The most expensive thing you can do with a satellite is talk to it. For an Earth observation satellite snapping pictures, that means every pixel it beams down to a ground station costs money, time, and precious bandwidth. The obvious fix is to not send the pixels at all. Send the answer instead. That is the quiet, hardware-laden bet being made by Spiral Blue, a Sydney-based company that has spent the last six years figuring out how to put a server rack in a shoebox and launch it into orbit.
Their product, the Space Edge Computer, is exactly what it sounds like: a hardened computer that slots into a small satellite and runs AI models on the imagery as it is captured. The company claims this can increase a satellite's effective data capacity by as much as 20x [NewSpace Index]. Instead of downlinking a terabyte of raw forest imagery, the satellite could send a few megabytes of processed data pinpointing areas of deforestation or fire risk.
A hardware wedge in a software-defined sky
The space industry is increasingly software-defined, with constellations managed from the cloud and data pipelines that stretch from orbit to AWS. Spiral Blue's wedge is a deliberate step back from that abstraction, into the physical reality of radiation, vacuum, and launch vibrations. Their flagship unit, the SE-1, is built around an NVIDIA Jetson module, a common choice for edge AI on Earth but a rare sight in space. The company touts it as the first Australian space edge computer and the first NVIDIA Xavier NX to go operational in orbit [Spiral Blue blog].
This is not just an engineering flex; it's a commercial strategy. By building on a familiar, high-performance computing architecture, they lower the barrier for customers who want to run their own code up there. Their 'Your Code In Space' (YCIS) service promises companies can gain 'space heritage', a crucial credential for defense and institutional buyers, in a fraction of the traditional time and cost [SPACE & DEFENSE, 2026].
Traction measured in orbits, not just dollars
Spiral Blue's progress is best tracked not on a cap table, but on a launch manifest. The company has designed, built, and launched at least nine of its Space Edge Computers, with three more scheduled [Satsearch]. These aren't just prototypes; they're flying on missions with partners like Satellogic and SatRevolution, processing real data. Their funding story reflects a deeptech grind more than a venture sprint. They've secured a $2.1 million seed round led by Amplify Partners [Crunchbase, 2026], but their early lifeblood has been targeted contracts and grants. These include a $640,000 contract from the Australian Defence Innovation Hub to develop 'Vessel Detect' for maritime surveillance and a grant from the Australian Space Agency [Microsoft News].
The crowded orbital edge
Spiral Blue is not alone in seeing the value of off-planet compute. The competitive field is a mix of pure-play space compute firms, cloud giants, and defense contractors.
| Competitor | Primary Focus | Key Differentiator |
|---|---|---|
| Spiral Blue | Space Edge Computers | Hardware + software stack for in-orbit AI on NVIDIA Jetson |
| Loft Orbital | Satellite-as-a-Service | Manages entire satellite bus and mission for customers |
| Satellogic | Earth Observation Constellation | Vertically integrated, from satellites to analytics platform |
| AWS/Azure | Cloud Ground Stations & Analytics | Dominant earth-side data pipelines and AI services |
Spiral Blue's answer is to stay focused on the compute layer as a horizontal enabler. A partnership with Microsoft's Space Startup Launchpad integrates their Cobalt imagery platform with Azure, letting customers spend cloud credits on satellite tasking [Microsoft Australia News Centre, 2026].
The unit economics of not downloading
The financial logic of edge computing in space is brutally simple. Downlink bandwidth is a finite and expensive resource. Processing data in orbit condenses it, sometimes dramatically. If a satellite can turn 1 terabyte of imagery into 50 gigabytes of actionable alerts, it has effectively created 950 gigabytes of virtual bandwidth. That's capacity that doesn't need a bigger antenna, a more powerful transmitter, or extra ground station time. For Spiral Blue to graduate from a promising hardware vendor to a foundational space infrastructure company, it must out-execute incumbents in making its platform the default choice for new satellite builds.