Artificial intelligence (AI) has become part of nearly every public conversation. For some, it signals automation and disappearing jobs; for others, it promises unprecedented gains in efficiency. The labor market is undeniably shifting fast — roles are evolving, some are fading, and new skill sets are in demand. But beyond disruption, AI is also building entirely new sectors that did not even exist a few years ago.
AI Earth Observation. Image from Canva
One emerging frontier is the so-called agent economy — a model in which tasks are no longer carried out solely by humans, but by AI “agents” that can divide responsibilities, coordinate with one another, cross-check results, and hand off outputs to other systems. Another fast-growing field is Earth Observation, where satellite data, powered by AI, is turned into forecasts, early risk warnings, and decision-ready insights for businesses and public institutions alike.
Why Is This Happening Now?
For a long time, progress was constrained by three factors: limited data availability, immature algorithms, and insufficient computing power. Solutions often had to be rebuilt from scratch for each individual case. A model that worked in one context failed in another; one data type was suitable, another required additional training. This meant lengthy tuning, numerous custom examples, and costly resources — and many ideas remained just ideas.
The landscape has shifted. Satellite data is no longer scarce, technologies have come of age, and next-generation AI models make it possible to deploy solutions at unprecedented speed. What once required months of model training can now, in many cases, be achieved by clearly defining the task, selecting the right data, and validating the outcome.
A straightforward business principle follows: what you can measure, you can manage. And what you can forecast, you can de-risk.
Earth Observation Is Poised for Growth
The space economy is often associated with rockets and satellites. In reality, it has two distinct components. The first is engineering and infrastructure — building satellites, rockets, and organizing launches. The second consists of services that emerge once satellites are in orbit and their signals begin creating everyday value.
According to 2024 estimates, the engineering segment is worth approximately €63 billion, while the services segment reaches about €408 billion. However, nearly the entire services market is dominated by navigation — around 96.2%. Navigation operates in smartphones, vehicles, logistics, agriculture, aviation, and financial time synchronization — which is why it feels “self-evident.” Meanwhile, Earth Observation accounts for only about 1.2% of the services market.
Just 1.2%? Yes.
But precisely for that reason, it is one of the most promising areas for growth. When satellite data, strengthened by AI, begins to transform from mere “images” into forecasts and actionable decision tools, a new economic layer emerges.
Using satellite imagery and artificial intelligence to analyze urban growth. Photo: AIML.
One thing is crucial to understand: AI, on its own, solves nothing. Real value emerges only when technology meets domain expertise. It is experts who know where the true problem lies, what mistakes actually cost, which solutions are feasible in practice, and what complies with regulatory realities.
The biggest breakthroughs happen not where algorithms are strongest, but where technological power converges with deep, hands-on knowledge of the field.
Three Real-World Environmental Applications
One of the clearest examples of AI-powered Earth Observation delivering real-world impact can be found in environmental protection. This is the field where Dr. Valentas Gružauskas, Head of the Artificial Intelligence Methods Laboratory at the Faculty of Mathematics and Informatics at Vilnius University, is putting theory into practice. Together with his team, he develops AI solutions that turn raw satellite data into meaningful indicators, forecasts, and decision-ready insights.
The research team of the AI Methods Laboratory. Photo: AIML.
Peatland Moisture and Fire Risk
Peatlands may look quiet and remote, but their condition directly affects wildfire risk, carbon emissions, and landscape stability. The problem is speed: moisture levels can change rapidly, while ground-based measurements are often too sparse to capture the shift in time. The answer comes from radar satellite data (SAR), which can “see” through clouds and operate day and night. AI then fuses these signals with additional datasets, producing forecasts grounded in data — not intuition.
Cloud Cover and Seeing Beyond It
Optical satellite imagery frequently becomes useless under thick cloud cover. Radar data can bypass this obstacle, but it is far less intuitive to interpret. AI bridges that gap, combining inputs from multiple satellite systems into a coherent analytical picture. The result: informed decisions can be made even when the human eye sees nothing at all.
Urban Heat Islands
Cities trap heat — and understanding where and how that happens is critical for infrastructure planning, energy efficiency, and urban development. Thermal satellite imagery reveals broad heat patterns, but its resolution is often too coarse for street-level decisions. AI enhances that detail, effectively transferring drone-level precision onto large-scale satellite datasets. In the near future, many urban planning decisions may rely on satellite data and AI alone.
When Satellite Analysis Becomes a Conversation
Until recently, working with satellite data meant relying on a rare combination of expertise: professionals who understood environmental or infrastructure systems and could also navigate complex datasets and modelling tools.
Now, a new shift is underway. Satellite analysis is becoming conversational. A user poses a question in plain language, and the system translates it into analytical insights. What once required layers of technical mediation is turning into an interactive dialogue.
This doesn’t side-line experts — it amplifies them. Specialists can get to the heart of an issue faster, while decision-makers gain clearer, more accessible insight into complex realities.
Where Else Is It Applied?
AI-powered Earth Observation isn’t just transforming environmental protection — its reach extends across multiple sectors:
- Energy — monitoring infrastructure and planning for future needs
- Finance and insurance — modelling risks from floods, droughts, and wildfires
- Agriculture and forestry — tracking conditions and measuring sustainability
- Municipal governance and real estate — analysing urban growth and development
The principle is straightforward: AI unlocks possibilities where solutions used to be too costly, too slow, or too complicated to implement.
Will AI Replace People?
Today, technological leverage is so immense that a single professional with deep domain expertise can achieve what once took an entire organization. But there’s a catch: you have to be the one who knows your field better than anyone else.
AI doesn’t “know” what truly matters in your sector. It can process data, spot patterns, and generate forecasts — but the insight, the meaning, comes from humans: those who grasp the real problem, understand the cost of mistakes, know what can realistically be implemented, and navigate the regulatory landscape.
The pressing question isn’t whether AI will replace people. The real question is: which people, harnessing AI, will transform their markets?
Listen to Dr. Valentas Gružauskas’ interview on the LRT program “Labas rytas, Lietuva” from 27:23 in the recording (interview is in Lithuanian language).
About the Author
Dr. Valentas Gružauskas is an Associate Professor and Senior Researcher at Vilnius University and Head of the Artificial Intelligence Methods Laboratory (AIML). His research focuses on deep learning, multimodal machine learning, remote sensing, trustworthy AI governance, risk management, and conformity assessment.
He is also Head of “AI Conformity & Research Consulting,” working on the implementation of the ISO/IEC 42001 standard, practical regulatory compliance, and securing EU funding for private R&D projects. He actively participates in the European Commission’s AI Board and the Lithuanian Standards Department.
More about the laboratory: https://aiml.lt/