Ubitium, a semiconductor startup developing a universal processor for physical AI applications, has secured US$14 million in new Seed funding. The round was led by Ibex Investors, with participation from OpenOcean, existing investors KBC Focus Fund, Runa Capital, and Inflection, and Quantum Systems co-founder Florian Seibel. The investment brings Ubitium's total Seed funding to US$18 million since the company was founded in 2024, following first silicon that demonstrated signal processing, AI inference, and real-time control on a single architecture.
Addressing the Computing Demands of Physical AI
Autonomous machines such as drones, robots, vehicles, and industrial equipment must perceive, think, and act continuously, often on battery power and in real time. Today, developers typically split these workloads across a digital signal processor for sensing, a neural processing unit for AI, and a central processing unit for control, which adds data transfers and engineering complexity. Ubitium is replacing that stack with one reconfigurable array of programmable elements that allocates computing resources at runtime across sensing, AI inference, and real-time control.
Proven in Silicon
The company's first chip, the UB108, contains 256 processing elements and was manufactured using Samsung's 8nm process. It has 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, with 4,096 processing elements, and plans to tape out and deliver evaluation units in 2027.
Commercial Targets and Developer Benefits
Initial applications will focus on high-resolution radar and audio systems, which bring demanding signal processing and AI onto one processor in markets where embedded chips already sell in volume. Autonomous systems, including drones, are the target applications for the complete perceive, think, and act loop. The architecture combines unified memory with standard RISC-V development tools, allowing manufacturers to use a common processor and toolchain instead of multiple specialized chips.
Investor Confidence in Reconfigurable Computing
Quantum Systems co-founder Florian Seibel said autonomous machines must process surroundings, make decisions, and control themselves on board within strict power and weight limits. Ibex Investors partner Jeffrey Peters said the reconfigurable architecture can deploy the latest AI advances on a single piece of silicon with one set of development tools. OpenOcean general partner Tom Henriksson called Physical AI a ground-up hardware revolution and described Ubitium as the foundational processor for this new era.
Founding Team and Intellectual Property
Ubitium was founded in 2024 by CEO Hyun Shin Cho and CTO Martin Vorbach, a pioneer in reconfigurable computing with more than 200 patents licensed by major semiconductor companies. The startup holds 17 granted patents covering its core technology and reports team experience from Intel, Nvidia, Apple, and Texas Instruments, alongside more than 250 career patents and over 350 peer-reviewed publications. Headquartered in Düsseldorf, Germany, Ubitium is backed by Ibex Investors, OpenOcean, KBC Focus Fund, Inflection, and Runa Capital.
The new capital will support commercialization of the UB410 as Ubitium works to make intelligence and perception a standard component of every machine, similar to the way the microprocessor standardized computing. With first silicon already proven, the company is positioned to offer a flexible alternative to the patchwork of specialized chips used in autonomous systems today. Initial customers in radar and audio provide a near-term route to market, while autonomous systems remain the broader target for the complete perceive, think, and act loop.