There was a time when remote sensing was mostly about capturing imagery and letting a human analyst figure out what was in it. That time is over. The convergence of machine learning, cloud compute, and sensor technology has turned remote sensing from a data collection discipline into an automated intelligence pipeline — one where AI models classify, extract, and interpret geospatial data at speeds and scales that were impossible five years ago.

The Compute Problem Behind Every Pixel

Where the Industry Is Heading

Managing the Cost of Geospatial AI

The post When Remote Sensing Meets AI: The Compute Cost Nobody Is Talking About appeared first on Spatial Tech.

The post When Remote Sensing Meets AI: The Compute Cost Nobody Is Talking About appeared first on Spatial Tech.

The post Geospatial vs Geographic Data: Understanding the Distinction That Shapes Every GIS Workflow appeared first on Spatial Tech.

The post Geospatial vs Geographic Data: Understanding the Distinction That Shapes Every GIS Workflow appeared first on Spatial Tech.

The post Geospatial Digital Twins Guide 2026 | 3D Data, Standards & Open Data appeared first on Spatial Tech.

The post Geospatial Digital Twins Guide 2026 | 3D Data, Standards & Open Data appeared first on Spatial Tech.

The post Smart Sensors Guide 2026 | IoT for Geospatial & Infrastructure appeared first on Spatial Tech.

The post Smart Sensors Guide 2026 | IoT for Geospatial & Infrastructure appeared first on Spatial Tech.

Richard Andersson is a Partner at NodeNordic, a Stockholm-based AI consultancy that works with organisations across the Nordics to implement automation and AI-driven workflows. His background spans industrial engineering at KTH Royal Institute of Technology and over eight years in software development at firms including Forefront and Solita before co-founding the consultancy practice. We spoke with him about how AI automation is being adopted across data-intensive industries, where the real efficiency gains are, and what geospatial organisations should consider before investing in automation infrastructure.

The automation opportunity in data-intensive industries

Spatial Tech: You work with clients across a range of industries. Where are you seeing the strongest demand for AI-driven automation right now?

Richard Andersson: The pattern is consistent across sectors — organisations that process large volumes of structured or semi-structured data are reaching a point where manual workflows simply cannot scale. We see this in financial services, in logistics, in public sector administration, and increasingly in industries that deal with spatial and environmental data. The common thread is that these organisations have data pipelines that were designed for a different era of volume and complexity. The tools worked when you were processing hundreds of records or images. When that becomes hundreds of thousands, the economics break down unless you automate.

Spatial Tech: How does that translate specifically to geospatial workflows?

Richard Andersson: Geospatial is a fascinating case because the data is inherently complex — you are dealing with multi-dimensional datasets, temporal layers, coordinate reference systems, and outputs that need to be both analytically rigorous and visually communicable. Many of the organisations we speak with are still running semi-manual classification pipelines for satellite imagery, or using rule-based systems for feature extraction that were state of the art a decade ago but are now limiting throughput. The opportunity is not just in making existing processes faster — it is in enabling entirely new categories of analysis that were not economically viable before.

Where automation delivers real value

Spatial Tech: Can you give a concrete example of where AI automation has fundamentally changed an operational workflow for a client?

Richard Andersson: Without naming specific clients, I can describe a pattern we see regularly. A municipality or infrastructure operator has a requirement to monitor asset conditions across a geographic area — roads, utilities, vegetation encroachment, building envelopes. Traditionally this involves periodic manual surveys, field inspections, and a lot of spreadsheet-based reporting. What we help organisations do is build automated pipelines where drone or satellite imagery feeds into classification models that flag anomalies, generate condition reports, and route maintenance priorities into existing operational systems. The human expertise shifts from data processing to decision-making, which is where it should have been all along.

Spatial Tech: That sounds like it requires significant upfront investment in model development.

Richard Andersson: It used to. Two years ago, building a custom classification model for a specific use case required months of labelled training data preparation, specialised ML engineering, and considerable compute resources. Today, foundation models and transfer learning have compressed that timeline dramatically. As an AI konsult Stockholm, we are seeing projects that previously required six months of development reach production readiness in six to eight weeks. The tooling has matured to the point where the bottleneck is no longer model development — it is organisational readiness to integrate automated outputs into existing decision processes.

The Nordic advantage in applied AI

Spatial Tech: Why do you think the Nordic region has become particularly active in applied AI and automation?

Richard Andersson: Several factors converge here. The Nordics have high labour costs, which creates a strong economic incentive to automate repetitive tasks. There is a deep tradition of engineering excellence and a workforce that is comfortable with technology adoption. The public sector in Sweden, Norway, and Finland is relatively progressive about digital transformation compared to many European counterparts. And there is a strong university pipeline — KTH, Chalmers, Aalto, NTNU — producing engineers who understand both the theoretical foundations and the practical implementation challenges.

For geospatial specifically, the Nordics have additional advantages. Sweden and Finland have extensive public geographic data infrastructure through Lantmäteriet and the National Land Survey of Finland. Norway has invested heavily in maritime and offshore spatial data through organisations like the Norwegian Mapping Authority. This means there is a rich foundation of spatial data that is accessible, well-documented, and ready to be enhanced with AI-driven analysis.

Spatial Tech: How do you see the relationship between AI consultancies and geospatial technology providers evolving?

Richard Andersson: I think we are moving towards a model where geospatial technology companies provide the data infrastructure and domain-specific tools, while AI consultancies help organisations build the automation layers on top. The geospatial vendors understand coordinate systems and projection mathematics and sensor calibration better than anyone. What they sometimes lack is deep expertise in production ML systems — model monitoring, drift detection, automated retraining pipelines, integration with enterprise IT infrastructure. That is where firms like ours add value. The best outcomes happen when domain expertise and ML engineering expertise work together rather than one trying to replace the other.

What geospatial organisations should consider

Spatial Tech: For geospatial organisations considering their first AI automation project, what advice would you offer?

Richard Andersson: Start with a workflow that is genuinely painful and genuinely repetitive. Do not start with the most complex or most visible project — start with the one where your team spends the most time doing work that does not require human judgment. Classification tasks, data validation, report generation, anomaly detection in time-series data — these are all excellent starting points because the success criteria are clear and the impact on team productivity is immediate and measurable.

The second thing I would say is to be realistic about data quality. Every AI project ultimately depends on the quality of the input data. If your imagery archive has inconsistent metadata, or your asset database has gaps, fix those problems first. Automating a broken process just gives you broken outputs faster.

And the third point is to think about integration from day one. The most impressive classification model in the world is useless if its outputs cannot flow into your existing GIS platform, your asset management system, or your reporting workflow. The integration architecture matters as much as the model architecture.

Spatial Tech: Richard, thank you for your time and insights.

Richard Andersson: Thank you. It is an exciting time for anyone working at the intersection of spatial data and AI — the capabilities are advancing faster than most organisations realise, and the opportunities for those who move early are significant.

The post How AI Automation Is Reshaping Geospatial Workflows: A Conversation with Richard Andersson appeared first on Spatial Tech.

The post How AI Automation Is Reshaping Geospatial Workflows: A Conversation with Richard Andersson appeared first on Spatial Tech.

The expansion of AI workloads is creating massive demand for physical infrastructure — and the companies that build electrical systems are scaling up to meet it. Siemens recently announced more than $165 million in new manufacturing investments across North and South Carolina, adding facilities and hundreds of jobs specifically aimed at producing the electrical equipment that powers data centres and AI factories.

The investment is a clear signal of where the infrastructure bottleneck sits right now. While much of the AI conversation focuses on chips, models, and software, the physical layer — power distribution, switchgear, busway systems, and prefabricated electrical assemblies — is where capacity constraints are increasingly felt. Data centres consume enormous amounts of electricity, and the equipment needed to deliver, manage, and protect that power supply has to be manufactured, assembled, and installed before any server rack goes live.

What Siemens is building

The Carolina investments include multiple new and expanded facilities. In Raleigh, North Carolina, a new 131,000 square foot facility will assemble integrated power delivery solutions — prefabricated systems designed to reduce on-site installation time for data centre operators. In Wendell, also in North Carolina, a new site will localise production of medium voltage protection and automation devices, while an existing facility expands switchgear production capacity.

In South Carolina, a new facility in Spartanburg will handle lighting panel production and distribution, while an expanded site in nearby Roebuck increases busway manufacturing capacity with new paint, epoxy, and plating lines. Across both states, the investments are expected to create around 350 new manufacturing jobs.

These projects add to nearly $700 million Siemens has committed in recent years to expanding U.S. electrical manufacturing capacity, including facilities in California and Texas.

Why electrical infrastructure is the AI bottleneck

The narrative around AI infrastructure tends to focus on semiconductor supply and GPU availability. But the physical systems that deliver power to those chips are just as critical — and in many cases harder to scale quickly. A data centre is ultimately an electrical facility. Every rack of GPUs requires reliable power delivery from the grid through multiple layers of transformation, distribution, switching, and protection equipment.

As AI workloads grow more power-intensive — with individual training clusters now consuming tens of megawatts — the demand for electrical infrastructure is scaling in parallel. Prefabricated and modular power delivery systems, like those Siemens is producing in its new Raleigh facility, are becoming increasingly important because they allow data centre operators to bring capacity online faster than traditional site-built approaches.

Siemens positions its data centre and AI infrastructure portfolio as a chip-to-grid solution — covering simulation, automation, cooling optimisation, and electrical distribution as an integrated stack rather than a collection of individual components.

The workforce dimension

Manufacturing capacity alone does not solve the problem if there are not enough skilled workers to operate the facilities and install the equipment. Siemens is addressing this through a workforce development initiative called Careers Electric, launched in North Carolina with a $9.25 million investment from the Siemens Foundation. The programme is designed to expand access to electrical training and create pathways into manufacturing and installation careers — roles that are essential to actually delivering the infrastructure that AI depends on.

What this signals for the broader market

Siemens is not the only company scaling up electrical manufacturing for data centres, but the size and specificity of these investments reflect how confident major industrial players are that AI-driven power demand is not a short-term cycle. The company reported record levels of data centre-related electrical equipment orders, and these new facilities are a direct response to that demand.

For the smart infrastructure sector more broadly, the pattern is clear: AI is not just a software story. It is an industrial story — one that runs through power grids, manufacturing plants, and the physical systems that keep data centres operational. The companies that can build and deliver that infrastructure at speed will capture a significant share of the value being created by the AI buildout.

Around 620 words. Category: Smart Infrastructure. The Siemens link is embedded naturally in the middle of the post as a contextual reference to their data centre portfolio. Want me to draft the next two posts?

The post Siemens Bets Big on U.S. Manufacturing to Meet Surging AI Data Centre Demand appeared first on Spatial Tech.

The post Siemens Bets Big on U.S. Manufacturing to Meet Surging AI Data Centre Demand appeared first on Spatial Tech.

The issue is straightforward. A machine learning model trained to classify agricultural crops using Sentinel-2 satellite imagery will perform well over farmland. But apply that same model to an area of open water, and it will generate nonsensical outputs — confidently labelling ocean pixels as wheat fields or vineyards. For an experienced GIS analyst, this kind of error is easy to spot and correct. But in automated workflows where no human is reviewing intermediate results, these failures can propagate silently through entire analysis pipelines.

This is not a theoretical concern. Research has shown that even well-regarded models like BigEarthNet, trained on Sentinel-1 and Sentinel-2 data, can swing from over 85 percent accuracy in optimal conditions to as low as 20 percent in unfavourable scenarios. The gap between best-case and worst-case performance is enormous, and most users have no way of knowing which end of that spectrum they are operating at for any given query.

The documentation problem

Compounding this reliability issue is a documentation gap. Models published on platforms like HuggingFace and Kaggle are often poorly documented — at least not in a machine-readable format that a processing platform could use to automatically validate whether a model is appropriate for a given dataset and region. In practice, this means users need to manually inspect model specifications, preprocess input data with custom Python scripts, and make judgement calls about applicability. That effectively limits the use of these models to specialists with both domain expertise and programming skills.

For geospatial AI to scale beyond expert users, platforms need to handle this validation automatically.

Model fencing as a solution

A research collaboration between Constructor University and rasdaman GmbH in Bremen, Germany, funded by the EU’s EFRE programme, is working on exactly this problem. The project, called FAIRgeo, introduces a concept called “model fencing” — automatically restricting AI model inference to the spatial, temporal, and thematic contexts where reliable results can be expected.

The approach works by enriching model metadata with machine-readable information about where and when a model is valid. When a user submits a query, the platform checks parameters automatically before execution: correct satellite source, correct spectral bands, appropriate patch size, and geographic applicability. If the model is being asked to operate outside its validated comfort zone, the system can flag the issue or prevent execution entirely.

Early results are promising on the usability front as well. What typically requires over a hundred lines of Python code can be reduced to a two-line datacube query, with the platform handling data selection, preparation, and tiling automatically. Performance benchmarks also show the integrated approach running faster than traditional Python implementations in most cases.

Why this matters beyond research

As AI becomes embedded in operational geospatial workflows — from agricultural monitoring to urban planning to climate risk assessment — the consequences of unreliable model outputs grow more serious. Decisions about land use, disaster response, and infrastructure investment increasingly depend on automated analysis of satellite data.

The geospatial industry needs standardised approaches to model validation and applicability metadata. Efforts like FAIRgeo, which is contributing its findings to OGC working groups on data quality and coverage standards, point toward a future where AI on Earth observation data is not just more powerful, but meaningfully safer.

The post Why AI Models Need Guardrails When Applied to Earth Observation Data appeared first on Spatial Tech.

The post Why AI Models Need Guardrails When Applied to Earth Observation Data appeared first on Spatial Tech.

Enterprise campus networks are under pressure. The combination of growing bandwidth demands, increasing device density, operational technology integration, and sustainability targets is forcing organisations to rethink how their physical network infrastructure is built. Nokia and Alcatel-Lucent Enterprise are responding with a deepened strategic alliance that combines optical fibre technology with secure campus networking — targeting the kinds of environments where connectivity failures have real operational consequences.

The partnership, now in its fifth year, integrates Nokia’s Optical LAN fibre infrastructure with ALE’s enterprise networking solutions for in-building and campus connectivity. The result is a converged fibre-based architecture capable of carrying multi-gigabit data speeds across complex facilities while reducing energy consumption and total cost of ownership compared to traditional copper-based network designs.

Why fibre is gaining ground in campus environments

Most enterprise campus networks still run on copper cabling for the last segment of connectivity — from switch closets to end devices. This architecture has served well for decades, but it is increasingly strained by the demands of modern campus operations. IoT sensor networks, high-density WiFi, CCTV systems, building management platforms, and operational technology applications all compete for bandwidth on infrastructure that was designed for a simpler era.

Fibre-to-the-edge architectures eliminate many of these constraints. Optical fibre supports significantly higher bandwidth over longer distances, requires fewer intermediate network layers, and consumes less energy than equivalent copper deployments. For large campus environments — hospitals, resorts, logistics facilities, transport hubs — the reduction in physical infrastructure also translates to meaningful space savings.

The Nokia-ALE approach consolidates what would traditionally be separate network layers into a single fibre backbone. At Ikos Resorts in Greece, for example, the combined solution runs guest WiFi, CCTV, voice communications, and building safety sensors through one converged high-availability architecture — replacing what would previously have required multiple parallel network infrastructures.

The operational technology angle

What makes this partnership particularly relevant for critical infrastructure is the operational technology integration layer. Modern logistics facilities, manufacturing plants, and transport networks increasingly depend on automated systems that require deterministic, low-latency connectivity. Automated warehouse systems, robotic material handling, real-time asset tracking, and industrial control systems all need network infrastructure that is not just fast but reliably available.

ALE’s contribution to the partnership includes automated device onboarding, asset discovery and classification, virtual network segmentation, and continuous monitoring — capabilities designed to handle the complexity of environments where hundreds or thousands of connected devices need to be securely managed without manual intervention. Virtual segmentation is particularly important in mixed-use environments where IT traffic and operational technology traffic need to coexist on the same physical infrastructure without interfering with each other.

Deployment track record

The partnership has now been deployed across more than 100 enterprises globally, spanning hospitality, healthcare, transport, and logistics. Notable projects include Grand Paris Express, Montreal Railways, Pantai Jerudong Hospital in Brunei, and Wembley Park in the UK — all environments where network reliability is not optional and where the consequences of downtime extend beyond inconvenience into safety and operational risk.

What this signals for enterprise networking

The broader trend is clear: enterprise campus networks are converging. The era of separate infrastructure for IT, OT, building management, and security systems is giving way to unified fibre-based architectures that carry everything on a single physical layer. This convergence is driven by economics — fewer network layers means lower cost — but also by operational necessity. Managing five separate network infrastructures across a large campus is unsustainable as device counts and bandwidth demands continue to grow.

For organisations in logistics, healthcare, manufacturing, and transport — sectors where both connectivity and physical infrastructure intersect — the Nokia-ALE model represents the direction enterprise networking is heading: fewer layers, more bandwidth, lower energy consumption, and a single converged platform capable of supporting both IT and operational technology workloads.

The post Nokia and Alcatel-Lucent Enterprise Push Fibre-Based Networks Into Critical Infrastructure appeared first on Spatial Tech.

The post Nokia and Alcatel-Lucent Enterprise Push Fibre-Based Networks Into Critical Infrastructure appeared first on Spatial Tech.

Announced in early 2026, Ericsson’s 5G Advanced location services deliver sub-10 centimetre outdoor accuracy using Real-Time Kinematics technology and sub-one metre indoor precision through integrated indoor positioning solutions — all running natively on 5G Standalone networks without requiring additional sensors, overlay infrastructure, or device-side applications.

Why network-based positioning matters

Traditional positioning approaches face a fundamental architectural limitation: they rely on signals from satellites orbiting roughly 20,000 kilometres above the Earth’s surface. In open-sky conditions, this works well enough. But inside buildings, in dense urban canyons, underground, or in industrial facilities, satellite signals degrade rapidly or disappear entirely. This is precisely where many of the highest-value positioning use cases exist — warehouse automation, hospital asset tracking, manufacturing floor logistics, and indoor navigation.

Previous attempts to solve indoor positioning have typically involved deploying separate infrastructure: Bluetooth beacons, ultra-wideband anchors, WiFi fingerprinting systems, or dedicated sensor networks. Each of these adds cost, complexity, and a separate technology stack that needs to be maintained independently of the primary communications network.

Ericsson’s approach embeds positioning as a core capability of the 5G network itself. If a facility already has 5G coverage — which an increasing number of factories, hospitals, ports, and campuses do — then positioning becomes available as a network function rather than a separate deployment. This is a significant architectural simplification for enterprises that need both connectivity and location services.

The accuracy gap is closing

Sub-10 centimetre outdoor accuracy is notable because it brings network-based positioning into the range previously reserved for survey-grade GNSS receivers and RTK correction services. For applications like autonomous vehicle guidance, precision agriculture, and drone operations, this level of accuracy opens up use cases that were previously impractical without specialised equipment.

Indoor sub-metre accuracy, while less precise than what ultra-wideband systems can achieve, is sufficient for the majority of enterprise use cases: tracking assets across a warehouse floor, monitoring personnel in a hospital, managing equipment across a manufacturing facility, or enforcing geofencing policies across a logistics campus.

The combination of indoor and outdoor positioning within a single unified system is arguably more significant than the raw accuracy numbers. Most real-world operations span both environments — a package moves from an outdoor loading dock to an indoor sorting facility, a patient transfers from an ambulance to a hospital ward, an autonomous vehicle transitions from a public road to an indoor parking structure. Seamless handover between indoor and outdoor positioning without switching between separate technology systems eliminates a longstanding operational pain point.

What this means for the positioning market

Ericsson is positioning this offering as a monetisation opportunity for mobile operators — enabling them to sell precision location as a service to enterprise customers across manufacturing, healthcare, automotive, public safety, and drone operations. The business model shifts positioning from a device-side capability to a network-side service, with operators acting as positioning infrastructure providers rather than just connectivity providers.

For the broader geospatial and positioning industry, the trajectory is clear: 5G networks are becoming positioning infrastructure in their own right, not just communications pipes. As 5G Standalone deployments expand globally, the installed base of network-based positioning infrastructure will grow in parallel — potentially creating a ubiquitous indoor-outdoor positioning layer that complements rather than replaces satellite-based GNSS.

The competitive implications for traditional positioning technology providers — from GNSS receiver manufacturers to indoor positioning system vendors — will become clearer as enterprise adoption scales through 2026 and beyond.

The post Ericsson’s 5G Advanced Location Services Push Positioning Accuracy Below 10 Centimetres appeared first on Spatial Tech.

The post Ericsson’s 5G Advanced Location Services Push Positioning Accuracy Below 10 Centimetres appeared first on Spatial Tech.