

Discover more from TerraWatch Space Insights
8-Things Earth Observation: September/October Conference Edition 🌍 🛰
Analysing some key developments in EO industry from the month that passed
Hey! Welcome to the latest edition of 8-Things EO. Ever since I sent out the previous edition, I have been attending one conference after the other, non-stop throughout September and the first week of October. So, I am calling this one the ‘Conference Edition’ and no surprise, there were a lot of developments in Earth observation. Let’s unpack some interesting ones!
Sponsored by
Three Curated Things
A non-exhaustive summary of some key developments in Earth observation since the last edition.
1. Funding, Contracts and Financial Stuff 💰
PlanetWatchers, one of the few SAR analytics software companies focusing on the agriculture market, raised a $11M series A round.
GRASP Global, a French startup that will launch a cubesat constellation to characterise air pollution from space raised a €2M pre-series A round.
Pano AI, a California-based startup providing wildfire monitoring based on artificial intelligence raised a $60M series A round (potential EO user).
4M Analytics, an Israeli geospatial startup raised $30M to map US underground utility infrastructure (potential EO user).
Ideon Technologies, a Canadian subsurface imaging solution provider for the mining industry raised a $16M A round (potential EO user).
NOAA announced nine awards totalling $13.8M for new weather observation sensors, including $4M to Spire for hyperspectral microwave sensor.
Spire was also awarded a NASA contract to provide data from the SNOOPI mission that will evaluate the use of radio signals to measure soil moisture.
Satellogic signed a contract with the Government of Albania to offer its satellites through its ‘constellation-as-a-service’ offering and announced a contract to monitor all subnational protected areas on the planet.
Two Irish companies, Davra and Treemetrics, were awarded over €3M in total funding to develop solutions for the mining and forestry sectors.
Blacksky has won a $14M contract from the US National Geospatial-Intelligence Agency to monitor global economic activity, leveraging its Spectra AI platform.
The US National Reconnaissance Office signs agreements with six commercial providers of space-based radio frequency data: Aurora Insight, HawkEye 360, Kleos Space, PredaSAR, Spire Global and Umbra Lab.
Orbital Insight was awarded a $950M (😲) contract by the US Air Force to develop AI-powered multi-sensor geospatial analytics.
2. Announcements, Partnerships and Strategic Stuff 📈
Planet has shared more details on its upcoming hyperspectral satellite constellation which will have 400 spectral bands at a 30 meter resolution.
Spire has a new customer for it's space-as-a-service offering, GHGSAT - it will launch 16 cubesats with greenhouse gas monitoring payloads.
Finnish EO company Iceye and Spanish EO startup Satlantis will launch a tandem optical-SAR satellite constellation - what looks like a step towards creating an indigenous EO constellation from Spain.
Open Cosmos announced OpenConstellation, a shared space infrastructure for EO, what looks like a bespoke satellite-as-a-service offering.
Google has acquired Israeli air quality monitoring startup BreezoMeter, a company that uses Copernicus data, for over $200 million.
ESA selects Harmony as the newest Earth Explorer mission - two identical satellites with receive-only SAR and thermal-infrared sensors contributing to global natural risk monitoring. ICYMI, the past ones have been very successful.
BCG acquired Quantis, an environmental consultancy specialising in building science-driven sustainability solutions(including using EO like this one).
Microsoft, Planet and The Nature Conservancy launched the Global Renewables Watch to map and measure all utility-scale solar and wind installations using EO.
NOAA launched Argos-4, a wildlife tracking satellite (developed in collaboration with CNES), and getting ready to launch the Joint Polar Satellite System-2 weather satellite on November 1.
Azure Orbital Ground Station, Microsoft’s ground station as a service, has partnered with Pixxel, Muon Space, and Loft Orbital.
Cape Analytics, a property insurance provider leveraging EO and Canopy Weather, a weather provider have launched ‘Hail Roof Damage Rating.’
3. Reports, White Papers and Interesting Links 🔗
This course from EUMETSAT, ECMWF and others on AI for Earth Monitoring.
This paper explaining how EO could be used for tracking the Paris Agreement.
This quality assurance framework for EO - an important step for the industry.
This (past) training from EUMETSAT on atmospheric composition from satellites.
This open-source online platform for near-real time monitoring of dams and environmental impacts in the Mekong Basin (discovered through SatSummit).
This piece on Chinese EO satellites (a good complement to this episode).
This Google Earth Engine remote sensing 101 book (open access).
Interested in geospatial? In the Geomob podcast, Ed Freyfogle and Steven Feldman profile interesting geospatial projects - from start-ups, industry giants, government, academics, or hobbyists.
From Jack Dangermond, co-founder of Esri to the anonymous author of the open source intelligence newsletter, Line of Actual Control, they have had several interesting conversations. Check it out and subscribe to their podcast!
Five Things From The Conference Season
Major talking points (and misses) from the conferences I attended and participated in the last month - World Satellite Business Week, International Astronautical Congress, SatSummit, EU Space Week and Space for Food Security.
4. Commercial Earth Observation Market: The Dilemma
I gotta start from this argument (which actually was the highlight of the World Satellite Business Week) between Payam Banazadeh, CEO of Capella Space and Joe Morrison, VP of Product at Umbra Space, as this is what I want conferences to be about - more arguments, less agreement; more thought-leadership, less self-glorification; more ensuing discussions, less moving on to the next conference.
What’s interesting about this situation is that you can make a case for both sides: as a rational market analyst I can understand Payam’s (Capella Space) point of view that the focus of the EO companies should be serving their top customer: the government. While, being a wishful evangelist, I am shaking my head in agreement with Joe (Umbra Space) on the need for EO companies to focus on the commercial sector.
The problem is that (let me throw some consulting jargon at you) the customer persona, the willingness to pay and the buying habits of both these segments vary so much that they both fall at two ends of the spectrum. Generally speaking, governmental customers have in-house EO data analysis teams, have a bigger pocket and would want to monitor large areas, but ‘zoom in’ on-demand, while commercial customers typically don’t have an in-house remote sensing teams, are very price sensitive and want to monitor selected areas continuously. As you would imagine, the delivery infrastructure for both these segments would vary, the data pipelines would look different and the amount of pre-processing needed could potentially cannibalise the other segment (in terms of resource allocated).
So, what’s the solution? My take is that there will be some EO data companies that will not be built for the government, they will be built for the commercial sector. Typically, they would be “verticalised” aimed at solving a problem within an industry, at least to start with, that can build for a specific customer persona, with a defined willingness to pay and a particular buying habit, which will allow them to hopefully scale. Another solution would be for the EO companies to open up their archives, but that’s a whole different conversation!
5. Inadequate Focus On Data Distribution, Standards and Interoperability
I was (not really) surprised at how the major conferences have a lot to say about the hundreds of satellites and sensors that were launched and about to be launched, but not so much about what happens after they are launched. Thankfully there was SatSummit that had enough importance given to this area, but I would like to see more conferences start discussing these subjects:
How is EO data distributed today? Is the data acquired interoperable with other types of EO data - within one sensor and among other sensors? What is the standard format for sharing the data? Will there be one standardised API to distribute analysis-ready data to the intermediate users? And, do we have enough people in the world who know how to work this data?
I put out this tweet a while ago about my wish to have this “all-purpose, sensor-agnostic engine” for distributing and processing EO data - the responses I got on LinkedIn and Twitter were interesting. Several folks in the Dissemination layer are attempting this in a number of ways, albeit incrementally, along with open source initiatives like STAC, COG, Pangeo among others. There was also what I call “a necessary experiment” from Descartes Labs in this area, which came to an end (or not?) recently, but I am convinced this is something that we need to solve, if we have any hopes of making Earth observation mainstream.
6. The Lack Of A Holistic Conversation - Sticking To Silos
I have written about the need for a holistic approach in EO before (#8) - one that would focus on solving problems through EO instead of forcing an EO technology to solve problems. But, I think most discussions we have during conferences are still stuck in silos: by sensor (SAR, hyperspectral, infrared), by medium (satellite, aerial, drone) and by segment (weather, IoT, navigation, communication).
Most problems that I can think of require holistic solutions, independent of sensor or medium. A real-time wildfire monitoring solution requires a combination of weather, satellite imagery and IoT data; a global flood mapping tool would use a combination of satellite, aerial and drone imagery; an effective crop monitoring solution in a developing country would need high-precision navigation to deliver insights via mobile network operators. How long until we move away from silos and move towards partnerships?
The faster we embrace a holistic approach to problem-solving in Earth observation the faster we will grow and the more impact we can have. The longer we continue with a siloed approach, the higher the complexity for the end-user.
7. Backward Vertical Integration: An Inevitability?
This is an interesting subject and again, it doesn’t get much attention. What does the future of EO adoption look like? Would organisations buy EO data from the Acquisition layer and build remote sensing teams? Or, would they procure EO-derived products from the Intelligence layer and integrate the insights? Or, would it be a mix of both? I wrote about these strategies in this thread.
But, I want to highlight the outcome of some conversations I had during the Space for Food Security conference: Some large organisations are more likely to build in-house remote sensing teams and buy EO data as a primary approach, while there are several Intelligence companies offering analytics, insights and applications.
The reason? Lack of standards and reliability. I think they may have a point - today, we have several solutions in the market for any use case - be it crop intelligence, deforestation monitoring or flood insurance. Each of these solutions have their very own crop identification algorithm, tree species detection technique and building footprint analysis method. As an end-user you are left wondering how to choose between the options? Do you need to make up your own mind or is there going to be a global authoritative voice for EO-related products? An ISO for EO, if you will!
I had Yana Gevorgyan, the head of the Group on Earth Observations, on the podcast and we touched on this subject - on GEO’s role in being that voice. But, will GEO have the expertise to be an authority across all sectors and all use cases? Today, if an organisation, say, wants to integrate insights from satellite data for providing finance to smallholder farmers, they are left with a complex researching, due diligence and procurement process to handle. So, in some cases, these end-user organisations might choose to build their own data pipelines, ground validation and data processing tools than depending on an external service providers, for the sake of confidence. Certainly feels like we should be having more conversations on this topic, doesn’t it?
8. More Validation For The Need For An Advisory Service Layer in Earth Observation
My conversations with corporates & end-user organisations, space agencies, non-profit organisations and EO service providers continue to validate my thesis that there needs to be a proper advisory service layer in EO. There are two ways this layer can operate: an advisory service layer that will serve as a complement to the market with the adoption of EO or the more complicated advisory solution layer that will end up competing with the current offerings in the EO market.
The advisory service layer is an enabling layer that works independently, objectively and holistically and is based on a model that focuses on leveraging existing solutions instead of reinventing the wheel. In this case, the focus is on enabling the adoption of EO data, so more strategic EO consulting for the end-user. An example would be the set of tasks TerraWatch Space performed for one of our customers, as detailed below:

The advisory solution layer is focused on solving problems by developing custom solutions for the customer - the kind I expect McKinsey (see slide #9) and EY to provide based on my understanding of their offering. In this case, the focus is on building a solution for a customer leveraging EO data, potentially exploiting their in-house teams. Some technology consulting companies may choose to differentiate on the Acquistion layer (see Accenture’s investment in Pixxel), while others may choose to differentiate through partnerships with the Intelligence layer (see BCG’s partnership with Jupiter Intelligence). We are going to see more similar developments. Watch this space!
So long!
Aravind - Founder, TerraWatch Space