The appeal of hyperspectral satellites has long been apparent.
By measuring the intensity of light in each pixel in hundreds of narrow wavelength bands, hyperspectral sensors allow users to see methane emissions in the atmosphere, fertilization in the soil and determine whether they’re looking at camouflage or canopy.
It’s a way to “decode the Earth at a chemical level,” said Shoaib Iqbal, CEO and co-founder of Australian hyperspectral startup Esper Satellite Imagery.
Constellations of small satellites have promised global insights that are impossible to obtain with airborne hyperspectral sensors and revisit rates that improve on government hyperspectral missions that often cost hundreds of millions of dollars.
But for more than a decade companies have grappled with the complexity of calibrating hyperspectral sensors and turning the raw data into information customers value.
Now, investors and entrepreneurs see a light at the end of the tunnel.
“The market is transitioning from being supply-led, where companies were proving what the technology could do, to demand-led, where governments and enterprises are increasingly recognizing what problems it can solve,” said Mark Boggett, the CEO of the investment group Seraphim Space. (Seraphim has invested in Pixxel of India and the United States, Canadian startup Wyvern and constellr of Germany.)
Research firm Analysys Mason expects the market for satellite-based hyperspectral imaging, including data and downstream analytics, to grow at a compound annual rate of 60%. The research firm’s Earth observation report published in 2024 forecasted growth of the hyperspectral market from $3.14 billion to $99 billion over a decade.
Artificial intelligence
One reason for the optimistic forecast is AI.
“Five years ago, extracting value from hundreds of spectral bands required specialist scientists,” Boggett said. “Today, AI models are increasingly able to identify patterns within highly complex datasets.”
AI is key for making sense of the information U.K.-based Open Cosmos will obtain when its OpenConstellation 1.0 delivers daily, high-resolution hyperspectral and high-resolution multispectral imagery covering 3 million square kilometers, said Alberto Perez Casinelli, the company’s vice president of data.
Similarly, AI is helping Xplore with satellite operations, software development and quickly identifying the spectral pattern associated with a particular material. “Innovation for hyperspectral is happening at an advanced pace because of AI,” said Lisa Rich, chief operating officer for hyperspectral imagery provider Xplore.
For Planet, AI “is helping with the production of code,” said Jeff Guido, Planet vice president for space mission delivery. “But as far as interrogating the dataset itself,” human expertise remains essential because “scientifically rigorous processes oftentimes are needed.”
Early customers
Government agencies have been among the first customers to acquire space-based hyperspectral data.
“What was once viewed primarily as a commercial Earth observation capability is increasingly being recognized as a strategic national asset,” Boggett said. “Governments are starting to appreciate that hyperspectral data can provide unique insights that conventional optical imagery simply cannot, from identifying specific minerals and monitoring critical infrastructure, to detecting environmental contamination, supporting food security and enhancing defense and intelligence operations.”
Hyperspectral data can detect camouflage and reveal decoys for military equipment. Stealth vessels, designed to avoid detection, can also be exposed with hyperspectral, said Kuva CEO Jarkko Antila.
In the U.S., NASA and the National Reconnaissance Organization have awarded contracts to commercial hyperspectral data providers. Kuva US, Orbital Sidekick, Planet, Teledyne Brown and Wyvern are among the 14 companies competing for multiple awards through November 2028 in NASA’s Commercial Satellite Data Acquisition (CSDA) program with a maximum contract value of $476 million. U.S. government researchers, contractors and federal, state and local government agencies can request data, which NASA gathers through task orders.
Approximately $75 million of the $476 million CSDA ceiling had been obligated under the CSDA program as of Aug. 20, according to the federal government’s USASpending website.
For NASA, commercial data complements observations drawn from the Earth Surface Mineral Dust Source Investigation, a NASA imaging spectrometer mounted on the International Space Station.
“The neat thing about commercial constellations is they are taskable. NASA missions typically are not,” said Melissa Martin, NASA’s CSDA program manager. “Commercial hyperspectral data improves the spatial and temporal resolution and enhances continuity for key observables.”
The intelligence community’s National Reconnaissance Office began evaluating the hyperspectral capabilities of BlackSky, HyperSat, Orbital Sidekick, Pixxel, Planet and Xplore in 2023 to determine how the commercial observations fit with the intelligence agency’s remote-sensing architecture. Through the initial contracts, the NRO analyzed each firm’s observation capabilities, business and cybersecurity plans.
In May, the NRO tapped Pixxel for a follow-on award to assess its on-orbit constellation and purchase data.
The market has expanded internationally as well. The European Union selected Finland-based Kuva Space’s Hyperfield-1 microsatellite constellation as an Emerging Copernicus Contributing Mission in 2023. Kuva also won a 5 million euro contract in 2023 1.8 million euro contract in 2024 from the European Space Agency Civil Security from Space program.
Similarly, Wyvern has been under contract with the Canadian government since the Edmonton, Alberta, startup began selling hyperspectral data in 2024. Wyvern now operates a fleet of five satellites.
The Chinese government is backing commercial hyperspectral satellite developers as part of its broader investment in Earth observation.
Commercial market
Outside of government, the commercial market for hyperspectral data is emerging.
Mineral exploration and agriculture technology customers tend to be knowledgeable about applications for hyperspectral data.
“A big part of the customer discovery journey of Wyvern is understanding who can take the data and run and which customers are worth shepherding along,” said Callie Lissinna, Wyvern co-founder and vice president for strategic relations.
Kuva provides hyperspectral data to commodity traders and reinsurance firms seeking information on crop health and soil moisture. Customers also evaluate biodiversity, monitor algae growth, search for minerals and identify chemical leaks with Kuva data.
Pixxel’s six-satellite Firefly constellation provides daily observations for agriculture, mining, energy, defense and environmental monitoring applications.
Planet’s Tanager-1 satellite launched in 2024 detects and quantifies methane and carbon dioxide emissions with a hyperspectral instrument built by NASA’s Jet Propulsion Laboratory for a philanthropically funded public-private coalition led by the nonprofit Carbon Mapper.
“Carbon Mapper has created a portal where anyone can see the methane detections that have been performed with Tanager and other instruments,” Guido said. “We want to make it intuitive. Anyone that has a browser can get those insights that they need.”
For now, Esper remains focused on mining. The Australian startup combines hyperspectral data from government and commercial satellites with subsurface geological observations to help customers identify mineralization zones, monitor production and rehabilitate tailing dams, where mining byproducts are stored.
Open Cosmos works with industry partners to create products for agriculture, maritime, mining and oil-and-gas applications. With a constellation that includes Earth observation and communications satellites, “we can provide real-time rich data,” Perez Casinelli said. “Our strategy is a combination of sensors, high-resolution [electro-optical] and hyperspectral, connected to our telecommunication constellation.”
Xplore applies data from XCUBE-1 launched in late 2024 to customers working in defense and intelligence, agriculture, forestry and space domain awareness.
Innovation ahead
In spite of significant progress made in the hyperspectral sector, “there is room for improvement on sensor technology, optics, constellation design, communications architecture and data processing on the ground,” Lissinna said. In addition, companies will keep striving to improve tasking and revisit rates until their customers “have access to anything they need.”
Initially, some entrepreneurs and investors may have underestimated the technical challenges.
“There’s a misconception that you can simply buy a hyperspectral instrument, bolt it onto a satellite, launch it and expect world-class imagery,” Rich said. “The quality of the imagery ultimately depends on the entire system: optical design, spacecraft stability, thermal management, calibration methodology, image processing, software, mission operations and countless engineering decisions that determine whether the data is merely usable or truly exceptional.”
With the growing popularity of hyperspectral data, analysts think the sector may be poised for mergers or acquisitions.
“Consolidation is a distinct possibility,” said Claude Rousseau, Analysys Mason research director. “These hyperspectral players don’t have scale right now. Most of them are still very small with respect to the rest of the market. But they have some very interesting data.”
Few doubt that demand for hyperspectral data will grow.
“I’m bullish on hyperspectral in the long term,” said Scott Herman, Earth observation technology specialist and SatVu chief technology officer. “But we haven’t had that high-resolution hyperspectral combined with the killer use cases for it that would cause it to break open.”
Like synthetic aperture radar (SAR), hyperspectral adoption will take time because it requires “tools, tradecraft and training,” unlike electro-optical imagery, which is easier to interpret, Herman said.