Meissa Spatial AI is moving from commercial construction sites into South Korea’s defense aerospace sector through a new project with Korea Aerospace Industries, or KAI, to develop photorealistic 3D terrain maps for military flight simulators.
The South Korean spatial intelligence company said it has begun developing a Spatial AI 3D Map Solution for KAI’s flight simulation systems. The roughly two-year project is scheduled to begin in the second half of 2026 and will cover simulators for fixed-wing aircraft, rotary-wing aircraft and unmanned aerial vehicles.
Under the project, Meissa will develop an integrated system for creating, updating and delivering large-scale 3D terrain environments based on real-world imagery.
The project is significant because it transfers technology already deployed commercially in construction into a defense training environment. Meissa says its spatial technology has been used across hundreds of construction sites, while its platform is deployed by major Korean builders. The company’s own website describes its technology stack as combining drone and satellite data collection, proprietary 3D modeling and computer vision AI.
The partnership also reflects a wider shift in military simulation. As aircraft systems become more sophisticated, realistic terrain is becoming an increasingly important part of virtual training and mission rehearsal.
Meissa Spatial AI Moves Into Defense
Meissa’s technology was initially commercialized around real-world industrial environments rather than military simulation.
The company developed a platform that uses drones to capture physical sites and converts the collected imagery into measurable digital environments.
Its construction platform combines automated drone scanning, 3D mapping and AI-based analysis to support functions including construction progress, quantity measurement, quality management and safety monitoring.
According to Meissa, its construction SaaS has been deployed at more than 300 construction sites.
The new KAI project takes the underlying capability — reconstructing real physical spaces digitally — and applies it to a very different problem.
Instead of creating a digital representation of a construction site for project management, Meissa will create large virtual environments that pilots can fly through inside simulators.
The fundamental technology is related, but the scale and operational requirements are substantially different.
Military simulation needs terrain that is visually convincing, geographically accurate, continuously updateable and capable of being delivered efficiently to multiple simulation systems.
Why Terrain Matters in Flight Simulators
A flight simulator is more than a digital cockpit.
Accurate aircraft controls and avionics are essential, but pilots also rely heavily on the outside environment.
Terrain influences navigation.
Buildings, mountains, roads and other landmarks help pilots determine position and direction.
Terrain also affects how pilots approach targets, select flight paths and react to threats.
For helicopter and UAV operations at lower altitude, individual structures and terrain features can become particularly important.
A simulator with accurate aircraft behavior but unrealistic surroundings therefore provides an incomplete representation of an actual mission.
Meissa argues that the value of modern flight simulation increasingly depends on the realism of its terrain database.
Older simulation environments could rely more heavily on simplified representations of buildings and terrain.
Modern systems are increasingly expected to reproduce surface detail, textures and environmental characteristics much more closely.
That creates demand for higher-resolution mapping technology.
Satellite and Drone Data Serve Different Missions
One of the most interesting elements of Meissa Spatial AI is its combination of satellite and drone imagery.
The two data sources serve different purposes.
Fixed-wing aircraft can travel over enormous areas at relatively high altitude.
A simulator for these aircraft may therefore need terrain covering an entire country or large operational region.
Capturing that amount of territory entirely with drones would be impractical.
Satellite imagery provides a more suitable foundation for large-scale geographic coverage.
Helicopters and UAVs operate differently.
They may fly at low altitude between buildings or around complex terrain.
For those missions, individual structures, surface details and smaller geographic features become much more important.
Drone imagery can provide significantly more detailed local information.
Meissa plans to combine these two layers rather than choosing one source for every application.
Building a High-Resolution World Model
Combining satellite and drone data is not as simple as placing one image on top of another.
The sources have different resolutions, perspectives and error characteristics.
They therefore require separate reconstruction pipelines before they can be aligned.
Meissa says its system will integrate both types of data into a common coordinate framework.
The result is intended to be a single high-resolution World Model where detailed operational zones connect with much larger geographic environments.
For example, a simulator could use satellite-derived terrain across a broad operational area while incorporating drone-derived detail around a strategically important city, airfield or facility.
The pilot would experience the environment as one continuous world rather than separate datasets.
This hybrid approach could make it possible to balance geographic scale with local visual detail.
3D Reconstruction Is Core to the Project
The project relies heavily on Meissa’s 3D reconstruction capabilities.
Photographs and satellite images are fundamentally two-dimensional.
To create an environment that a simulator can navigate, those images need to be transformed into three-dimensional geometry.
Meissa says it has developed its own 3D reconstruction engine for this purpose.
Its commercial platform already converts drone imagery from construction sites into measurable 3D environments.
The defense project expands that process to a much larger scale.
Instead of reconstructing one construction site, the system may need to process broad terrain areas containing cities, roads, mountains and infrastructure.
Automation becomes increasingly important as scale increases.
A process that requires extensive manual modeling may work for a small simulation area but becomes costly when terrain needs to cover large regions and remain current.
AI Can Reduce Manual Mapping Work
This is where AI becomes relevant.
Spatial AI combines machine perception with geographic and three-dimensional information.
Computer vision can identify structures and patterns within imagery.
Reconstruction algorithms can infer geometry from multiple observations.
AI can then assist in processing and interpreting the resulting environment.
Meissa describes its core capabilities as data collection and integration, 3D modeling and Vision AI.
The company’s broader objective is to convert physical environments into quantitative digital information that software systems can use.
For flight simulation, automation could reduce the amount of manual work required to construct and maintain virtual terrain.
That becomes especially important when the maps need frequent updates.
Military Terrain Cannot Remain Static
Real-world environments continuously change.
New buildings appear.
Roads are modified.
Industrial facilities expand.
Military installations change.
Urban redevelopment can alter entire districts.
A flight simulator using old geographic data may therefore represent an environment that no longer exists.
For basic aircraft handling, that may not always matter.
For mission rehearsal, it can matter considerably.
Pilots preparing for operations in a particular region need a virtual environment that resembles current conditions.
Meissa says terrain map update cycles are becoming shorter as simulation requirements increase, shifting from annual updates toward updates measured in months.
Automating the reconstruction pipeline could therefore become as important as the visual quality of the original map.
A highly realistic map that takes years to update would have limited value for rapidly changing operational environments.
Mission Rehearsal Raises the Requirements
The KAI project goes beyond general flight training.
Meissa says the system is intended to support mission rehearsal, where pilots can practice missions in environments resembling actual operational areas.
That increases the importance of geographic fidelity.
A pilot may need to practice approaching a specific location.
Terrain may affect radar visibility, navigation and route planning.
Urban structures can influence low-altitude operations.
A realistic virtual environment allows crews to become familiar with these conditions before flying an actual mission.
Mission rehearsal is particularly relevant for scenarios that are difficult or dangerous to reproduce with real aircraft.
Training for emergencies, hostile environments or specialized missions can expose crews and equipment to unnecessary risk if conducted entirely in real flight.
Simulation allows those scenarios to be repeated.
Simulators Can Reduce Training Constraints
Real aircraft are expensive training platforms.
Every flight consumes fuel and contributes to maintenance requirements.
Aircraft availability can be limited.
Weather can interfere with schedules.
Some emergency scenarios cannot be safely reproduced in actual flight.
Simulation provides a way around many of these limitations.
Pilots can repeat difficult procedures without consuming aircraft flight hours.
They can practice failures that would be dangerous to create deliberately in a real aircraft.
Multiple crews can also train against the same scenario.
For military operators, simulators therefore become part of the aircraft system rather than an optional training accessory.
Meissa’s announcement argues that simulator quality increasingly contributes to the competitiveness of aircraft programs themselves as customers demand more advanced training systems alongside the aircraft they purchase.
KAI Could Internalize More Simulator Technology
The partnership also has a strategic technology dimension for KAI.
KAI has previously indicated that it wants to combine Meissa’s automatic 3D map generation technology with its flight simulator hardware.
Reporting around KAI’s investment in Meissa described the goal as internalizing a simulator digital-twin solution capable of producing more realistic virtual training environments.
This could reduce dependence on externally sourced software and mapping technologies.
Defense systems often remain in service for decades.
Long-term dependence on foreign software can create issues involving licensing, upgrades, cybersecurity and technical support.
An internally controlled mapping pipeline gives KAI greater flexibility to maintain and adapt simulation environments.
Meissa specifically says its project is intended to reduce the licensing burden and operational uncertainty associated with foreign commercial engines and manual update processes.
Streaming Technology Supports Multiple Simulators
Creating a large 3D world is only part of the technical problem.
The system also needs to deliver that environment efficiently.
High-resolution 3D terrain can contain enormous amounts of data.
Loading an entire geographic model locally onto every simulator would create storage and update challenges.
Meissa plans to use 3D spatial streaming technology to support simultaneous connections from multiple simulators.
Streaming allows relevant geographic information to be delivered as needed rather than requiring every piece of the world model to be loaded at once.
The concept is similar to technologies used in large-scale digital maps and online virtual worlds.
For military simulation, however, reliability and security requirements can be considerably stricter.
The ability to serve multiple simulators from a common terrain system could also simplify updates.
Instead of manually updating individual systems, new terrain data can potentially be distributed through a centralized pipeline.
Sensors Need Their Own Virtual World
Human vision is not the only perspective that matters in modern military aviation.
Aircraft increasingly rely on electro-optical and infrared sensors.
A realistic training environment therefore needs to reproduce how terrain appears through those systems as well as through a cockpit window.
Meissa says the planned platform will support both daytime and nighttime environments along with infrared and electro-optical sensor simulation.
That capability is relevant for aircraft, helicopters and unmanned systems conducting surveillance or targeting missions.
An object that is easily recognizable in daylight may appear very different through an infrared sensor.
Pilots and operators need experience interpreting those representations.
A high-quality terrain model can therefore serve multiple visual and sensor channels within the same training environment.
UAV Simulation Is Becoming More Important
The inclusion of unmanned aircraft is particularly notable.
Military UAVs are becoming increasingly important across reconnaissance, surveillance, communications and combat operations.
Unlike traditional aircraft training, UAV simulation may involve both aircraft control and sensor operation.
Operators can need to navigate terrain while simultaneously interpreting camera or infrared imagery.
Low-altitude UAV missions can also demand very detailed environmental models.
Buildings, roads and individual terrain features may directly influence route planning.
Drone-derived mapping is well suited to this requirement because it can capture local areas at much higher resolution than broad satellite imagery.
The combination of detailed drone data and broad satellite coverage could therefore become particularly useful for unmanned aircraft mission rehearsal.
Meissa Has Already Proven the Technology Commercially
One reason the KAI partnership stands out is that Meissa is not developing its spatial technology exclusively for defense.
Its underlying system has already been commercialized.
The company’s construction platform uses drones and proprietary 3D mapping technology to digitize job sites and analyze the resulting data.
Meissa says its SaaS has been deployed across more than 300 construction sites and is used by major Korean construction companies.
Its website identifies functions including automatic site scanning, automated analysis, construction progress monitoring, quantity management and safety oversight.
That commercial experience provides an important technical foundation.
Construction sites are large, complex and continuously changing.
Capturing them repeatedly requires scalable data processing and reliable 3D reconstruction.
Those requirements overlap with some of the challenges involved in maintaining simulation terrain.
Defense Brings New Requirements
Commercial success does not mean the technology can be transferred directly into defense without modification.
Military systems introduce additional requirements.
Security is one of the most obvious.
Geographic datasets associated with operational areas may be sensitive.
Systems need controlled access and secure data management.
Availability is another concern.
Training infrastructure may need to function in restricted networks without relying on ordinary public cloud services.
Performance requirements can also differ.
A construction platform may allow users to examine a static 3D model.
A flight simulator needs to render terrain continuously as an aircraft moves through the environment.
Latency, frame rate and streaming performance therefore become critical.
Meissa says its focus will be adapting technology developed across hundreds of commercial sites to the operational and security requirements of defense.
KAI Is More Than a Customer
The relationship between the two companies extends beyond this individual project.
KAI is a strategic investor in Meissa.
In late 2025, Meissa completed a KRW 9.7 billion pre-IPO financing round, bringing cumulative investment at the time to approximately KRW 35 billion. KAI was identified as Meissa’s second-largest shareholder and had invested roughly KRW 8 billion.
Meissa’s latest company description says cumulative investment has since reached KRW 37 billion and identifies KAI as both a major shareholder and strategic partner.
This relationship helps explain why cooperation is expanding across several technology areas.
KAI provides aerospace hardware and defense-system expertise.
Meissa provides spatial data processing, satellite imagery, drone mapping, computer vision and 3D reconstruction.
The combination allows the companies to pursue digital technologies that sit between physical aerospace platforms and software.
Satellite Technology Creates Another Connection
The simulator project is not the only area where Meissa and KAI overlap.
Meissa has also been developing satellite data processing technology.
Previous reporting on KAI’s investment described a satellite operations and preprocessing solution designed to collect and analyze imagery produced by KAI’s next-generation medium and small satellite programs.
This creates an interesting technical connection.
Satellite imagery can support both geospatial intelligence and simulator terrain generation.
A company capable of processing satellite imagery at scale can potentially reuse parts of that technology across multiple applications.
For Meissa, defense simulation therefore sits within a broader spatial-data strategy rather than representing an isolated product.
Its business increasingly spans construction, public-sector applications, satellite imagery and defense.
Spatial AI Is Becoming an Industrial Layer
The project also illustrates the expanding meaning of AI.
Public discussion often focuses on generative AI systems that produce text, images or software.
Spatial AI addresses a different problem.
It helps machines understand and reconstruct physical environments.
This can involve computer vision, mapping, localization, geospatial analysis and 3D modeling.
The resulting digital representation can then be used by humans, robots, drones or simulation systems.
Meissa describes its mission as recreating real-world spatial information in a quantitative digital environment so that different industries can make better decisions.
Construction is one application.
Military simulation is another.
Robotics, autonomous vehicles, disaster response and infrastructure management can potentially use similar spatial intelligence.
Digital Twins Extend Beyond Factories
Digital twins are often associated with factories and industrial machinery.
But a digital twin can also represent geographic environments.
In a flight simulator, the objective is not merely to display attractive scenery.
The virtual environment needs to preserve meaningful relationships between terrain, structures and geographic coordinates.
This is particularly important when training is connected to a real operational area.
The closer the virtual environment corresponds to physical reality, the more useful it can become for planning and rehearsal.
KAI’s previous statements around Meissa have explicitly connected automatic 3D mapping with the internalization of simulator digital-twin technology.
That makes the current project part of a wider aerospace digitalization strategy.
Domestic Technology Could Reduce Foreign Dependence
For South Korea’s defense industry, domestic software capability has strategic importance.
The country has developed increasingly competitive aerospace and defense hardware.
But modern defense platforms depend heavily on software.
Simulation, AI, digital twins, satellite processing and autonomous systems are becoming increasingly important parts of the overall product.
If key software remains dependent on overseas vendors, domestic hardware capabilities alone may not provide full technological independence.
Meissa’s project addresses one part of that issue.
By developing the terrain generation and service pipeline with domestic technology, KAI could gain greater control over updates, customization and integration.
This does not mean every component of the final simulator will necessarily be domestically sourced.
But expanding local software capability can reduce dependencies in strategically important areas.
Better Simulation Could Support Aerospace Exports
Simulation technology may also influence KAI’s export competitiveness.
Aircraft customers do not purchase only the physical aircraft.
Training systems, maintenance, logistics and software support are all part of the package.
A sophisticated simulator can help an overseas customer train pilots without consuming large amounts of real aircraft flight time.
If the simulation environment can also be customized to reproduce the customer’s own territory, its value increases further.
Meissa’s satellite-and-drone approach could potentially support that type of customization.
Satellite data can provide broad geographic coverage, while drone imagery can add detailed areas when appropriate and available.
A reusable mapping pipeline could make it easier to create different training environments for different customers.
That possibility is particularly relevant as KAI expands its presence in international aerospace markets.
The Two-Year Development Period Will Be Important
The project is scheduled to run for roughly two years beginning in the second half of 2026.
During that period, Meissa will need to demonstrate that technology developed for commercial spatial analysis can satisfy aerospace simulation requirements.
Several technical challenges will matter.
Large geographic datasets must be reconstructed accurately.
Satellite and drone models need to align.
Terrain must stream efficiently.
Multiple simulator systems may need simultaneous access.
Visual environments need to support different times of day and sensor types.
Updates must be manageable without rebuilding the entire environment manually.
Security requirements also need to be incorporated.
The result will provide a clearer indication of whether Meissa can successfully expand from commercial spatial AI into defense software.
Meissa Spatial AI Opens a New Defense Market
The KAI project represents more than a conventional software contract.
It demonstrates how technology developed for one physical industry can migrate into another when the underlying problem is similar.
Construction companies need accurate digital representations of changing real-world sites.
Military flight simulators need accurate digital representations of changing real-world terrain.
Both problems require data collection, spatial alignment, 3D reconstruction and efficient visualization.
Meissa has spent years building those capabilities around drones, satellite imagery and computer vision.
KAI now wants to apply them to fixed-wing, helicopter and UAV simulation.
The project remains in development, so its operational performance has not yet been demonstrated.
Claims about improvements in training realism and reduced dependence on foreign software will ultimately need to be evaluated after deployment.
Still, the direction is clear.
Meissa Spatial AI is moving beyond construction technology into aerospace and defense, while KAI is seeking to combine increasingly sophisticated software with its aircraft and simulator hardware.
If the two-year program succeeds, the result could give KAI a domestically developed pipeline for converting satellite and drone imagery into continuously updated 3D training environments.
For Meissa, it would establish a new commercial use for technology already tested across hundreds of physical construction sites.
And for South Korea’s defense industry, it provides another example of AI expanding from standalone software into the digital infrastructure surrounding aircraft, drones and next-generation training systems.
