What is Earth observation data worth per square kilometre and per revisit, who pays for it, and does artificial intelligence change who captures the value?
The short answer. satellite imagery is getting cheap, and the money is moving elsewhere. Archive imagery from Planet's SkySat fleet lists at $6 per square kilometre and Satellogic's at $4, against $17–26 for 30 cm archive imagery from resellers; constellations now collect far more than they sell. Value sits in two places. The first is guaranteed access: assured tasking at $40 per square kilometre and above, and sovereign deals in which a government buys capacity or whole satellites, the fastest-growing line for Planet, BlackSky and ICEYE. The second is answers rather than pixels: change detection, monitoring and alerts sold per asset or per decision. Buyers are still mostly governments, but the growth has moved from US intelligence, where the NRO's commercial imagery budget was cut in the 2026 request, to European and Asian defence ministries. Radar (SAR) is the growth engine: ICEYE reported more than €250 million of revenue in 2025 and raised money at a valuation above €10 billion in June 2026. Free data from Sentinel and Landsat and open AI foundation models lower the cost of analytics, which shifts margin to whoever owns the customer's decision.
Ten findings
| # | Finding | Chapter |
|---|---|---|
| 1 | Imagery prices are falling at the low end: SkySat archive $6/km², Satellogic archive $4/km²; constellations sell a small share of what they collect | Chapter 11, Chapter 50 |
| 2 | Guaranteed access is the premium product: assured tasking costs several times archive, and sovereign capacity deals are the largest new contracts | Chapter 11, Chapter 62 |
| 3 | Governments buy most EO; the growth has moved from US intelligence to European and Asian defence | Chapter 17, Chapter 19 |
| 4 | SAR is the growth engine: ICEYE >€250M revenue (2025), €1.5B backlog, valuation >€10B (June 2026) | Chapter 44 |
| 5 | Planet reached record revenue of $116.1M in one quarter (+58%) and guides to $430–441M for its fiscal 2027; backlog ~$815M | Chapter 38 |
| 6 | Open data is a hard price floor: Sentinel-2 at 10 m and Landsat at 30 m are free, so paid data must beat them on resolution, revisit or latency | Chapter 13 |
| 7 | AI foundation models (AlphaEarth, Prithvi, Clay) make analytics cheaper to build; the scarce asset becomes labelled outcomes and customer workflow | Chapter 52, Chapter 54 |
| 8 | On Setcoin's model a 30-satellite high-resolution constellation needs to sell about a fifth of its capacity at ~$23/km² just to cover its costs | Chapter 50 |
| 9 | Europe is building sovereign EO (Germany's ~€1.7B SAR constellation with ICEYE, Italy's IRIDE, the EU's EOGS studies) | Chapter 21 |
| 10 | The market is $4–8B a year depending on definition; a disciplined estimate of satellite data plus first-level analytics is ~$5–6B in 2025 | Chapter 15 |
Exhibit 1.2
Guaranteed access, not resolution alone, sets the top of the ladder
The price ladder: list price per km², 2026 · $ per km²
Key numbers
| Measure | Value | Source |
|---|---|---|
| Planet revenue, Q2 FY2027 | $116.1M (+58%) | Planet, Sep 2026 |
| BlackSky revenue, Q2 2026 | $33.3M (+50%) | BlackSky, Aug 2026 |
| ICEYE revenue, 2025 | >€250M (unaudited) | ICEYE, Mar 2026 |
| SkySat archive price | $6/km² | Planet price list, Sep 2026 (as reported) |
| Satellogic archive price | $4/km² | Satellogic, as reported |
| NRO EOCL programme | ~$4B over ten years | Reporting, 2025 |
| ICEYE satellites in orbit | 57 | Setcoin catalogue analysis, 20 Oct 2026 |
| Planet Dove family in orbit | 115 (660 launched since 2013) | Setcoin catalogue analysis |
What changed in 2026
Three developments in 2026 shaped this study. First, sovereign buying became the main source of growth. Germany's SAR constellation with Rheinmetall and ICEYE entered production, Sweden and Poland expanded their ICEYE systems, and Planet, BlackSky and Satellogic all reported that dedicated capacity and satellites for governments drove their revenue. The March 2025 suspension of Ukraine's access to US-licensed imagery, though it lasted only a week, became the reference point for governments explaining why they want control.
Second, the listed operators turned the corner on profitability. Planet, BlackSky and Satellogic all reported positive adjusted EBITDA in 2026 after years of losses, and ICEYE reported EBITDA above €100 million for 2025. The industry moved from raising money to fund satellites to funding them from contracts and cash.
Third, AI changed the economics of analytics. Foundation models such as Google's AlphaEarth, NASA and IBM's Prithvi and Clay made it cheaper to turn imagery into answers, and Prithvi ran in orbit for the first time. The scarce assets are now labelled history, accuracy and integration into customers' decisions, which favours operators and analytics firms that own those relationships.
Against these, two pressures grew. the price of archive and flexible tasking imagery kept falling as fleets expanded, and US intelligence signalled lower commercial imagery budgets. The companies that grew in 2026 were those that turned satellites into contracts and data into decisions.
The interactive models
| Model | Question it answers | Chapter |
|---|---|---|
| Revisit calculator | How often can a constellation see a point on Earth? | Chapter 9 |
| Constellation cost per km² | What must a constellation charge per square kilometre to cover its costs? | Chapter 50 |
| Analytics value chain | How much of an analytics subscription goes to the imagery supplier? | Chapter 52 |
| Buyer's data cost | What does monitoring an area cost a year, by resolution and revisit? | Chapter 65 |
| Government share stress test | What happens to an EO company's revenue if US intelligence buys less? | Chapter 67 |
How the study is organised
| Part | Chapters | Content |
|---|---|---|
| I. What EO is and what it costs | 2–15 | Method, sensors, the value axes, the fleet, revisit, prices, open data, market size |
| II. Who buys | 17–36 | US intelligence, defence elsewhere, Europe, Asia, agriculture, insurance and finance, energy and infrastructure, climate, maritime, Ukraine |
| III. Who sells | 38–48 | Planet, BlackSky, Vantor, ICEYE, other SAR, other optical |
| IV. Economics and AI | 50–67 | Constellation cost, the value chain, foundation models, edge AI, platforms, pricing models, sovereign deals, buyer's cost, government stress test |
| V. Rules, supply and capital | 69–84 | Regulation, ground segment, launch and parts, consolidation, capital, hyperspectral and thermal, very high resolution, RF sensing |
| VI. Synthesis | 86–98 | Scenarios, implications, checklist, risks, signals, questions, method |
Reading paths
| Reader | Suggested path |
|---|---|
| Investor or lender in an EO company | 1, 11, 38–44, 50, 67, 90 |
| Corporate buyer of imagery or analytics | 1, 5, 11, 13, 25–31, 65 |
| Defence and government buyer | 1, 17–23, 36, 62, 69 |
| AI and analytics company | 1, 52–60 |
What this study does not do
It does not value companies or recommend securities. Company figures are as reported; Setcoin calculations show their arithmetic.