Earth Observation in 2026: Planet, Maxar, BlackSky, Capella, ICEYE
Planet, Maxar (Vivace), BlackSky, Capella, ICEYE, Synthetaic, Orbital Sidekick — the 2026 earth observation business. Imagery-as-a-service vs raw imagery, on-satellite AI, downlink reduction.
Earth observation in 2026 is a real business, not a pre-revenue research category. Planet’s daily-revisit constellation, Maxar’s high-resolution legacy (now owned by Vivace), BlackSky’s tasking-as-a-service, Capella’s synthetic aperture radar, ICEYE’s SAR constellation, and the analytics-and-applications layer on top all add up to a market that’s spending less time selling imagery and more time selling decisions.
The economic shift that defines the 2026 business: customers don’t want pixels, they want answers. EO-as-a-service, on-satellite AI processing, and downlink reduction are the structural changes driving the next phase.
Planet: daily revisit, the volume play#
Planet operates the largest commercial earth observation constellation: roughly 200 PlanetScope Dove satellites plus the SkySat tasking constellation. The PlanetScope mission images the entire Earth landmass daily at 3-5 metre resolution, with SkySat providing on-demand higher-resolution (50cm) tasking.
Planet’s customers in 2026 span agriculture, defence, finance, energy, infrastructure monitoring, insurance, environmental monitoring, and government. The recurring-revenue model — annual access to the imagery archive plus tasking credits — is the structural advantage over one-off pixel sales.
The strategic shift through 2024-2025 has been moving up the value stack. Planet Insights Platform packages analytics rather than raw pixels — change detection, road and building extraction, crop classification, deforestation alerts, methane plume detection. The pixels are increasingly the input to an analytic product, not the product itself.
Maxar / Vivace: the high-resolution incumbent#
Maxar Intelligence (the Maxar Technologies brand) operated the WorldView constellation — the highest-resolution commercial imagery available (30cm-class on WorldView Legion, less on older satellites). In 2023 Advent International took Maxar private, and in 2025 the business was acquired into Vivace, a defence-tech platform combining several intelligence-and-imagery assets.
Maxar’s customer base is dominated by defence and intelligence. The Legion constellation, deployed through 2024-2025, is the most capable commercial high-resolution imagery system. For applications where 50cm-class imagery is required, Maxar remains the default.
The 2026 question for the Vivace-owned Maxar is whether the private-equity ownership accelerates the analytics-and-applications strategy or whether the cash-flow optimisation of the defence business takes priority. Both are coherent strategies.
BlackSky: tasking and rapid revisit#
BlackSky operates a smaller high-resolution constellation (~20 satellites) optimised for rapid revisit and on-demand tasking. Where Planet aims for daily coverage and Maxar for highest resolution, BlackSky aims for “image this specific place every 90 minutes.”
The BlackSky model is tasking-first: customers (defence, financial intelligence, supply chain monitoring) order imagery of specific locations on specific schedules. The Spectra platform pairs that with analytics — vessel detection, vehicle counts, change alerts.
BlackSky’s 2024-2025 contract wins with the US Department of Defense and allied governments have anchored the revenue base. The commercial customer growth has been steadier than dramatic; the defence pivot defines the company’s current shape.
Capella Space and ICEYE: the SAR specialists#
Synthetic Aperture Radar (SAR) — imaging through clouds and at night — is the other half of the earth observation market. Two companies define the commercial SAR space:
Capella Space, San Francisco-based, operates ~10 SAR satellites with sub-50cm resolution capability. Capella’s strength is on-demand tasking with rapid turnaround; customers can get imagery hours after request rather than days.
ICEYE, Finland-based, operates a larger SAR constellation (~30 satellites) and has expanded into flood mapping, dark vessel detection, and insurance applications. ICEYE’s commercial momentum has been strong — particularly in insurance (flood and storm damage assessment) and government (military and civil protection).
SAR is undervalued by customers used to optical imagery. The advantage of seeing through clouds is enormous in tropical and Northern European regions where optical revisit is bottlenecked by weather. SAR is also the only realistic answer for nighttime imagery and for change detection independent of lighting conditions.
Synthetaic, Orbital Sidekick, and the analytics layer#
The 2026 analytics-and-applications layer has matured to the point that many of the most interesting EO companies don’t operate satellites at all.
Synthetaic operates the RAIC platform, which applies machine learning at scale to imagery from any provider — finding aircraft, ships, oil tanks, encampments across continental-scale imagery in minutes. The company’s public demonstrations (finding the downed Chinese spy balloon in 2023, mapping Russian military movements) established its credibility.
Orbital Sidekick operates hyperspectral satellites — capturing imagery across hundreds of spectral bands rather than just RGB — and focuses on energy infrastructure monitoring, particularly methane plume detection. Hyperspectral is a smaller market but the per-customer revenue is high.
Descartes Labs (acquired by EarthDaily in 2023) operates a geospatial analytics platform layered on imagery from multiple providers. The platform abstracts the imagery source from the analytic, letting customers buy answers rather than pixels.
Allen Institute for AI’s Skylight, Global Fishing Watch, SkyTruth, and other non-commercial analytics platforms also matter. The civil-society and academic analytics layer is consequential in the EO ecosystem.
EO-as-a-service vs raw imagery#
The business model split in 2026:
Raw imagery sales — buy pixels of a specific area at a specific time. Per-image pricing. The classic Maxar business. Still important for high-value one-off needs (intelligence, journalism, specific events).
Imagery archive access — annual or multi-year licence to query the imagery archive. Planet pioneered this model. Smooths revenue, encourages broad analytic exploration.
Tasking-as-a-service — customer specifies what to image and when; vendor delivers. BlackSky’s core offering, increasingly part of Capella, ICEYE, and Planet’s models too.
Analytics-as-a-service — customer specifies the answer needed (e.g., “tell me when new construction starts at these coordinates”); vendor handles imagery, ML, and delivery. The fastest-growing segment.
The economic gravity is moving up the stack. Customers are willing to pay more for answers than for pixels, and the analytics layer captures more of the value than the imagery layer.
On-satellite AI and downlink reduction#
The technical inflection that matters in 2026: AI inference running on the satellite, not on the ground. The conventional EO pipeline downlinks raw imagery, processes it on ground servers, then delivers results. That pipeline produces enormous downlink demand — modern EO satellites generate terabytes per day per satellite.
On-satellite AI flips the pipeline. The satellite runs inference at capture time and downlinks only the relevant tiles or extracted features. Bandwidth cost falls dramatically, time-to-insight shrinks, and the same downlink budget supports a much larger constellation.
Players moving on this: Planet’s “Insights at the Edge,” ICEYE’s on-satellite SAR processing, Orbital Sidekick’s hyperspectral filtering, and a number of startups providing on-orbit compute. NVIDIA Jetson and specialised radiation-tolerant AI chips are showing up on satellite buses for this purpose.
The economic implication: on-satellite AI is the prerequisite for the next generation of EO economics. Without it, downlink costs become a constraint on constellation scale; with it, the constraint moves to compute and antenna.
What the 2026 customer landscape looks like#
The customer mix has broadened beyond the historical defence-and-intelligence dominance:
- Agriculture — Planet, John Deere, Climate FieldView, Granular use EO at scale
- Insurance — ICEYE, Cape Analytics, Plnar, Galaxy use SAR and optical for property and natural-catastrophe pricing
- Finance — hedge funds use Planet, Maxar, BlackSky for alternative data signals
- Energy — methane monitoring (MethaneSAT, Orbital Sidekick, Carbon Mapper), pipeline monitoring, oil-tank inventory
- Supply chain — vessel tracking, port monitoring, factory activity
- Environmental — deforestation, glacier melt, water resource management
- Defence and intelligence — still the anchor customer base
The diversification is good for the industry’s stability — no single customer segment is large enough to crash the market.
What we tell clients#
For clients evaluating earth observation in 2026:
- Define the answer, not the imagery — start from what decision the EO data will inform
- Multi-vendor by default — different providers fit different use cases; lock-in to one is rarely the right call
- Analytics platforms abstract the imagery layer — Planet Insights, Synthetaic, Descartes, Orbital Sidekick can deliver answers without the customer touching pixels
- Plan for SAR alongside optical — cloud cover and nighttime gaps are real, and SAR closes them
Our data engineering and AI/LLM integration work touches this where clients integrate EO-derived signals into broader analytics platforms.
The 2026 picture#
Earth observation in 2026 is a real, diverse, growing market. The pixel-sales business is mature and slowly shrinking in proportional importance. The analytics-and-applications layer is the growth story. On-satellite AI is the technical inflection. Defence remains the anchor; agriculture, insurance, finance, and energy are the diversification.
For most enterprises, EO has gone from “interesting” to “operational input.” The companies that figure out how to translate imagery into decisions will continue to capture disproportionate value over the next decade.
Related reading#
- Satellite constellations in 2026: Starlink, Kuiper, OneWeb, IRIS²
- Starlink economics in 2026: subscribers, V3, Direct-to-Cell
- AI in agriculture: precision farming patterns
Earth observation in 2026 is an answers business, not a pixels business. If you’re integrating EO-derived signals into your analytics platform, our data engineering team builds the integration spine. Tell us about the use case.