Authors: Elena Vrabie and Bogdan Iordache

“What big satellites weighing several tons have been doing for 30 years, we are doing with a 50 kg platform,” says Tristan Laurent, CEO of Absolut Sensing. The French startup is shrinking hyperspectral sensing to a size that can go to orbit on SpaceX’s rideshare program. The result is a next generation of climate-tech solutions that not only monitor methane leaks but also analyze them, turning high-resolution Earth observation into a commercially viable product.

Tristan Laurent has a background in finance, management, and business administration, with prior experience in IT and satellite projects at RHEA Group, ALTEN, and Open Cosmos. He co-founded Absolut Sensing with the mission of building technologies that allow human activity to preserve, rather than destroy, the planet.

Absolut Sensing develops miniaturized, cryogenically cooled hyperspectral instruments for Earth observation. These instruments combine optical systems with AI-driven data processing to measure and predict greenhouse gas emissions. As a startup within Groupe Absolut, it leverages decades of expertise in cryogenics and aerospace to efficiently scale its satellite platforms.

In this interview, Laurent talks about Absolut Sensing’s successful launch of GESat GEN1 aboard SpaceX, the technological innovations that enable compact hyperspectral monitoring, and the global need for accurate methane data. He covers AI-based forecasting, commercial partnerships, ESA’s Copernicus program, and policy drivers. 

Underline Ventures: In January this year, Absolut Sensing validated the performance of its hyperspectral sensors by launching the GESat GEN1 aboard a SpaceX Falcon 9 rocket. What did it mean for you and your team, and how do you see this achievement in the context of future missions? 

Tristan Laurent: For us, it meant going from a promise to a real company. We have now been offering commercial services since April. Our business model is based on selling data, and until you have a satellite, you are not generating any revenue. 

The launch was pivotal because it put us in the group of companies that have an operational asset. Some questions we had were: what the actual performance we could reach was, how good that performance was for the customer, and what the customer’s feedback was. We got that, and we are excited because the customer is super happy.

Now, we will make very little change to the technology. We are just going to produce many more of these satellites moving forward, and we have a clear plan ahead.

UV: Can you walk us through Absolut Sensing technology innovation – what is now possible when it comes to hyperspectral imaging, sensors, and cryogenic cooling?

TL: Two major things changed in the past five years that made what we are doing possible. One is on the software side, and it’s the appearance of AI, actual AI. The second is on the hardware side, the ability to miniaturize what we call the integrated detector cooler assembly (IDCA).

Think of your fridge at home and how bulky that is, and now you are trying to make that super small because the bigger it is, the more expensive it is to launch to space. You also want it to consume less power because you are power-constrained in space.

The hardware innovation was miniaturizing the cryogenic cooler (IDCA) and dividing the power consumption by more than two for the same amount of cooling, which in our case is about 140 Kelvin at the focal plane. That made it possible to do what big satellites weighing several tons have been doing for 30 years, but on a 50 kg platform. A one-ton satellite costs nine figures in CAPEX, and we didn’t have that kind of money.

On the software side, we generate about 1TB of data per day. You need to process and translate raw data into products the client will buy, such as methane emission maps, image sites, and leak locations. You also need to combine satellite data with weather and high-resolution mapping data, and sometimes ground sensor data. Then you have to find patterns. If you have this type of data that tells me this, then this will likely happen. 

We are not just monitoring; we are also forecasting methane and carbon emissions of particular assets, like a barrel of oil, and the equivalent of natural gas. All of this requires huge amounts of data crunching. That is why the team is made up of half hardware engineers and half data and computer scientists, and machine learning engineers.

UV: How does your modeling and machine learning system turn data into forecasts?

TL: We take existing integrated mass enhancement models, which translate a given concentration of methane into an emission rate. You go from X amount of methane in a column of atmosphere to that much emission in kilograms per hour.

Then we extrapolate that into the future and include meteorological parameters in a chaotic system. There is humidity, albedo (how reflective a surface is), solar elevation, and all sorts of things. We built what we call a physics-guided machine learning model with about 1,000 parameters. It’s a fairly simple AI model, and it allows us to improve the forecast on those emission profiles by a factor of 40.

Right now, our model works for oil and gas assets. Those assets are called point sources, like flares that burn. We are expanding to wide-area sources, such as landfills and other places that emit over a wide area.

UV: Why did you choose bare metal instead of cloud for your infrastructure?

TL: It would have been definitely easier to go with a cloud provider like AWS. We spent quite a bit of time early in our beginnings relying on cloud providers. But we are running a marathon, not a sprint, and having a cost structure that is scalable moving forward was important.

Cloud is ok when you are up to a certain amount of data that you have to process on a given day. But we ran an estimate of 100TB. Every business model that is below that threshold, I think, should totally go full cloud and not go into the hassles that come with dealing with their own architecture. Anything beyond that costs too much money. 

UV: What has been your experience building hardware in Europe?

TL: We are fortunate that in France, we have a strong talent pool for cryogenic engineers. The technology is made of three big things. There is the sensor, which in our case is a short-wave infrared sensor sensitive to photons in the 650-nanometer range. Then there is the optical system you put in front of the sensor, which allows you to look at what you want to look at. And the cryogenic element, which cools the focal plane, is the combination of the optical system plus the sensor.

The cooler a system is, the slower the atoms move, which creates less noise. What you want to optimize in satellite imagery is the signal-to-noise ratio, which impacts your performance. The noise is generated locally. If your optics are at 30 degrees, you generate a lot of noise, and the thermal stability is bad.

UV: You have collaborated with space industry partners such as SpaceX, Kongsberg NanoAvionics, and TotalEnergies. How have these collaborations helped accelerate hardware validation and scale EO for monitoring missions?

TL: Obviously, not having to build our own rockets is super good to scale. Another thing is that SpaceX is predictable. Back in the day, you would book a launch, and it would have a three to four-month delay. Now, with the transporter program, it is essentially a bus. It is rarely late, and it works. Launch is a binary event, so reducing that risk is critical for scaling.

For satellite platforms, there is still a craftsmanship problem. Every satellite platform is still kind of unique, and that slows scaling. We don’t build satellites; we build a camera that senses what is in the air, and we put it on a satellite. 

On the software side, we rely on OVH for data processing and storage. We run bare metal architecture; it’s not like with AWS, where you tell them what you want and then you don’t have to do anything. We have a team of six DevOps engineers handling the architecture. 

UV: Why is methane monitoring urgent, and how impactful is it for climate action?

TL: There’s this IPCC report every year that talks about how we are doing in terms of the climate-agreement goal, and the key element is that methane is the short-term solution to reduce global warming. Now it’s no longer a question of whether global warming is happening. It’s a question of how bad it’s going to be. 

Methane is 81 times more potent in its ability to trap heat. So if you take one molecule of CO₂ (carbon dioxide) and one molecule of methane, from a global-warming and heat-trapping perspective, one molecule of methane is far more impactful. There is a lot less methane emitted compared to CO₂, which is why we keep talking about CO₂ emissions, but we don’t know that much about methane’s potency. Methane also accounts for one-third of global warming worldwide. So although it’s less in absolute quantity than CO₂, which accounts for close to two-thirds, methane is still super important. 

When you look at how much humans are responsible for methane emissions versus CO₂ emissions, we actually have much better leverage on methane because it’s something we can easily fix. A lot of it comes from the oil and gas industry, and it comes from aging infrastructure – flares, vents, pipelines, unsealed wells, and things that don’t really need a lot of CAPEX from an operator to improve. 

Starting the company was a combination of the market opportunity. Now, there is a forcing function for companies to decrease methane emissions, as they are being taxed on the amount of methane they release into the atmosphere, along with the economic benefits. For an oil and gas operator, they want to sell as much natural gas as possible, and any methane leak is natural gas going away. It’s estimated to be about 1.7% of revenue loss every year, which is a huge amount when you look at the monetary value.

People should care about methane both for the global-warming potential and because, from an economic perspective, it is not like you have to trade off a clean solution that is less energy-efficient or that has lower economic output. It’s just a matter of showing people where the issue is and how impactful the solution is.

UV: How challenging is detecting methane at scale?

TL: It is a global problem across more than 110 countries that have oil and gas infrastructure. Many emissions happen in geopolitically sensitive areas where you can’t send people or drones, which is preferred, because ground data would be more accurate. So satellites that are flying over 550 km above our heads are the only way to know what’s happening.

Next comes the technical challenge. It’s a chaotic system. It’s not the same as taking a picture of a parking lot and counting cars. We are measuring a concentration of gas in the air, which is impacted by lots of parameters – temperature, turbulence of wind directions, and humidity. You have to model this, and you can’t do it without AI. The company’s next phase is producing more satellites and improving our models to get even better at solving this.

UV: How will space-based sensing change in the coming years?

TL: Traditionally, satellites (from NASA, ESA, JAXA) are very good at modeling large systems, like the weather forecast. Beyond 72 hours, weather forecasts lose precision because spatial resolution isn’t good enough. The reason is that those satellites have a wide swath and a 1-10 km spatial resolution.

Commercial satellites have higher resolution and now complement those big satellites by zooming into very small areas. In our case, we are looking at individual valves on drilling sites, if a leak is coming from a particular valve. 

Moving forward, big satellites will continue to model large systems, and commercial satellites will zoom in locally, focusing on specific points. It will be complementary. And as far as the industry overall, the remote sensing industry is still about 80% driven by defence and will continue like so. Climate is still a small chunk, but things are picking up. 

During the Biden administration Inflation Reduction Act had two components: methane taxation above a threshold and how companies needed to report methane emissions, and financial incentives for operators that were going in the right direction in various forms. 

When Trump came in, the tax was removed, but incentives increased due to the oil and gas industry. What is super exciting is that we are also seeing similar momentum in Europe with the EU Methane Law, which was proposed 18 months ago, and this is going to move the market up. 

UV: Can you describe your current go-to-market strategy? How does your sales process differ from commercial to governmental?

TL: We sell to everyone who is not on the sanctions list. We are three years old, and not many big oil and gas companies know about us. The market is fragmented and mostly driven by small independent operators. 

We have two customer bases. One is the operators, and the second is the government. Operators ask us to validate their inventories that they then send to governments. Governments look at all sites and validate the amount of methane being emitted. We play on both sides.

Our business model is that we sell through a direct approach, via our web platform called PLUM, where customers can go from product discovery to buying in four clicks. And we also have 12 resellers covering the U.S., the Middle East, Asia Pacific, and Europe. We go through distributors because our sales team is one person: me. We don’t need more because the platform handles it. 

We are in a niche where the sweet spot is about 50-50% between commercial and government. Right now, we are more governmental than commercial, but next year we aim to be at a balance because what we sell has an immediate economic value to the customer – they can sell more by having fewer leaks. 

The sale process between the commercial is different from that government, and the civil government is different from that defence. In my small experience, defence customers take years to acquire; they will give you a small pilot first, and require top security and quality assurance to build trust. Those processes slow commercial growth. 

Our current architecture through OVH, if we were to address a defence customer, would not work; we would need to redo our entire infrastructure, and that would not be good from a scalability perspective. The right way to do both, like SpaceX and Airbus are doing, is to spin off defence into a separate organization.

UV: What can you tell us about the ESA’s Copernicus contract?

TL: In the beginning, our strategy was to fly under the radar. You would not see a lot of communication about Absolut Sensing for the first three years. We said we would focus on delivering and let our customers talk about how serious we are and how good the products are. 

The Copernicus and ESA teams are amazing. They trusted us very early when we were 12 people, and the satellite had not been launched yet. They helped us with data quality and were an amazing customer. A best practice is not to put a satellite in space unless it is paid for. The $5M Copernicus contract allowed us to commit to launch. Commercial customers are not going to give you money until there is a product to buy. Now we are moving from a data demonstration contract to an operational contract with Copernicus. 

UV: What’s the GESat constellation architecture you are planning for 2027?

TL: We were selected by the EIC accelerator for €17.5M in blended financing, with €15 million in equity, and we are fundraising on the back of that. The overall plan is a €30 million round. The next satellite launch is in 2027. The initial batch is of three satellites, but the full constellation will be of 12. Three satellites are enough to cover all the oil and gas and the Copernicus needs.

We are also expanding into other verticals, like flooding, working with insurance companies, because these events are going to be more common and are going to create losses. So the ability to predict these at the local (city or area) scale, not the country level, is going to be important. We will be using the same tech stack on the focal plane and one additional hardware element: a linear variable filter, something that you put in front of your sensor that allows you to select precisely what you want to look at. 

UV: From your experience building and launching satellites, what should aspiring space entrepreneurs know?

TL: It’s stressful. It’s intense. Aspiring entrepreneurs might not want to do it after hearing what I have to say. In the first two years, we were working nonstop, Monday through Sunday. We actually have the record in Europe for launching the fastest first satellite from company creation to launch. It was three years, and we did it with a team of 12.

One advice I can give is to definitely hire the best people and map out, from the get-go, the critical skills you need, both technical and program management-wise. There are many administrative projects you need to tackle, such as obtaining insurance and licenses.