- Satlyt raised $8 million seed funding to develop AI software for satellites in orbit
- Funding led by Non Sibi Ventures, with former astronaut Bernard Harris involved
- Satlyt builds software to run AI across satellites owned by various companies
Satlyt, a space technology startup developing software to run artificial intelligence workloads in orbit, has raised $8 million in seed funding as investors back the idea of turning satellites into distributed computing platforms.
The funding was led by Houston-based Non Sibi Ventures, with partner and former NASA astronaut Bernard Harris helping shape the firm's investment thesis.
Satlyt was co-founded by Rama Afullo, a former Google cloud computing and SpaceX Starlink product manager, and operates from Sunnyvale, California, and Nairobi.
The company is taking a different route from startups and major technology companies building dedicated orbital data centres. Instead of manufacturing spacecraft, Satlyt is developing software that can run AI models across satellites owned and operated by different companies.
TechCrunch reported that Satlyt plans to use the fresh capital to advance this software-driven approach as demand for computing capacity in space develops.
Also Read | 'Picture Abhi Baaki Hai': Vineeta Singh Reacts To Critics Slamming 80% Valuation Haircut
Satlyt has already tested its technology on two demonstration missions. Its software recently ran Google DeepMind's Gemma model aboard a Momentus-operated spacecraft. The system reduced the size of a transmission related to software errors by more than 60%, potentially lowering data-transfer costs for satellite operators.
The startup is also preparing for another mission aboard a spacecraft developed by Indian company TakeMe2Space. The mission will involve NASA, space surveillance startup Stellerian and TakeMe2Space, with each testing different computing applications.
One of Satlyt's key targets is reducing the amount of information satellites need to send to Earth. AI processing in orbit can help spacecraft analyse sensor data or address certain operational problems before transmitting information to ground stations.
The company eventually aims to connect computing capacity across separate satellites and create a shared cloud infrastructure in orbit. It expects to attempt that model next year.
Also Read | Why An Indian-Origin Engineer Left A Rs 2.89 Crore Meta Job To Build An AI Startup
Non Sibi Ventures sees the approach as less dependent on the economics and launch constraints surrounding large orbital data centres. Satlyt instead expects its opportunity to grow with the broader satellite market.
Afullo aims to have Satlyt's software running on 20% of satellites by the end of the decade, as more spacecraft adopt GPUs and other advanced processors.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.