EarthDaily aims to cut the work between satellite observations and useful decisions
Burn scars from the Bald Range wildfire near Summerland, BC, Canada. Image acquired September 17, 2026 (image: EarthDaily)
Satellite observations can reveal crop water stress before plants visibly change colour, giving buyers, insurers and lenders time to reassess risk. Across thousands of fields, however, detecting those signals requires more than frequent images: users need measurements they can compare without repeatedly cleaning, correcting and aligning the data themselves.
EarthDaily Analytics aims to reduce the work required to turn repeated satellite observations into useful decisions. With eight satellites in its EarthDaily Constellation and an automated processing system, the company aims to make broad-area monitoring practical for organisations that cannot inspect every image individually.
“We are an AI and analytics company first, and that's the reason we exist: to turn Earth observation data into answers organisations can act on,” says Don Osborne, EarthDaily’s CEO and president, in a written interview with MoveTheNeedle.news.
EarthDaily's history goes a long way toward explaining this focus on analytics. The company was formed in Vancouver in 2021 with backing from private equity firm Antarctica Capital, whose affiliates acquired satellite-project assets and the Geosys agricultural analytics business from UrtheCast following its filing for creditor protection. The constellation builds on a programme UrtheCast began in 2016, while Osborne, a mechanical engineer by training, became UrtheCast’s CEO in 2018 after senior roles at Canadian space technology company MDA.
Inside the EarthDaily satellite constellation
While EarthDaily has designed its satellites to monitor large areas frequently, it accepts some limits on the fine detail they can capture. Each EarthDaily satellite carries 16 imaging assemblies that collect data across 22 spectral bands while imaging a strip of Earth 240 kilometres wide. Osborne says this balance is particularly useful in agriculture, where repeated observations can reveal crop conditions changing over a matter of days.
The company specifies native pixel spacing of five metres for visible and near-infrared bands, 95 metres for shortwave infrared and 120 metres for thermal infrared. Delivered products differ: its AI-ready surface-reflectance product includes nine bands sampled at five or ten metres, rather than all 22 bands at five metres.
Osborne describes nearly daily coverage across most land areas and says two more satellites are ready to launch in 2027. In its 26 October announcement of a NASA data-evaluation contract, EarthDaily gives a more specific coverage figure: a two-day cycle with eight satellites, rising to about 92% of Earth’s landmass daily, excluding Antarctica, at full operations.
What those passes reveal depends on the wavelengths measured. Shortwave infrared responds to moisture in soil and vegetation, while thermal measurements provide temperature information. Other bands have a supporting role: as Osborne explains, “some compensate for the atmosphere and improve cloud masking”.
EarthDaily relies on Loft Orbital and Airbus for spacecraft infrastructure, and ABB for optical instruments and payload integration, allowing it to concentrate resources on processing and analytics.
Preparing Earth observation data for AI
Is a crop deteriorating, or has a change in sunlight made it look different? For an AI model comparing satellite images, that distinction is crucial: differences in atmosphere, viewing angle or sensor response can trigger false alarms or hide early signs of stress.
EarthDaily tackles the problem in orbit and on the ground. Its satellites look straight down and observe locations at roughly the same local morning time. Osborne says the company works with calibration specialist L-1 Standards and Technology and checks its measurements daily against eight science missions to detect drift.
Automated processing then corrects atmospheric effects and image geometry, masks cloud, shadow and haze, and aligns pixels with the same ground locations across observations. At roughly 100 terabytes of acquisitions a day, according to Osborne, doing this image by image would be impractical.
The result is intended to let customers put data into models with less preparation. EarthDaily reports compliance with the Committee on Earth Observation Satellites’ Analysis Ready Data framework (CEOS-ARD), while its “AI-ready” description emphasises consistency between observations. Reliable inputs help a model detect meaningful changes, although its conclusions still require validation.
EarthDaily first applied this processing approach to public Sentinel-2 and Landsat imagery. Connecting its own observations with those archives could give customers a longer record against which to judge emerging changes.
Agriculture and maritime monitoring
For agriculture, that approach helps reveal how crops are responding to changing conditions. Osborne distinguishes this work from measuring climate directly: “We don’t have weather sensors; we combine satellite observations with weather information to understand how conditions such as drought and heat are affecting crops.”
He cites the 2022 Brazilian soybean season, when the team observed deteriorating vegetation associated with low rainfall in Center-South Brazil. According to Osborne, it lowered yield forecasts in December, before forecasts from Brazil’s CONAB and the US Department of Agriculture reached similar levels later that season.
Although the example predates the constellation, it illustrates how earlier warnings could inform purchasing, sourcing or production-risk assessments.
Beyond agriculture, geospatial intelligence company SynMax has subscribed to EarthDaily imagery for maritime intelligence. The agreement demonstrates adoption, although Osborne cautions: “It's early in that rollout, so I don't have measurable results to point to yet.”
Whether those applications become dependable operational services will determine the commercial value of the system. As Osborne puts it: “Success for us means data that holds up the same way in production as it did in a demo, sector by sector.”
Liked this article? You can support our independent journalism via our page on Buy Me a Coffee. It helps keep MoveTheNeedle.news focused on depth, not clicks.