Bloomberg TV Bulgaria invited Synthgen to discuss our agricultural AI work, the €124,000 Netherlands grant, and how synthetic data accelerates farming robotics.
We at Synthgen are excited to share that we recently had the opportunity to speak on Bloomberg TV Bulgaria's show "UpDate." During the interview, we discussed our journey, our vision, and how we're using synthetic data and AI to transform agriculture. (bloombergtv.bg)
The idea behind Synthgen started with a simple goal: to generate wildlife images for AI training. From there, we realized that synthetic data has much broader potential - especially in domains where collecting real data is hard or costly. That's how we pivoted toward agriculture.
Recently, we were honored to receive a €124,000 grant from the Netherlands Enterprise Agency. This funding supports our work with Luxeat Robotics - a Dutch company building farming robots that use lasers to eliminate weeds. It's a collaboration that brings together their real-world robotics and field data with our synthetic-data generation and AI training capabilities.
With synthetic data and automated AI pipelines, we expect that farm robots can be trained and deployed up to 10× faster than before - because AI doesn't need breaks and synthetic data can simulate many more scenarios than real-world data alone.
Our goal is to make this technology accessible for everyone - from large agribusinesses to small farms and regional producers. We believe mechanized farming should not be limited by dataset size or computing resources.
The grant will mainly support our growing team, the necessary hardware, and cloud infrastructure - because AI workloads require significant compute power.
Synthetic data is "simulated data" - data that look and behave like real-world data but don't come from actual events. In practice, we take real samples (e.g. photos of plants, soil, fields), then use AI to generate tens, hundreds, or even thousands of variations. Another AI validates them. This approach massively multiplies the data available for training models - and reduces the need for costly, time-consuming manual collection or labeling.
Because agriculture involves varied crops, soils, lighting, weather, and environmental conditions - things that are hard to capture exhaustively - synthetic data allows us to cover edge cases, rare events, or hard-to-replicate scenarios. That versatility can make a big difference when training AI-powered robots for real-world farming.
Our long-term vision goes beyond agriculture. We plan to expand to other industries, from sectors like construction, to healthcare, semiconductor manifacturing and beyond.
We're grateful for the attention from Bloomberg TV Bulgaria - and even more excited about what's ahead. We look forward to sharing more updates as our work progresses!
