In the heart of Europe’s tech landscape, a new player called simmetry.ai just took a bold step toward solving one of the toughest problems in artificial intelligence development: the lack of high-quality training data. On February 13, 2026, the Osnabrück, Germany-based company announced it secured a €330,000 funding injection from NBank, the investment and development bank of Lower Saxony through its High-Tech Incubator (HTI) programme. This support will help the startup expand a synthetic data platform that promises to reshape how engineers build and train AI systems.
Artificial intelligence thrives on data. For tasks like detecting objects in images or understanding complicated environments, models need thousands — often millions — of annotated examples. In sectors like agriculture, food processing and industrial automation, capturing all those real-world scenarios proves costly, time-consuming or even impossible. simmetry.ai aims to ease that burden by generating photorealistic synthetic data that mirrors real conditions, dramatically reducing the need to gather millions of real photos and videos.
Founders of simmetry.ai include CEO Kai von Szadkowski, CTO Anton Elmiger, and Prof. Dr. Stefan Stiene. They spun the company out of the German Research Centre for Artificial Intelligence (DFKI) in 2024, driven by the belief that synthetic data can unlock AI potential in fields where standard datasets fail to capture the variability of real-world conditions.
Tackling a Core AI Bottleneck
simmetry.ai grew from research into a solution that directly addresses what many AI developers describe as the biggest bottleneck in model creation: data scarcity. Industry insiders estimate that teams spend more than 80 % of their time and budget just collecting and preparing data before they can even begin training algorithms — a task that drains resources and delays deployment.
In agriculture, for instance, developers must train AI to recognize a wide range of crop types, weed species, and environmental conditions. Capturing every possible lighting angle, soil type and growth stage in real life would require years of data collection. Instead, synthetic data lets teams simulate these scenarios in silico, generating fully annotated images that reflect edge cases and rare events. simmetry.ai’s platform handles tasks like semantic segmentation, object detection, 3D pose estimation and regression, which are essential for modern computer vision models.
The startup’s technology generates diverse datasets across multiple sensor modalities, meaning it can mimic data from visual cameras, depth sensors and other imaging tools used in robotics, autonomous equipment, quality inspection systems and other complex industrial applications. This versatility makes the platform useful far beyond agriculture — extending to manufacturing floors and logistics environments where accurate perception drives automation.
How Synthetic Data Accelerates Innovation
At its core, simmetry.ai’s platform creates photorealistic, fully annotated training data tailored to each customer’s needs. Annotation matters because most machine learning workflows require labeled data — for example, identifying where an object appears in an image or determining the exact location of a weed in a field. With conventional data, engineers must painstakingly label thousands of images manually, a process both expensive and prone to human error. Synthetic data eliminates most of that manual work, offering precision and scale out of the box.
The startup’s approach brings several clear advantages:
- Cost Efficiency: Firms can skip costly field data collection, especially in agricultural or industrial settings where capturing every meaningful scenario can be prohibitive.
- Speed: Teams speed up model development by generating large custom datasets in hours or days instead of months.
- Robustness: Synthetic data lets models train on edge cases — rare but critical events that might not appear in real world samples but significantly impact field performance.
Imagine training an AI system to spot a broken part on an assembly line. In a real factory, you might struggle to capture enough images of defective items to teach a model reliably. With synthetic data, developers can fabricate thousands of variations, positioning defects in different orientations, lighting conditions, and backgrounds to ensure the model learns robustly.
Strategic Focus and Future Growth
The team at simmetry.ai chose agriculture as their first focus area because it combines heavy data requirements with high potential societal impact. Smarter AI could help farmers optimize crop yields, reduce water and pesticide use, and improve sustainability. But conventional training datasets fail to capture the complexity of outdoor environments, which vary wildly across seasons, weather, and crop types. simmetry.ai’s technology makes it possible to simulate precisely those uncontrolled variables.
Looking ahead, the company plans to use the NBank funding to build a scalable, user-friendly platform that serves AI developers across industries. The goal goes beyond one-off projects — simmetry.ai wants to become the backbone of synthetic data generation for teams worldwide, helping them iterate faster and innovate more confidently.
The roadmap includes:
- Expanding support for more sensor types and environmental conditions
- Improving automation and customization tools so non-experts can tailor datasets easily
- Scaling infrastructure to support larger datasets and more simultaneous users
- Building partnerships with larger AI toolchains and development platforms
By focusing on flexibility and accessibility, simmetry.ai embraces a broader vision of democratizing data generation — so even smaller AI teams can build world-class models without huge data budgets.
Broader Implications for AI Development
If synthetic data becomes widely adopted, it could transform the economics of AI development. Right now, companies with deep pockets and vast data inventories hold a significant edge. But tools like simmetry.ai lower that barrier, letting even resource-constrained teams train models that confidently handle rare conditions and complex environments.
This shift matters most in sectors like agriculture, manufacturing and robotics, where real-world trials prove costly or risky. Instead of sending engineers into fields or factories to collect datasets manually, teams can simulate those environments and produce precisely annotated images ready for training. This not only makes projects more efficient but also reduces safety risks and environmental disruptions linked to extensive field testing.
Final Thoughts
The €330,000 funding from NBank marks a meaningful milestone for simmetry.ai. It validates the startup’s vision and gives the team runway to expand and refine its technology. More importantly, it highlights a growing recognition that synthetic data stands as a cornerstone for future AI breakthroughs — particularly in areas where traditional datasets fall short. As AI spreads beyond tech labs into real-world applications in agriculture, manufacturing and beyond, tools that fill data gaps will become increasingly vital. Simmetry.ai has positioned itself at the forefront of this movement, aiming to help developers innovate faster, build better models, and solve problems that once seemed out of reach.
Also Read – SaaS Startups vs IT Services: The Real Shift