Syntiant and Novi Space Achieve Low-Power AI Inference in Orbit

Event summary

  • Syntiant and Novi Space demonstrated rapid AI model development and optimization for satellite-based computer vision in low Earth orbit (LEO) on March 26, 2026.
  • The test leveraged Syntiant’s quantized neural network models running on Novi’s satellite and SP240 space-edge computer, featuring an AMD Versal adaptive SoC.
  • The models successfully performed onboard detection of objects such as ground vehicles and ships, reducing latency and improving operational efficiency.
  • The models were trained and optimized using Syntiant’s proprietary development tools and could be retrained and redeployed in less than 24 hours.

The big picture

This demonstration highlights the growing trend of integrating AI into space-based systems to reduce latency and bandwidth requirements. Syntiant’s ability to optimize AI models for severe resource constraints positions it as a key player in the emerging market for intelligent edge AI in aerospace. The collaboration with Novi Space underscores the strategic importance of partnerships in advancing space technology and data analytics.

What we're watching

Scalability
How Syntiant’s low-power AI solutions will scale across other satellite and space-based applications.
Competitive Dynamics
Whether Syntiant can maintain its lead in low-power AI inference as competitors enter the space technology market.
Regulatory Compliance
The pace at which regulatory frameworks will adapt to the deployment of AI in space, potentially impacting operational efficiencies.