Photo: EDGX

Europe’s space industry is undergoing a renaissance right now. With governments looking to spend more money on space and defense, this investment is starting to reverberate across the supply chain. One company to keep an eye on this area is EDGX, a new Belgian AI space startup, that raised 2.3 million euro ($2.7 million) in funding in August last year. This funding will be used accelerate the commercialization of EDGX Sterna, a next-generation edge AI computer for satellites.

The company has also closed a multi-unit deal with a satellite operator worth 1.1 million euro ($1.3 million) and expects even more deals this year. Nick Destrycker, co-Founder and CEO of EDGX told Space Security Sentinel, the company expects to announce multiple new customer deals during the course of the year. “Our focus right now is on execution and preparing the ground for future opportunities,” he said.

Satellite operators are one of the key target customers for edge AI computers for satellites, but not the only ones. The company will focus on the entire supply chain of large constellation projects, which includes satellite operators, manufacturers, and payload developers. In addition, it also wants to serve startups and universities that have ambitious projects and exciting use cases, because some will grow into larger satellite missions in the future.

The Belgian company also hopes to can do some business further afield even if the initial target will be Europe. Europe is its anchor market and the natural starting point for EDGX, but the company is already looking into the U.S. market. “Beyond that, we are actively looking at other high-growth regions where constellations are expanding and space initiatives are taking shape,” adds Destrycker.

With a lot of focus on defense markets in Europe, this could be a key target market for the company.

Destrycker says that defense missions like intelligence, surveillance, and reconnaissance (ISR), signals intelligence (SIGINT), and space situational awareness (SSA) are a strong fit for the company’s Sterna edge computing platform, because these require the highest levels of onboard processing performance. “In these cases, the ability to generate actionable insights directly in orbit, without delays, is critical for the end users such as DoD’s. We already see potential customers in this segment approaching us with interest,” comments Destrycker.

He says what makes Sterna different is that it is both scalable and developer-friendly, with strong documentation and openness that make it easy to integrate and use. It sits natively in the Nvidia ecosystem, which means customers can port existing AI models without heavy adaptation, accelerating adoption.

“Sterna is built for space from the ground up, with radiation mitigation software and proven robustness that operators can trust in orbit. In practice, it brings the power of Earth-class AI processing into a space-qualified envelope. There are no compromises: from the developer’s side, everything feels familiar and well-documented, and from the space side, it has the reliability and resilience needed for missions. This balance of performance, openness, and space readiness is what sets Sterna apart,” Destrycker adds.

There has been a lot of talk recently about how the worlds of AI and space could merge together. Over the next two years, Destrycker says he expects to see the first real ‘servers in space’ come online, with satellites capable of running multiple applications in parallel. He believes that defense and institutional customers will drive this adoption first, but commercial constellations will follow quickly because AI will be necessary to remain competitive on cost and performance. “AI-enabled onboard processing will move from being experimental to becoming a standard part of satellite missions,” he adds.

Destrycker believes that AI is going to fundamentally transform the space industry.

“Today, most satellites still send large amounts of raw data to Earth for processing, which is costly and slow. In the near future, satellites will function as in-orbit data centers, running multiple AI models simultaneously and generating insights directly on board. This will reduce latency, save bandwidth, and enable entirely new use cases. We believe that outdated ground-centric architectures will start being replaced within this decade by AI-enabled, flexible platforms,” he says.

In terms of what would make it a successful year for the company, Destrycker believes its first successful in-orbit demonstration campaign will be number one on the list. Secondly, it hopes for the closure of multiple new customer deals of a similar size to its first one. “Those two milestones will confirm both the technical strength of our product and the commercial demand behind it,” he says.