This universal processor combines CPU, GPU, DSP and FPGA in one chip
- One processor architecture for AI, embedded systems and robotics
- The scalability of RISC-V ensures seamless expansion between various applications
- Seed funding fuels Ubitium’s drive to redefine chip technology
For more than fifty years, the semiconductor industry has relied on the Tomasulo algorithm, introduced by IBM in 1967, to build specialized CPUs, GPUs, and other chips tailored to specific computing tasks.
Now Ubitium, a hardware startup founded by semiconductor veterans, has developed a universal RISC-V processor that consolidates all computing workloads on a single, affordable chip.
This technology is especially important for embedded systems and robotics, where hardware costs often limit the use of advanced computing solutions.
Removing the boundaries between specialized computing tasks
Designed to be scalable, Ubitium’s universal processor supports a portfolio of chips that vary in size but share the same microarchitecture and software stack, allowing customers to expand their applications without changing their development processes.
The processor’s workload-independent design makes it suitable for any computing task and helps simplify hardware requirements.
Ubitium has raised $3.7 million in seed funding, which will accelerate the development of prototype chips and initial development kits, with plans to launch the first commercial processors by 2026.
“The $500 billion processor industry is built on restrictive boundaries between computing tasks,” said Hyun Shin Cho, CEO of Ubitium.
“We’re erasing those boundaries. Our universal processor does it all – CPU, GPU, DSP, FPGA – in one chip, one architecture. This isn’t an incremental improvement. It’s a paradigm shift. This is the processor architecture that powers the AI era required.”
Cho further stated that the company envisions a future where a single processor design can perform tasks ranging from small embedded systems to high-performance computing without specialized hardware modifications.
“For too long we have accepted that making devices intelligent means making them complex. Multiple processors or processor cores, multiple development teams, endless integration challenges – today that changes,” he added.