Düsseldorf, Germany-based semiconductor startup Ubitium has raised $14 million in Seed funding to develop a universal processor for Physical AI applications, including robotics, autonomous vehicles and drones.
The round was led by Ibex Investors, with participation from OpenOcean, existing investors KBC Focus Fund, Runa Capital and Inflection, and Florian Seibel, co-founder and co-CEO of Quantum Systems, who invested personally. The financing brings Ubitium’s total Seed funding to $18 million since its founding in 2024.
Combining AI, Sensing and Control on One Processor
Founded in 2024 by CEO Hyun Shin Cho and CTO Martin Vorbach, Ubitium develops reconfigurable processors designed for machines that need to perceive their surroundings, process information and perform physical actions.
Autonomous machines typically use separate processors for different functions: digital signal processors for sensing, neural processing units for AI inference and CPUs for control.
Ubitium is developing an architecture that combines these functions within a single array of programmable processing elements. Computing resources can be allocated dynamically as workloads change, reducing the need to transfer data between specialised chips.
The architecture uses unified memory and standard RISC-V development tools, allowing manufacturers to work with a common processor and software environment.
“Physical AI machines bring a whole new set of computing challenges, and the inference isn’t the hardest one,” said Hyun Shin Cho, CEO and co-founder of Ubitium. “Sensing the real world and controlling the machine in it are just as hard, and today those still run on separate chips. We built one Physical AI processor that does all three, proved it in silicon in a year, and this funding turns it into a product.”
Preparing the First Commercial Processor
Ubitium has already demonstrated its architecture through the UB108, a chip containing 256 processing elements manufactured using Samsung’s 8nm process.
According to the company, the chip successfully demonstrated signal processing, AI inference and real-time control running on the same reconfigurable array.
Ubitium is now preparing its first commercial processor, the UB410, which will contain 4,096 processing elements. The company plans to complete the chip’s tape-out and deliver evaluation units in 2027.
Initial customer applications focus on high-resolution radar and audio systems, where signal processing and AI inference need to operate together. Ubitium is also targeting autonomous systems, including drones, as future applications for its complete sensing, AI and control architecture.
The company holds 17 granted patents covering its core technology, with additional applications pending. The new financing will support its transition from demonstrated silicon to a commercial processor.