Ubitium 融資 $14M，開發面向實體 AI 的通用處理器

Semiconductor startup Ubitium has raised $14 million in new seed funding to develop its universal processor for autonomous machines and other physical AI applications. 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 investment brings Ubitium's total seed funding to $18 million since its founding in 2024 and follows the development of its first chip, which demonstrated signal processing, AI inference and real-time control using a single processor architecture.

Ubitium is developing a processor for physical AI applications, including robots, drones, vehicles and industrial equipment. These systems need to process sensor data, make decisions and control physical movements in real time, functions typically handled by separate specialised processors. This requires data transfers between chips, increasing system complexity.

To address this, Ubitium is developing a reconfigurable processor that combines these functions within a single architecture. Its programmable processing elements can dynamically allocate computing resources between sensing, AI inference and real-time control, depending on the machine's workload.

Commenting on the challenges of developing processors for physical AI, Hyun Shin Cho, CEO and co-founder of Ubitium, said:

Physical AI machines bring a whole new set of computing challenges, and the inference isn't the hardest one. 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.

The architecture combines unified memory with standard RISC-V development tools, allowing manufacturers to use a common processor and software toolchain instead of multiple specialised chips. By reducing data transfers and simplifying software development, Ubitium aims to lower the complexity and cost of building intelligent machines.

Ubitium'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.

The company is now preparing its first commercial processor, the UB410, which will feature 4,096 processing elements. Initial applications will focus on high-resolution radar and audio systems, while autonomous systems, including drones, are among the target applications for the complete architecture.

Ubitium plans to use the new funding to advance the UB410 towards commercialisation, with the chip's tape-out and delivery of evaluation units planned for 2027.