Who we are
VeriMake is a research and engineering team working at the frontier of embodied intelligence. We see embodied intelligence as a universe being born — the Embodiverse. Silicon and steel are its matter, agents are its life, and learning is its evolution.
What we mean by silicon life. A machine becomes silicon life when it can do three things: sense and think on its own silicon, act in the physical world through a body, and keep learning from experience — and pass what it learns on to the next generation. Chips, nervous systems, bodies and communities are how we build each of these, layer by layer.
Silicon life, AI and robots. AI gives silicon life its mind; a robot gives it a body. Neither alone is yet life — a mind with no body cannot act, and a body that only runs a fixed program cannot grow. Silicon life is where the two meet: a mind on its own silicon, in a body, that keeps learning in the real world.
Vision — Building a universe for silicon life.
Mission — Driving the evolution of silicon life — from chips to nervous systems, bodies and communities.
Values — Believe it and make it.
(Believe it: we trust the power of belief, which turns imagination into reality. Make it: we learn by building, and ideas become real when we put our hands on them.)
Right now — we build the chips, data and bodies that let robots learn in the real world, and we open them to everyone who makes.
What we’re building now
- Edge accelerators. Domain-specific hardware that runs AI workloads fast and at low energy — from low-precision arithmetic for large language models to the Vina-GPU and Vina-FPGA docking accelerators.
- Real-world data. Our in-house EGO egocentric capture device is in volume shipment, collecting the first-person experience robots learn from.
- Robot bodies. The Classroom Assistant Robot, a humanoid teaching assistant built to work safely beside teachers every day.
- An open community. The VeriMake Community, where makers share work on RISC-V, FPGA, embedded systems and AI.
Our mission, unpacked
Driving the evolution of silicon life — from chips to nervous systems, bodies and communities.
We believe silicon life grows in layers, each one built on the one beneath it — and we work on every layer.
Data is the experience of silicon life. A single robot learns from what it has lived through; a community learns from what all of them have lived through.
- Chips — where it starts. Intelligence needs a substrate. We design domain-specific accelerators on FPGAs and edge hardware that run demanding workloads fast and at a fraction of the energy, from low-precision arithmetic for neural networks onward. Our accelerators have already sped up molecular docking for drug discovery (Vina-GPU, Vina-FPGA) — proof that the same design skills carry over to the workloads of silicon life.
- Nervous systems — sensing and thinking. On top of that compute we build perception and decision-making that run at the edge: multimodal sensing, speech and vision, and human–computer interaction, so machines can understand the world and the people in it.
- Bodies — acting in the world. Intelligence becomes embodied when it can move and touch. We build robot platforms and intelligent hardware, such as our Classroom Assistant Robot, that work safely alongside people every day.
- Communities — shared experience, the layer we are starting on. No life evolves alone. Our first steps are shared data — EGO devices gathering real-world experience that many robots can learn from — and an open community of builders. Networks of robots that cooperate and learn from each other are where we are heading next.
Running through every layer is evolution, and data is its fuel: systems that keep learning in the real world, so that each generation goes further than the last.
Beyond the main line
Silicon life will not only work for people — it will create with them. We take this up through human–AI co-creation projects such as Lyra, an AI music-creation tool founded by Chenye Wang, with Yanxiang Zhu as mentor and collaborator.
Where the name comes from
The summer after second grade, our team leader Yanxiang Zhu watched an anime called Robotech. The vastness of space, the giant starships, and above all the Veritech fighters — jets that transform into humanoid mecha — captivated him and opened the door to science fiction.
In the Robotech universe, “Veritech” is explained as Variable Engineering and Robotic Integration Technology. We took its first part, Veri, and added Make:
VeriMake — bringing the imagination of anime into the real world.
Community
We run the VeriMake Community, where we share projects and experience in RISC-V, FPGA, embedded systems, AI and HCI.
Collaborate
Researchers, engineers and organizations interested in embodied AI, domain-specific accelerators and robotic systems are welcome to reach out — see Contact.