Orbital Stack Framework: Where the Space Economy's real commercial value is building

In November 2025, a 60-kilogram satellite carrying a single NVIDIA H100 GPU launched into orbit on a SpaceX rideshare mission. Within days it did something no machine had ever done before: it trained a large language model in space. That satellite belonged to Starcloud, a startup founded in 2024 with backing from NVIDIA. Crusoe, one of the fastest-growing AI infrastructure companies in the US, signed on to become the first public cloud provider to run workloads in space by early 2027.

Many found this as an "interesting novelty” whereas it should be tagged as "strategic inflection."

The space economy is not a launch story. At this time it is an infrastructure story. What gets built on that infrastructure and who controls it is the actual strategic question. And right now, that question is being answered at a pace the mainstream ecosystem hasn't caught up with.

The infrastructure layer is being built in real time

The next decade will likely see the foundation of a true off-world economy. Advances in reusable rocket technology, satellite servicing, and in-space manufacturing are creating entirely new business models with the emerging cislunar economy spanning low Earth orbit to the Moon presenting key opportunities in infrastructure development, satellite servicing, and resource extraction.

LEO communications constellations are projected to capture 50% of global broadband revenue by 2030, shifting $200 billion in TAM from terrestrial networks. Constellation capex exceeds $50 billion annually through 2028.But the more consequential shift is happening above the connectivity layer. On January 30, 2026, SpaceX filed an FCC application for what it calls the "SpaceX Orbital Data Center System" with up to one million satellites operating at 500–2,000 km altitude. For context, there are roughly 14,000 active satellites in orbit today from all operators combined. The filing stated: "Orbital data centers are the most efficient way to meet the accelerating demand for AI computing power."

Eric Schmidt acquired Relativity Space in March 2025 for approximately $800M, becoming CEO with a controlling stake. He confirmed the orbital data center thesis explicitly testifying before Congress that data centers could go from consuming 3% of US electricity to 99% without space-based alternatives.

Introducing the Orbital Stack Framework

The founders and investors who will capture disproportionate value in the space economy are those who think in layers rather than sectors. We propose a four layer Orbital Stack Framework at which durable commercial value is being built above the Earth:

Layer 1- Launch & Access: 

The commodity layer. Reusable launch vehicles have cut per-launch costs by over 80% versus early 2000s levels. The commercial space launch market accelerates from $10.8 billion in 2026 to $18.6 billion by 2030 but this layer is commoditising fast, with new entrants from Asia and private firms rapidly narrowing the dominance of early leaders. Launch is a prerequisite, not a moat.

Layer 2-  Constellation Infrastructure:

 The connectivity and observation layer. With 150+ commercial Earth observation constellations in development globally, the decade ahead will be defined by interoperability, analytics-driven delivery, and the institutionalisation of EO as a mainstream geospatial data layer. Planet Labs secured a $260M contract from Germany in 2025 for satellite services supporting European defence and security. BlackSky's Gen-3 constellation delivers 35-centimetre imagery with AI-driven analytics for defence and intelligence customers recording revenue of $107 million with a backlog up 32% to $345 million. This layer is maturing rapidly and generating real commercial revenue today.

Layer 3-  Orbital Compute & AI Infrastructure:

 This is the emerging frontier. At GTC 2026, Jensen Huang announced the NVIDIA Space-1 Vera Rubin Module delivering 25x more AI compute for space-based inference versus the H100. Starcloud's next satellite launches in October 2026 with NVIDIA Blackwell platform integration, 100x the power generation of its first satellite, and projected unit economics that generate more cash than the satellite costs to build and launch. For Earth observation applications, processing data in orbit before transmission reduces bandwidth requirements by 90–95%, only relevant insights need to be sent to ground stations rather than raw imagery. The compute layer is where the next platform battle is being fought.

Layer 4- In-Orbit Services & Manufacturing:

 The long horizon. In-orbit servicing and manufacturing will extend satellite lifespans by 50%, adding $100 billion to the space economy by 2035. OrbitFab is building the first satellite refuelling infrastructure in space. Astroscale is developing debris removal and satellite life extension. Vast is developing Haven-1, the world's first commercial space station, operational by 2027–2028. Companies operating here are building the maintenance and manufacturing layer of a permanent orbital economy.

The sovereign dimension 

The orbital infrastructure race is not purely commercial. It is geopolitical and the two are inseparable.

China has launched experimental AI satellites as part of an orbital supercomputer initiative, deploying three test satellites in 2024 equipped with domestic AI accelerators, with plans for a 50-satellite constellation by 2028. China has proposed a 200,000-satellite constellation focused on data sovereignty and in-orbit processing.

Europe's ASCEND initiative, funded with €300 million through 2027, focuses specifically on European energy independence and data sovereignty through space-based infrastructure with a demonstration mission planned for 2026.

Rocket Lab landed an $805 million contract in December 2025 to deliver 18 missile warning and tracking satellites for the Space Development Agency the company's largest deal yet, nearly 50% larger than its entire 2024 revenue.

Governments are not funding space infrastructure out of scientific curiosity. They are funding it because whoever controls the orbital compute, communications, and observation layer controls strategic leverage on Earth over communications, defence, financial systems, and climate intelligence simultaneously.

For founders building in this space, sovereign contracts are not just a nice-to-have, they are a structural component of the business model in every layer above launch.

Where does the founder and investor opportunity sits

The mistake most ecosystem participants make when evaluating the space economy is anchoring on the headline - Rockets, Mars, Tourism, rather than the commercial logic building beneath it.

Earth Observation revenue is transitioning from data sales to high-value analytics and software-as-a-service. Satellite-as-a-Service business models are rising, supporting real-time insights and secure mission-driven data fusion. The value is migrating up the stack exactly as it has in every previous infrastructure cycle.

The founders worth backing in this environment are not the ones building the next launch vehicle. They are the ones building the application, analytics, and services layer that orbital infrastructure enables. Space-based climate intelligence. Orbital AI inference for defence applications. Satellite-derived supply chain monitoring. Hyperspectral imaging analytics for precision agriculture. In-orbit pharmaceutical manufacturing.

Spiral Blue has launched 10 NVIDIA computers into orbit onboard customer and partner missions, developing edge computing for space. Pixxel is building a constellation of hyperspectral Earth imaging satellites to detect and monitor environmental changes invisible to traditional cameras. These are not moonshot projects. They are early-stage commercial businesses building on a maturing infrastructure layer.

"The space economy's most valuable companies of the next decade will not be the ones that build the best rockets. They will be the ones that understand that the rocket is only the road."

What matters is what you build on the other side.

References : AdSkate StartUs Insights Tracxn MIT Technology Review United States Artificial Intelligence Institute IBM Wolters Kluwer 

Connect with us