Perciv AI, founded by Balazs Szekeres, Srimannarayana Baratam, Dr. Andras Palffy, is a Hungarian-founded startup delivering artificial intelligence (AI)-based software solutions to unlock the latent capabilities of radars. The company’s solution enables durable and cost-effective autonomous systems with Lidar-like performance even in harsh weather. In December 2024, the company secured €2.5 million in a seed round to transform radar perception for autonomous systems.
In the Startup of the Day column, Perceive AI’s team outlines the startup's concept, product, and future goals.
The Startup of the Day column on AIN.Capital is dedicated to tech projects from all sectors that originate from the CEE countries. If you would like to introduce your project, please fill in the questionnaire.
Tell us about your startup. How does it work?
Expensive sensors, such as LiDARs drive up costs for automated vehicles and robotics. Perciv AI cuts these expenses by replacing LiDARs with radars using our next-generation AI-driven radar and radar-camera fusion technology.
Perciv AI develops AI-driven machine radar perception software solutions for automated vehicles and systems, such as cars, trucks, tractors, robots, or smart cities, to understand their environments even in harsh circumstances, increasing safety for both the passengers and surroundings.
We believe that radars are not exploited to their limits, and just like cameras and LiDARs, they can be pushed beyond their traditional limitations with dedicated AI, and perform extremely well for a cheaper price - even in adverse weather conditions.
We provide the world’s first dedicated AI for radar perception, enabling autonomous vehicles from cars to robots to understand their environment in any condition. LiDAR-like performance at a fraction of the cost. With any radar, in any circumstances.
How did you come up with the startup’s idea? What was the reason/motivation behind it?
We have spent large part of our academic life (one PhD and two MSc theses) on this topic, i.e. how to make better *use* of radars with dedicated AI based perception, instead of making better radar hardware or improving its classical signal processing steps. Our results were very well received not just in the academy, but also highly relevant industrial players showed up and wanted to hire us. Given our exposure to startup life before, we started to play with the idea of what if we try to make such product in the wild (i.e. not in the lab), but not within a huge corporation, but on our own.
How long did it take to reach the prototype or MVP? What did you encounter?
The earliest version of the software was presented to selected customers within 6 months. However, in our field, and in AI in general, high quality data is the new oil. It took us some time (about 10 months) to build up our capabilities to calibrate, synchronize, record, and annotate multi-sensor datasets to train our dedicated AI models. This pipeline is so unique that we have encounters to license parts of it instead/along with our main product in different fields.
When exactly did you launch your product? Or when the launch is planned?
Our perception SDK is a software package with several modules, which is updated and extended with new features continuously. Early versions of the software was made available for selected partners in early 2024, and we plan to extend the user base in 25 significantly.
Tell us about the startup's business model. How do you monetize your product?
We operate on a license based model either on a per-sensor or a per-vehicle basis. In special use-cases, we have subscription models as well. The exact amount depends on the expected volumes, ordered features, and the exact operational domain.
What are your target markets and consumers?
Instead of focusing on a single market, e.g. ADAS (Advanced Driver Assistance Systems) in automotive, we have realized that our technology is widely applicable across industries. You can put a cheap radar sensor using our software on a delivery robot, forklifts, airport vehicles, tractors, warehouses, smart city applications, and so on - for Perciv, the technology remains the same. We even have partners in the aerospace industry.
If the startup has already launched the product, what are the results: metrics, income, or any clear indicators that can be evaluated.
Our traction speaks for itself - we had more than 20 companies as partners even before raising our seed round. Exact KPIs are different across use-cases, but to give you an example, in automotive applications, our technology beats classical radar based object detection approaches by 2-fold in certain metrics.
What about your team? How many people are working in the startup? If you’re looking for new employees, indicate whom exactly.
We are a team of 13 currently, 7 full time employees with 4 interns (who do their MSc thesis with us), along with a part time colleague, and 2 advisors. We are planning to double our size next year (2025), by hiring several Software Developers, Perception Engineers, Acceleration Engineers, but also some Business Developer help.
Have you already raised any investments? Provide us with more details on each funding round: the amount, investors, the purpose of the investment.
Perciv AI has raised 700k+ Eur in subsidies and grants, basically skipping a pre-seed round. We have just closed our seed round with a size of 2.5M eur. See the press release here.
What's next? Tell us about your future plans.
Perciv AI’s mission is to democratize autonomous technology by making perception safe, reliable, and affordable. In the next 2 years, using the recently received funding, we plan to expand its operations further and scale our radar-based perception solutions across various industries such as automotive, robotics, and aerospace. We will go into production with a agricultural drone manufacturer in 25 Q1, and plan to enter production with multiple AGR (Automated Ground Robots, e.g. forklifts, airport vehicles, delivery robots) companies in 2026.
