AVES reality
Autonomous cars are among the most complex physical AI systems ever built. They must operate safely in open, dynamic environments, yet validating them in the real world remains slow, expensive, and fundamentally limited.
AVES Reality addresses this bottleneck by building a foundational infrastructure layer for autonomous driving: accurate, simulation-ready 3D digital twins of the real world, generated automatically at global scale.
Using satellite and aerial imagery combined with geospatial AI, advanced computer graphics, and physics-aware modeling, AVES Reality reconstructs cities, road networks, and intersections as deterministic digital environments. These are not visual replicas, but procedurally generated worlds enriched with physical and semantic information that autonomous systems can reliably learn from, train in, and be validated against.
xista team
Closing the validation gap in autonomous driving
For years, autonomous mobility has been constrained by the gap between limited real-world testing and the enormous data volumes required to prove safety. Physical testing alone cannot cover rare edge cases, new geographies, or long-tail scenarios in a repeatable way.AVES Reality enables virtual validation through physics-correct, semantically consistent digital twins. Their platform allows customers — including major automotive OEMs — to generate simulation-ready maps of any operational area in minutes instead of months. This enables deterministic testing, repeatable scenario evaluation, and virtual roll-outs in new regions long before physical deployment.The result is a faster, safer, and significantly more cost-efficient path from development to real-world operation.
A base layer for physical AI and world models
As autonomous cars evolve toward higher levels of autonomy, they increasingly rely on world models that go beyond perception. These systems must reason about how objects behave, move, collide, and interact in the physical world. AVES Reality builds exactly this base layer. Their digital twins encode real-world constraints and physical consistency, making them suitable not only for autonomous driving but for a broader class of physical AI systems. In this context, AVES Reality is working with NVIDIA on reference workflows to generate high-fidelity synthetic environments for training and fine-tuning vision-language models and world foundation models. One early application includes intelligent AI agents for traffic management, using NVIDIA’s smart city frameworks. The company has already demonstrated real-world impact in city-scale projects such as Kaohsiung, Taiwan, showing how virtual environments are becoming essential for training physical AI systems under real-world constraints.
Founder vision
Florian Albert, Co-Founder and CEO of AVES Reality, explains: “As next-generation physical AI systems increasingly navigate, reason about, and interact with the real world, the ability to generate rich, deterministic digital training environments becomes a critical enabler for entire industries.”
Our investment rationale
At XISTA Science Ventures, we invest in companies that build core infrastructure for emerging technology waves. As Dr. Annu Gmeiner, Principal at XISTA Science Ventures, explains:
“For autonomous driving, the world itself is the ultimate test environment; yet accessing it has traditionally required years of mapping and preparation. AVES Reality removes this barrier by turning the planet into a living digital laboratory that can be used instantly for development and validation, dramatically accelerating global-scale autonomous deployment. We believe deeply in the founders’ clarity of vision and conviction to build core infrastructure that reshapes how autonomy is developed and brought to market.”
Our decision to invest in AVES Reality was driven by the convergence of:
- the urgent need for scalable, deterministic validation of autonomous cars,
- rapid advances in geospatial AI, simulation, and world modeling,
- and a technically strong founding team focused on infrastructure rather than point solutions.
AVES Reality operates at the intersection of space data, autonomy, simulation, and physical AI, a combination we believe will define the next generation of autonomous systems.
Financing
AVES Reality raised an oversubscribed €2.7 million Seed round, led by Matterwave Ventures, with participation from XISTA Science Ventures, xdeck ventures, and Lightfield Equity, alongside continued support from Bayern Kapital and the Fraunhofer Technologie-Transfer Fonds (FTTF). We are proud to support AVES Reality as they build planetary-scale infrastructure for the future of autonomous cars.