AI chips are no longer just training models in giant data centers on Earth but are also going to the Moon, analyzing images from lunar orbit before beaming back only the insights that matter. That’s exactly what Nvidia (NASDAQ:NVDA | NVDA Price Prediction) is doing.
While earthbound hyperscalers still drive the bulk of Nvidia’s growth, two announcements this week — one from Planet Labs (NYSE:PL) and one from Firefly Aerospace (NASDAQ:FLY) — show the chipmaker is extending its edge AI platform into space. That positions the stock for a large slice of the rapidly growing space economy, which McKinsey projects will expand from $630 billion in 2023 to $1.8 trillion by 2035.
Nvidia’s Jetson Platform Proves Itself in Orbit
Let’s start with the hardware that makes this possible. Nvidia’s Jetson Orin module delivers high-performance AI inference in a compact, power-efficient package built for size-, weight- and power-constrained environments. Planet Labs announced on April 7 that its Pelican-4 satellite successfully ran object-detection AI onboard using the Jetson Orin. The satellite captured an airport image over Australia and identified aircraft in moments. Raw data stayed in orbit; only compact results headed to the ground. This cuts downlink costs and latency for Earth-observation customers.
This morning, Firefly Aerospace revealed a collaboration with Nvidia to embed the same Jetson platform into its Elytra spacecraft for the Ocula lunar imaging service. Firefly’s announcement states the setup will process high-resolution lunar imagery in orbit during the Blue Ghost Mission 2, scheduled for late 2026. Elytra stays in lunar orbit for five years, enabling repeated mapping and change detection. In both cases, Nvidia’s software stack — built on CUDA — runs the AI models developed by each partner. No custom silicon required.
Edge Computing Solves Space’s Biggest Bottleneck
That brings us to why this matters for edge computing in space. Traditional satellites capture terabytes of data but face 2.6-second round-trip delays to the Moon and narrow bandwidth windows. Downlinking everything eats power and storage. On-orbit processing flips the script: capture, analyze, and downlink only actionable insights. Nvidia noted last month that its Space-1 Vera Rubin Module will deliver up to 25 times the AI compute of an H100 GPU, powered by solar energy for orbital data centers.
Although Advanced Micro Devices (NASDAQ:AMD) offers competing accelerators and an ecosystem for terrestrial data systems, Nvidia’s chips and CUDA ecosystem already dominate AI development here. It suggests developers building satellite software will stay locked in when they want to head to the Moon and beyond.
McKinsey’s space economy report breaks the $1.8 trillion space total addressable market into “backbone” services — satellites, launches, and core infrastructure — projected at $755 billion by 2035, plus “reach” applications that space tech enables across industries, adding another $1.035 trillion. Nvidia’s Jetson and Vera Rubin platforms slot directly into both layers: faster geospatial intelligence for defense and logistics, plus foundational compute for future orbital clouds.
Granted, space revenue today remains negligible — well under 1% of Nvidia’s $215.9 billion fiscal 2026 total. Radiation hardening and regulatory hurdles still exist. That said, these live demos on Pelican-4 and Elytra de-risk the technology faster than any lab test could.
Key Takeaway
In short, Nvidia is no longer just the data-center GPU leader. The Planet Labs and Firefly wins validate its edge AI platform as the default for space, opening a durable adjacency within the $1.8 trillion opportunity.
Fiscal 2026 free cash flow reached $96.6 billion, giving Nvidia ample room to invest without diluting shareholders. Its forward P/E stands at 16.39, a deep discount for a company with 65%revenue growth. Smart investors should treat this as an expansion of Nvidia’s core moat, not a replacement. The numbers show its orbital bet builds on proven strengths rather than betting the farm on the final frontier.