Industry Newsllmorbital computeh100starcloud
Starcloud Successfully Trains LLMs From Orbit
9.0
Starcloud, an NVIDIA-backed startup, said its Starcloud-1 satellite launched last month used an NVIDIA H100 to train Andrej Karpathy’s nano-GPT on Shakespeare and run inference on Google DeepMind’s open Gemma model. The company, founded in 2024, says this first orbital LLM work supports its plan for a 5-gigawatt solar-powered space data centre and highlights potential relief for Earth-based energy and cooling constraints.
Key Points
- 1Trains LLMs in orbit using NVIDIA H100, trained nano‑GPT and ran DeepMind's Gemma
- 2Demonstrates feasibility of orbital compute to ease Earth-based energy, cooling, and water constraints
- 3Encourages engineers to plan satellite-linked ML workflows and evaluate distributed orbital compute architectures
Scoring Rationale
High novelty and industry scope, but limited immediate operational details and long-term feasibility uncertainties remain.
Sources
Public references used for this report.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
High-Value Overnight OrdersEasyDelivered International ShipmentsMediumOn-Time Delivery Rate by CarrierHard
250 free problems · No credit card
See all Logistics & Shipping problems
