Episode: Above the Line, “Orbital Computing: The $1 trillion case for AI data centers in orbit”
Guest: Brendan Burke, Research Director, Semiconductors & Emerging Tech (The Futurum Group)
Host: Shay Boloor, Chief Market Strategist, Futurum Equities
Episode Published: July 28, 2026
Listen: YouTube
The Take
Orbital compute is usually sold as a physics story with five times the solar yield of a panel on Earth, no interconnect queue, and unlimited land. Burke makes it a cost story. A watt in orbit prices near $75 against $17 on the ground, before compute — 4.5x more, and most of that gap is the rocket. But NVIDIA’s guidance for next-generation superpod deployments points toward $100 billion per gigawatt, or $100 per watt. The bar orbit has to clear is moving toward it. With $3 trillion going into AI data centers in 2030 and a third off-grid, the question is not space versus Earth. It is where the cheapest marginal gigawatt sits.
What You’ll Hear
- Why the orbital thesis breaks first on something that has nothing to do with rockets.
- The five-subsystem bill of materials and the layer the market underrates.
- What $75 a watt buys, and which line item eats most of it.
- The launch price is where production compute starts going up.
- The 90-day signal that tells Burke whether the thesis is on track.
The Insights
Space Doesn’t Have to Get Cheaper if Earth Keeps Getting More Expensive
Orbital deployment runs roughly $75 per watt against $17 terrestrially. That 4.5x gap looks fatal until read against NVIDIA’s $100 billion per gigawatt figure for the most complex next-generation builds. Off-grid capacity on Earth carries its own load — gas turbines, switchgear, power equipment already in shortage — and as builds absorb more custom engineering, the terrestrial baseline climbs toward the orbital number, not away from it. Hyperscalers are pricing three options against each other, and one gets more expensive every quarter.
“$75 per watt is significant. It does start to actually line up against the hundred billion per gigawatt, $100 per watt that Jensen has mentioned.” — Brendan Burke, Research Director, Semiconductors & Emerging Tech, The Futurum Group
Launch Is the Necessary Condition. It Is Not the Sufficient One.
Most of the $75 per watt is the rocket, making Starship both the biggest lever and the biggest dependency. Cut launch 10x and orbital deployment lands near $20–30 billion per gigawatt — close enough that redundancy and in-orbit maintenance decide it. But launch alone does not close the gap. Three curves must bend at once: launch from $1,400 to $200 per kilogram, satellite hardware from $22 to $11 per watt, and power density from 40 to 80 watts per kilogram. Miss one, and the model breaks.
“It is the necessary condition right now. It’s not sufficient on its own… but it is an assumption that the industry is making — that the historical rate of progress of launch costs will be continued.” — Brendan Burke, The Futurum Group
The Most Underappreciated Layer Isn’t Power. It’s the Laser Link.
Every orbital conversation starts with energy — five times the panel yield, none of the fossil-fuel gear, gating off-grid builds. Asked which subsystem the market underrates, Burke picks interconnect. On Earth, scale-across is throttled by the glass available for fiber and the work of tying data centers together. In orbit, laser links face far less interference, so cluster coherence can exceed terrestrial fabrics. That is the difference between a gigawatt data center and a terawatt-scale cluster running one job in sync.
“That allows us to scale up the coherence of a cluster in space even beyond what we can do on Earth… gigawatt data centers versus terawatt scale clusters that are able to all communicate with one another and carry out jobs in sync.” — Brendan Burke, The Futurum Group
The Big Picture
The risk here is not a failed test flight. Burke reads reuse economics as something that could surprise to the upside within five years, possibly two, and argues for following the direction of travel rather than re-underwriting after every launch. The real exposure is concentration. SpaceX is the unlock today, and is about to carry the competing priorities of public markets. If the rest of the stack does not commit — hyperscalers with a space roadmap, silicon hardened for orbit, thin-film arrays, and higher-efficiency cells — orbital compute stays a single-company program instead of a terawatt-scale market. Google is looking into it, and Amazon has the makings of a strategy. What Burke watches over the next 90 days is who else publishes a roadmap and signs a pilot customer.
Listen & Resources
Listen to the full conversation: YouTube
Read the Full Report: Orbital Computing Can Reach $1 Trillion Addressable Market by 2030
Mentioned in This Episode
- Launch and vertical integration: SpaceXAI — Starship, Falcon 9 reuse
- Silicon in orbit: NVIDIA — Jetson-class GPUs proven for five-year lunar operation; Feynman-generation superpods
- Hyperscaler strategies: Google, Amazon
- Supply chain to watch: thin-film solar arrays, liquid radiators, laser interconnect
Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.
Other Insights From Futurum:
SpaceX S-1 Pivots Toward Orbital Data Infrastructure
Will Starcloud’s Orbital Data Centers Solve NVIDIA’s Terrestrial Energy Crisis?
Can AMD’s Edge Silicon Scale to the Trillion Dollar Orbital Opportunity?
Author Information
Brendan is Research Director, Semiconductors, Supply Chain, and Emerging Tech. He advises clients on strategic initiatives and leads the Futurum Semiconductors Practice. He is an experienced tech industry analyst who has guided tech leaders in identifying market opportunities spanning edge processors, generative AI applications, and hyperscale data centers.
Before joining Futurum, Brendan consulted with global AI leaders and served as a Senior Analyst in Emerging Technology Research at PitchBook. At PitchBook, he developed market intelligence tools for AI, highlighted by one of the industry’s most comprehensive AI semiconductor market landscapes encompassing both public and private companies. He has advised Fortune 100 tech giants, growth-stage innovators, global investors, and leading market research firms. Before PitchBook, he led research teams in tech investment banking and market research.
Brendan is based in Seattle, Washington. He has a Bachelor of Arts Degree from Amherst College.
