our customers, this is product, right? They're selling this.
Deep: the, methane that gets captured.
Ryan: Yeah,
Deep: natural gas is,
Ryan: Yeah. And they don't
Deep: wanna be leaking 25% of their product. Exactly. Yeah, that makes sense. is there any topics that we didn't cover that you think maybe are important to like, help understand your perspective?
Ryan: The two things that we covered that are, critical are one, we believe we're the most precise management, both in terms of quantity and geolocation, business that's out there, We can detect this stuff down to a couple of feet and down to a couple of kilograms, of gas an hour, which is, many times better than.
Just about any of our competitors. we continue to focus on industry and we want to be picks and shovels for industry and keeping that lane. We believe there's room for other players to play in the other lanes and that's great. And we have 15 years worth of data that we've only really started digging into what kind of data intelligence we can expose, how we can leverage all of that information to try to inform both our business moving forward, but also create more products for our customers that, help accelerate the curve to solution.
So, yeah,
Deep: One of the critiques that, the fossil fuel industry's attempts to become more environmental, gets from, the environmental side, is this, are you solving the problem but enabling, more emissions net?
So, in other words, to the extent that. Like, you know, a lot of those folks, clearly would just wanna like to eliminate, methane production or oil and gas extraction. which, you know. I think most folks understand is not gonna happen overnight. And, it's probably a multi-decade, affair and in the interim there's an awful lot of emissions.
we will always have a need for a lot of these products. even if it's just to make fleece jackets, you know, Even if we move the entire, you know, global energy grid off of oil and gas, it's still gonna be a need for it.
But how do you think about that? Let's, for catching a sake, uh, you know, the emitters enablement problem, like, does, is that something that you know, is ever talked about or in your circles and, how do you think about that?
Ryan: I mean, only in that, again, we picked our lane and we were gonna stay in it.
Whether Bridger is working, from the inside to try to improve the cleanliness of the product, whether we're there or not, I don't think affects overall the global demand for these products. for me, we've picked this lane and I personally am impartial to it to try to help businesses, improve their strategies to producing as clean a product as possible.
Solving for plastics globally. That's a bigger conversation and not one that I think, we're gonna be able to influence much one way or another. solving even things like waste management, right? We will always have trash and so anything we can do to help that industry be more clean, be more, environmentally, focused, I think, is a good thing.
I think there are plenty of other businesses that are gonna play that sort of, oversight and, external pressure management. we've just chosen not to take that position Yeah,
Deep: no, I mean, it makes a lot of sense. And honestly, I think you guys are doing a lot of. good, if there's no efficient way to monitor these emissions, then they get emitted at greater volume. it's really fairly simple on some level. sure you might in theory have more companies extracting methane, but I feel like that's. Divorced from the issue of how well it's done, like how well the emissions are captured.
So,
Ryan: yeah, completely agree. and again, I am a huge fan of working with industry as their partner To try to solve this problem. That doesn't mean people can't be watched dogging externally, but I definitely, you know, appreciate that position for us.
Deep: Well, thanks so much for coming on the show. I learned a ton about a topic I knew next to nothing about before. So yeah. Thank you so much.
Ryan: Yeah, great. Well thank you. Appreciate it. love to do it again sometime.