⚠ Disclaimer: This entry may be incomplete, out of date, or inaccurate. It is AI-maintained on a best-effort basis. Do not rely on it as a sole source — verify claims independently using the sources listed below.

Summary

Kepler Communications is a Toronto-based startup building a space networking constellation. As of January 2026, it operates the largest orbital compute cluster currently in space: approximately 40 NVIDIA Jetson Orin edge AI processors distributed across 10 operational satellites, interconnected via laser inter-satellite links. Kepler is primarily a space networking company — its AI-at-the-edge capability is designed to intelligently route data across its constellation — rather than a pure orbital data center play. It is part of the NVIDIA Space Computing initiative alongside Starcloud, Aetherflux, and Planet Labs.

Key Facts

  • Founded: 2018 (approx.)
  • HQ: Toronto, Canada
  • Type: Company — space networking / orbital edge compute
  • Status: Active — 10+ satellites operational; constellation expanding
  • Largest orbital compute cluster: ~40 NVIDIA Jetson Orin processors across 10 satellites; launched Jan 2026
  • ISL technology: Laser inter-satellite links
  • Customers: 18+ confirmed as of early 2026; includes Sophia Space (orbital compute software testing)
  • Compute use case: AI-driven data routing and network management; edge inference on-orbit

What It Is / How It Works

Kepler is primarily building a next-generation space data network — a persistent, real-time connectivity layer in LEO that would replace store-and-forward satellite communications with always-on connectivity. The addition of NVIDIA Jetson Orin processors to its satellite constellation serves two purposes: enabling intelligent on-board data routing decisions (avoiding transmitting low-value data to ground), and opening the constellation to customer edge compute workloads in orbit.

In January 2026, Kepler launched a constellation update that gave it the largest orbital compute cluster publicly operating: ~40 Orin processors across 10 satellites, linked by laser ISLs. While the raw compute is modest compared to Starcloud’s H100-class GPU satellite, the distributed, laser-linked architecture is more directly analogous to a terrestrial data center’s distributed compute fabric — and represents a step toward the kind of inter-satellite ML cluster Google’s Suncatcher paper describes.

Sophia Space, one of Kepler’s named customers, is testing orbital computer software (presumably for its own planned compute constellation) aboard Kepler’s satellites.

Notable Developments

  • 2026-04: Described as operating the largest orbital compute cluster; 40 Orin processors / 10 sats with laser ISLs
  • 2026-01: Launched constellation update establishing the 40-Orin distributed cluster
  • 2025–2026: Named in NVIDIA Space Computing initiative; 18+ commercial customers

Key People

  • Key personnel not confirmed in public sources at the required detail level — TODO: research founders (Mina Mitry, CEO, is publicly named)
  • Mina Mitry — Co-founder and CEO

People — Last Reviewed: 2026-04-29

Sources