So I have been thinking about a simplified way to describe what @terrawatchspace is and what we actually do - here's what I have come up with: TerraWatch helps you think better about Earth observation and we do so in 3 ways: 1. Newsletter = one-to-many thinking: Publish independent analysis and insights on EO markets, applications, policy and adoption 2. Advisory = one-to-one thinking: Work directly with governments, EO companies, startups, enterprises and investors on specific strategic, market and policy questions related to EO 3. Events (like EO Summit) = many-to-many thinking: Bring together the EO ecosystem and users across industries to exchange perspectives on EO market evolution, applications and adoption Three different but connected activities, but ultimately the same goal: helping people make sense of EO and what EO means for them 🛰️ 🌍

The thermal Earth observation market is entering a new phase as companies invest in scaling their capabilities. Recent moves from Hydrosat, constellr and SatVu point to very different strategies, with each company making different bets around resolution, coverage and revisit. This week’s TerraWatch Pro compares the three approaches and what they reveal about how the commercial thermal EO market is evolving. Become a paid subscriber to get weekly market briefings with insights and analysis on EO market and its applications. newsletter.terrawatchspace.c…

Embeddings is the new buzzword in Earth observation - several folks outside the EO bubble kept asking me for a simple explanation. So, we took a stab at helping demystify what EO embeddings are, how they work and why they are useful. This infographic is part of the latest @terrawatchspace deep-dive on EO foundation models & embeddings, how they are commercialised and what they mean for the future of the EO market. Full access for Premium subscribers and free preview for all (linked in comments).

Last week to register for our annual Earth Observation for Climate event, happening during Climate Week NYC. We will have two sessions focusing on risk, specifically wildfire risk and agricultural risk, with a group of fantastic speakers: 🌾 Agricultural Risk - Lilly Hancock from General Mills - ​Duncan Anderson from NVIDIA - ​Annalee Tai from Swiss Re - Raha Hakimdavar from Georgetown University 🔥 Wildfire Risk - ​Keith J. Masback from Earth Fire Alliance ​- Will Stikeleather from Marsh - ​David Saah from Spatial Informatics Group, LLC - Phil Green from SKYFOREST (an EarthDaily Analytics Company) We are grateful to EarthDaily for supporting the event. -- 📅 Thursday, September 24 🕔 5:00–7:00 PM 📍 Midtown Manhattan Register here: luma.com/lsr82ot1

The recently concluded Space Summit in Paris included two declarations on Earth observation - obviously, both got very little attention given that EO is always taken for granted and nobody cares until we have a problem. The two declarations are actually interesting- the first one set out the principles for space-based scientific including accessibility, openness, integrity and international cooperation. The second called the Copernicus programme a global public good and confirmed the need for its continuity. But does this move us closer to the global EO political governance framework that I have argued is missing? What can we learn from which countries signed these declarations, and which ones did not? In my new essay, I look at what the two declarations actually say, what they leave out, and whether a framework for open EO data can work even if it is not global.

New @terrawatchspace deep dive on Earth Observation Foundation Models (EOFMs) - an in-depth report on what they are and how they will change the EO market with several infographics. The piece covers a bunch of topics including: - What EOFMs and embeddings are (an attempt at demystifying them) - The three ways EOFMs are commercialised in EO - Which use cases are a good fit for EOFMs - How EOFMs could change the EO value chain - The EOFMs market landscape - How EOFMs will change the EO market and more.

Earth observation is such a fascinating sector work in because it is three things at once - a strategic asset focused on sovereignty, an enterprise market with scaling revenues and a public good focusing on open access and research (somewhat similar to the ongoing discussions with respect to AI). In reality, EO cannot and should not just be one of the three, but as we are currently seeing the market (and geopolitics) evolve, we are also asking ourselves if it can really be all three, properly. In this week's @terrawatchspace Pro newsletter, I laid out some thoughts on this concept - focusing on how essentially this amazing multi-purpose technology is going through an operational identity crisis now as the world around it changes. More to come on this from me!

It was a pleasure for @terrawatchspace to contribute to this Copernicus programme evaluation report along with colleagues at knowspace and Frontier Economics, assessing the impacts for UK after they rejoined Copernicus programme in 2024 after leaving in 2021. Participating in Copernicus is a long-term strategic decision for any country. While contract returns are modest for the UK (mainly due to the disruption after leaving), being a part of the programme is critical for UK's EO sector. Data from Copernicus missions is already deeply embedded in UK public services, climate research, and AI model development. There is also a common misconception that since “Copernicus data is free”, membership is not valuable. But as our report shows, being a Copernicus member provides various benefits: fewer download and access restrictions, archives, calibration and validation, crisis services, Contributing Missions data, and influence over future data capture. This is an interim report - it is still too early to judge the industrial return for the UK from rejoining Copernicus, but being in Copernicus ensures the UK remains a key player in the global EO sector and guarantees access to unmatched EO data.

To get exclusive weekly briefings with analysis and insights on Earth observation and applications, become a TerraWatch Pro subscriber. newsletter.terrawatchspace.c…

Commercial weather satellites have been around for more than a decade, but the market has been slow to develop. Since 2016, NOAA has awarded over $180M in contracts to procure data from weather satellite companies such as Tomorrow.io, Spire, PlanetiQ, Weather Stream, Muon Space etc. Yet, reliable, recurring demand for satellite-based weather data never materialised. But, all that could change with NOAA's planned $2.3B contract vehicle for data buys over the next decade, For weather satellite companies, that could finally provide the recurring government anchor the sector has lacked. So, does this segment now become a much bigger market? The answer may not be straightforward. Satellite-based weather is an unusual market because weather observations are treated as a public good: under the WMO Unified Data Policy, NOAA requires open licensing and broad redistribution rights for the commercial data it purchases (something I broadly agree with). So, in other words, NOAA is creating a market by becoming a major buyer of commercial data, while simultaneously limiting suppliers’ ability to sell the same data elsewhere. In this week's @terrawatchspace Pro newsletter, I discuss how NOAA's procurement has evolved from radio occultation to other sensors, why weather is an unusual market, how AI weather models could change the equation and the potential impacts NOAA's new multi-billion $ program.

I am convinced that the ongoing pivot of Earth observation towards the defense market will not remain limited to traditional defense use cases. The line will blur as climate change moves from being purely a sustainability/environment category to a strategic one, with security implications. As that happens, many non-defense EO applications, which seem to have been sidelined today, will become priorities again (but through a different lens): - crop and drought monitoring for food security, - snow, ice and water level monitoring for energy and resource security, - wildfire, flood, heat, risk for coastal and sea-level risk for critical infrastructure, real estate and insurance, - production monitoring across agriculture, mining and energy for commodity and supply-chain risk. Currently, it seems like the conversation in EO is whether to focus on defense applications or on climate applications (civilian/commercial). Maybe a better way to think about it - whether it is for investing in the next satellite or developing the next EO application - is to consider them as overlapping defense, civilian and commercial needs rather than treating them as separate markets. Focusing on defense may look like the safest place to be in EO today, but it also creates concentration risk around one market, one type of buyer and one demand cycle. The most sustainable EO businesses of the next decade would be those that can focus on a diversified portfolio of markets from a common core capability, but without forcing them all into the same product and delivery model (see my Defense Paradox essay).

I first created this Earth observation value chain 6 years ago, in order to be able to understand and analyse the EO market better. A lot of people said this is too complicated and was not necessary as EO is really just about providing data or services - I strongly disagreed as I was convinced we need this level of granularity to differentiate the trends and to be able to position companies better. Cut to 2026, each layer of the value chain has evolved: 1. Infrastructure has become the biggest revenue generating segment of EO due to sovereign demand and emergence of satellite-as-a-service and satellite-for-purchase models 2. Data is going through an evolution with increased demand for analysis-ready and increasingly AI-ready satellite data 3. Platforms are seeing a revival after the dip early on due to lack of demand, but AI-based platforms allowing data fusion, model fine tuning and custom embedding generation 4. Analytics as a layer is quietly starting to disappear due to two things - foundation models getting more capable of transforming pixels into information and edge computing enabling downlink of information not imagery 5. Insights is going through an evolution as LLM-based tools are becoming available capable of delivering a report or dashboard with geospatial agents running the processing in the background 6. Applications is highly competitive as the moat is not just having an EO-based product that solves a specific problem, but also integrating with operational workflows of customers, potentially leading to acquisitions. In the long-term, I expect: - the Acquisition segment to integrate better providing multi-sensor, low latency, AI-ready products, - the Processing segment to move to orbit in the best-case scenario and, - the Intelligence segment to achieve growth by possessing proprietary domain-specific data/models and customer needs understanding. -- Subscribe to the @terrawatchspace newsletter for Earth observation insights and analysis. Reach out to me for bespoke consulting work and advisory support.

Become a Pro subscriber and get a weekly briefing with exclusive EO insights and analysis. newsletter.terrawatchspace.c…

Google announced the Private Preview of Custom Satellite Embeddings, a new data product powered by the AlphaEarth Foundations AI model (currently in closed beta). I think of this as the first step towards: a) productising a research milestone announced last year of static, annually updated datasets to flexible, on-demand satellite embeddings updated every five days and b) distributing it for commercial use through the through the Google suite of products (Maps, Cloud etc.). This new development from Google looks like what Clayton Christensen called a disruptive innovation - a cheaper, innovative product that could displace many incumbents. I think embeddings could very well change how geospatial intelligence is delivered and consumed, subject to its actual performance and trust. In this week's @terrawatchspace newsletter, I analyse what this advancement means for both EO downstream companies & EO satellite data providers and how they can adapt.

Qwen3.8-Max is the best object detection VLM - satellite images - infrared images - documents - techical drawings - hand-sketched diagrams - crowded scenes - small objects ↓ check out some of those insane examples

Become a paid subscriber and get weekly market briefings. newsletter.terrawatchspace.c…

We have had quite a bit of consolidation in Earth observation over the past 5 years - we track at least 35 M&A deals in EO across the value chain from Data to Processing to Intelligence. These mainly include two types of acquisitions covering horizontal consolidation and vertical consolidation. Horizontal consolidation is when two similar EO businesses join up, at sensor or segment level - an EO satellite company buying another with a different sensor for synergies or growth, or an EO company moving down the value chain and buying an analytics company. Vertical consolidation is when a user of EO solution buys that company for strategic purposes and own that capability as its proprietary technology - think about a market leader in a specific vertical acquiring an EO downstream solutions firm, or an EO data company buying one to double down on a vertical. Analysing consolidation in EO over the past 5 years shows an interesting trend on why these deals happens and what to expect more of this year. We covered this in the latest edition of the TerraWatch Space Pro newsletter.

Interesting times in the Earth observation sector as we had two major acquisition deals announced recently. This was not about EO for defense for once, but rather the enterprise EO market - which I keep hammering on about. 📈 Verisk, a major insurance analytics and catastrophe modeling firm acquired EO-based intelligence provider McKenzie Intelligence Services for an undisclosed sum (likely low tens of millions as Verisk said it won't impact their financials). This should enable easier ingestion of EO insights into their underwriting and claims ecosystem. ⚡ Schneider Electric, a leading player in the energy sector focused on power distribution and grid infrastructure acquired EO-based vegetation monitoring solution provider AiDASH at an enterprise value of $350M. This would integrate satellite-based monitoring and wildfire risk assessment into their asset management software. Two quick thoughts: 1. These are textbook examples of vertical consolidation: when a major industry leader acquires a solution provider (usually a supplier) as they come to realise the technology - EO in this case - is strategic to their business. In both cases the acquirer had a working relationship: Schneider was a strategic investor and Verisk was a customer. While it's taken a while, industries are starting to understand that owning a proprietary EO-based solution is key to their growth and to support that, EO application companies are beginning to mature, both technically and commercially. This trend should continue and I am confident we will see more deals. 2. The exits of two EO downstream businesses - one start-up, MIS (40 ppl) and one scale-up, AiDASH (400 ppl) - are good news for the EO sector. It shows there is genuine opportunity for companies that are focused on solving and scaling specific problems with EO. But it also showcases another thing: we have had at least 6 EO companies acquired in the past year - some by industry leaders, some by space companies. The EO sector does not have many growing downstream companies left. Makes me wonder: Can there ever be actual, growing downstream businesses or, will they all be acquired by major industry players before they can fully scale? Will AI potentially change the game? A lot to cover in my next deep-dive on @terrawatchspace!