One Universal Robot Brain That Works on Any Machine
Neuromorphic is a universal robot brain that attaches to any robot body, turns plain-language instructions into autonomous workflows, and deployed its first customer from initial conversation to daily operations in one week.
Photo source:
Neuromorphic
Why Neuromorphic Exists
Deploying a robot today is slow, expensive, and
highly technical. Each robot typically requires weeks or months of custom
programming before it can perform even a single task in a new environment. If
the task changes, the programming starts over. If the robot body changes, the
entire software stack often needs to be rebuilt. This means that companies
interested in automation frequently find the setup cost and timeline more
burdensome than the labor the robot was meant to replace. Neuromorphic, a San
Francisco-based startup founded by three Turkish roboticists, built a universal
robot brain designed to eliminate that bottleneck entirely.
The company was founded by Ege Doğanay, Vatan
Aksoy Tezer, and Cem Şeref Toker, three engineers who first met as high school
teammates competing in the FIRST Robotics Competition, where they won multiple
championships together. They spent the following three years building robots
side by side before going on to research and engineering roles at Cambridge
University, Bilkent University, Stanford University, and Nanyang Technological
University. Neuromorphic was accepted into Y Combinator's Summer batch and received
$500,000 in initial funding. The company is also a member of the NVIDIA
Inception Program. Robotics researcher Özgür Soysal, currently pursuing his PhD
at Stanford, supports the team as an advisor.
How This Universal Robot Brain Works
Neuromorphic packages sensors, compute
hardware, and software into a single module that can be attached to any robot
body. The system does not require custom code for each machine. The same
universal robot brain runs on quadrupeds, wheeled robots, and humanoids. Once
attached, the brain handles perception, decision-making, and motor control
without needing to be reprogrammed for the specific hardware it sits on.
At the core of the system is an onboard large
language model orchestrator. It takes plain-language instructions, whether
typed into Slack, sent by email, or spoken over the phone, and translates them
into structured robotic workflows. Those workflows are assembled from a library
of reusable skills that cover core robotic capabilities: navigating spaces,
picking up objects, opening containers, inspecting areas, and using tools. Each
action is monitored in real time by sensor-based safety loops that verify execution
and intervene if something goes wrong. The result is a robot that can be told
what to do in everyday language and carry out the task autonomously, without a
robotics engineer standing behind it writing code.
Exploring the Practical Benefits of a Universal Robot Brain
The most striking demonstration of the system's
speed came with its first customer. A leading biotechnology company went from
its initial conversation with Neuromorphic to a signed commercial contract in
five days. Within one week, a robot was deployed in the company's wet lab,
operating autonomously around the clock. That robot has already completed over
1,500 tasks. The company did not need to hire a robotics team, write custom
software, or spend months integrating the system into its facility.
This speed is possible because the universal
robot brain separates the intelligence from the hardware. Traditional robot
deployments tie the software tightly to a specific machine. If the company
wants to add a second robot with a different form factor, the integration
starts from scratch. Neuromorphic's approach means the same brain can be moved
from one robot body to another, carrying its skills and learned behaviors with
it. A wet lab that starts with a wheeled robot performing sample transport
could later add a robotic arm for pipetting or a quadruped for facility
inspection, all running the same underlying intelligence without rebuilding the
software.
The Experience Behind the Founding Team
The founding team's combined track record in
robotics is what gives the early product credibility despite the company's
recent launch. Vatan Aksoy Tezer served as lead robotics engineer at Augmentus,
where he deployed motion planning algorithms to production with sub-millimeter
accuracy. Before that, he built navigation systems from scratch at KABAM
Robotics, deploying over 100 robots, and scaled automated storage and retrieval
systems at Rapyuta Robotics, bringing over 1,000 robots to production. He is
also a core maintainer of MoveIt, one of the most widely used open-source
motion planning frameworks in robotics.
Ege Doğanay conducted robotics and AI research
at Bilkent University and served as a research fellow at Nanyang Technological
University, publishing in leading AI and robotics conferences. He won HackMIT,
Stanford XR, and MIT Reality Hack. Cem Şeref Toker studied engineering at
Cambridge University, co-founded ErgoFlow to build LLM routing and training
acceleration, and previously developed precision calibration algorithms for
industrial robots at Zeeko, achieving 30% higher accuracy. The three founders
bring experience spanning academic research, open-source robotics
infrastructure, and commercial deployment at scale, a combination that few
early-stage robotics startups can match.
Where Universal Robot Brains Fit in the Robotics Industry
The robotics software market is entering a
phase where multiple companies are racing to build the intelligence layer that
sits between a robot's hardware and its tasks. Some are building vertical
solutions for specific industries. Others, like Neuromorphic, are building
horizontal platforms designed to work across different robot bodies and
different application domains. The bet behind a universal robot brain is that
the intelligence problem is separable from the hardware problem, and that
solving it once creates more value than solving it separately for each machine.
Neuromorphic's initial focus is wet labs and
physical industries where repetitive, monotonous, or hazardous tasks currently
require human workers to perform them manually. The company describes its model
as Robot-as-a-Service, making industrial robots more affordable and accessible
by removing the engineering overhead that has historically made deployment
prohibitively complex for most organizations. Whether a universal robot brain
can maintain reliability and safety across fundamentally different robot bodies
and work environments is the central technical question the company will need
to answer as it scales beyond its first deployment. What it has demonstrated so
far is that the concept works in at least one real-world setting: a
biotechnology wet lab where a robot is operating autonomously every day,
completing tasks it was never individually programmed to perform.
You have exceeded your free limits for viewing our premium content
Please subscribe to have unlimited access to our innovations.
Get The Latest In Innovation
Join Our Newsletter
Where bold ideas find bold readers.
