Data reaches the ground too late
Downlink, routing, and post-processing add latency when every minute affects response quality.
Orbital AI for urgent response
Process raw data in orbit before delay becomes damage
LeoTrek is building a distributed platform for AI workloads across low Earth orbit satellite clusters, so wildfire monitoring and other time-critical observation missions can move from raw data to action in minutes.
Mission focus
Fast orbital processing for wildfire monitoring and rapid alerts.
Deploy models, fuse sensing modalities, and prioritize urgent insights before full downlink.
In-orbit inference
Region-anchored monitoring
Minutes, not hours
Mission
Close the gap between sensing and action
Earth observation creates enormous value only when it arrives soon enough to change a decision. LeoTrek shifts compute closer to the sensor so satellite missions can support fast-moving events, persistent regional awareness, and new space-native software workflows.
Run urgent inference in orbit
Coordinate AI workloads directly across LEO assets and prioritize what matters most.
Actionable intelligence sooner
Deliver faster alerts, compressed insights, and better use of limited bandwidth.
Beyond wildfires
The same architecture fits disaster response, infrastructure monitoring, and other urgent EO workloads.
Why Now
The cost of waiting is measurable
Wildfires are the clearest example of why orbital compute matters. When insight arrives late, damage compounds across human safety, infrastructure, and climate.
Human and economic toll
Smoke exposure and damaged infrastructure already turn each fire season into a public-health and budget crisis.
Escalating climate events
Larger burn areas and massive CO2 release show these events are compounding, not stabilizing.
Minutes reshape outcomes
Detection latency changes containment options, dispatch quality, and how much land is ultimately lost.
More sensors, stronger onboard compute, and pressure for faster response make this the right moment to move EO intelligence from batch pipelines toward space-native execution.
Platform
Real-time intelligence in space
LeoTrek is designing a space-native orchestration layer for deploying, scheduling, and validating EO workloads across available orbital compute.
01
Distributed EO processing in LEO with performance guarantees
Run inference, filtering, and prioritization directly across orbiting satellite clusters, so urgent compute happens before full downlink.
02
Region-anchored monitoring
Keep continuity over fixed areas of interest and surface time-critical events as new passes occur.
03
Multi-source fusion and orchestration
Coordinate tasks across the constellation and combine thermal, optical, and contextual streams into one decision layer.
04
Cloud-like workflow deployment
Deploy models and workflows through a consistent interface instead of rebuilding software for each mission profile or satellite configuration.
05
Space digital twin
Model mission logic, workload placement, and orbital constraints before deployment so teams can validate latency, throughput, and failure behavior earlier.
06
AI workflow optimization
Prepare inference pipelines with pruning, quantization, and deployment-ready tuning so models fit onboard constraints without custom tooling each time.
Digital Twin
Mission simulation in context
Preview orbit nodes, workflow placement, and computation moving inside the target area before live deployment.
Simulation clock
Current UTC
Iberian Peninsula
Who We Serve
Built for both sides of the EO ecosystem
Satellite operators
Operators with software-defined satellites can use LeoTrek to expose spare in-orbit capacity through a multi-tenant AI-as-a-service model without disrupting primary missions.
- Monetize onboard compute and differentiate with value-added services.
- Offer a lower-friction interface for external AI workloads and multi-tenant access.
- Keep operational safeguards in place through orchestration, admission control, and mission-aware scheduling.
EO service providers and platforms
Service providers that need EO insights fast enough to support operational decisions while satellite data is still time-critical.
- Get detection and mapping results within minutes instead of waiting for full downlink and ground processing.
- Scale AI workflows beyond a single satellite with more predictable latency and execution behavior.
- Combine onboard EO data with third-party sources without routing every workload through cloud pipelines.
LeoTrek connects both sides: EO providers deploy workflows through a cloud-like interface, satellite operators supply the compute layer, and LeoTrek handles orchestration across available capacity.
Use Cases
Built for more than one emergency
Wildfire monitoring is the immediate wedge, but LeoTrek is relevant wherever Earth observation decisions are time-sensitive and bandwidth is scarce.
Wildfire response
Prioritize hotspots, fire spread indicators, and response-relevant updates without waiting for full raw-data transfer.
Disaster monitoring
Support floods, storms, and post-event damage assessment with faster orbital analytics.
Critical infrastructure
Monitor energy, transport, and industrial systems where earlier anomaly detection improves maintenance, planning, and incident response.
Continuous monitoring
Maintain regular visibility over locations of interest and detect meaningful change over time instead of treating every pass as an isolated event.
Research
Grounded in the Edge-Cloud Continuum Research
LeoTrek builds on published work in scheduling, serverless workflows, simulation, cost modeling, and hardware acceleration for AI workloads.
Hyperdrive
Scheduling serverless functions across orbital and ground-adjacent compute layers. IEEE/ACM SEC, 2024.
Read paper
IEEE EDGE 2025
Simulator demo
Stardust
A scalable simulator for multi-layer distributed computing environments. IEEE EDGE, 2025.
Read paperDatabelt
A continuous data path for serverless workflows across distributed compute tiers. Journal of Systems Architecture, 2025.
Read paperCosmos
A cost model for serverless workflows spanning multiple compute layers. IEEE SmartComp, 2025.
Read paperGaia
Hybrid hardware acceleration for serverless AI in distributed compute environments. IEEE/ACM BDCAT, 2025.
Read paperCurrus
A compound AI approach to distributed vehicle trajectory reconstruction. International Conference on the Internet of Things, 2025.
Read paper
Advisors
Schahram Dustdar
Full Professor and Head of the Distributed Systems Group at TU Wien, internationally recognized for his work in distributed systems, Edge-Cloud Continuum, Edge AI, and next-generation computing continuum.
Marie-Françoise Voidrot-Martinez
Expert in Earth Observation, geospatial interoperability, and operational meteorological systems, with leadership roles at the Open Geospatial Consortium and the Group on Earth Observations.
Contact
Open to pilot programs, operator partnerships, and research collaboration
LeoTrek is interested in pilot deployments, operator partnerships, commercial validation, and research collaboration around time-critical orbital intelligence.
LEO operators
Emergency response
Research partners
Investors
Part of the European deep-tech ecosystem
