About Paul Veth
A student of Physical AI, learning in public by studying the actuator technology that powers humanoid robots.
I am genuinely new to the world of humanoid robotics. I am not an engineer, not an academic, not an industry insider. I started ActuatorHQ because I wanted to understand how humanoid robots actually work at the hardware level, and I could not find coverage that was both accessible and technically serious.
So I decided to study it myself, in public. I read research papers, dig into patent filings, compare spec sheets, and try to connect what I find to the bigger picture. When I get something wrong, I correct it. When something is beyond my current understanding, I say so.
The result is a project that documents a learning journey. Some readers find it useful because a beginner asking basic questions sometimes surfaces insights that experts take for granted.
In 2025, I kept reading about humanoid robots in the news, but every article seemed to be either pure hype or so deeply technical that it assumed years of domain knowledge. I wanted something in between: serious coverage that a curious outsider could follow. When I could not find it, I started writing it myself. I picked the component layer as my focus because that is where the real engineering constraints live. An actuator determines what a robot can physically do. A gearbox determines how efficiently it does it. A thermal system determines how long it can keep going. These are concrete, measurable things, and studying them gave me a way into a complex field. ActuatorHQ started as a personal study project. I would read a paper, try to understand it, write about what I learned, and verify my understanding against the original sources. Over time, the articles got more detailed and the questions got sharper, but the approach stayed the same: study, write, check, share. I am still early in this journey. There is more I do not know than what I do. But that is exactly what makes learning in public valuable: the questions you ask when everything is new are sometimes the most useful ones.
Got curious about humanoid robots. Started reading papers and datasheets to understand how they actually work.
Could not find coverage that was both accessible and technically grounded. Decided to write it myself.
Launched ActuatorHQ as a study-in-public project, focused on the component layer: actuators, motors, drives, sensors.
Building a growing library of sourced analysis while continuing to learn. Still very much a student.
Most market analysis of Physical AI starts from the top , valuations, press releases, partnership announcements. ActuatorHQ starts from the bottom: the component specs. From torque density to backdrivability to thermal management thresholds, the hardware layer is where competitive advantages are built or lost. Every analysis I publish connects those component-level realities to the bigger picture: which companies can scale, which supply chains are vulnerable, and which technical bets are likely to pay off. Every claim is backed by data, datasheets, or primary sources. Nothing ships based on narrative alone.
Every analysis starts at the hardware layer. I read datasheets, catalog specs, and build comparison frameworks for actuator types , electric, hydraulic, quasi-direct drive, series elastic, harmonic drive , before drawing any market-level conclusions.
Specs don't exist in isolation. I track funding rounds, production scaling milestones, partnership structures, and go-to-market strategies for key players including Figure AI, Tesla Optimus, Unitree, Agility Robotics, Apptronik, Fourier Intelligence, and UBTECH.
Understanding who makes the motors, encoders, harmonic drives, force/torque sensors, and controllers that go into humanoid systems , and what that means for cost curves, bottlenecks, and competitive moats , is central to every ActuatorHQ deep dive.
Built and scaled multiple companies across technology markets , giving a ground-level understanding of how industries mature, where capital flows, and how component decisions become competitive advantages.
Hands-on experience building content engines, voice agents, and identity systems. Understanding AI not from the outside as an observer, but from the inside as someone who has made architecture and product decisions under real constraints.
Dedicated, public study of the Physical AI and humanoid robotics market , covering actuator specs, key player strategies, supply chain dynamics, and market structure across electric, hydraulic, and quasi-direct drive systems.
Every claim published through ActuatorHQ is grounded in data, datasheets, or verifiable primary sources. No narrative-first analysis. No speculation dressed as insight.
Two decades of watching technology markets move through hype cycles, shakeouts, and consolidation phases , applied directly to reading where the Physical AI market is and where it's heading.
The Physical AI space is saturated with breathless coverage that mistakes ambition for achievement. ActuatorHQ publishes nothing that isn't grounded in specs, sourced data, or verifiable market information. If the numbers don't support a claim, the claim doesn't ship.
Analysis at ActuatorHQ is written for people who build, invest in, or make decisions about hardware systems , not for people who just want to sound informed at conferences. The test for every piece: would an engineer or a founder find this genuinely useful?
Vague market commentary is everywhere. Specific, comparative actuator performance data mapped to real company architectures is rare. ActuatorHQ chooses specificity every time , concrete numbers, named companies, real trade-offs.
Studying the Physical AI market openly , sharing frameworks as they develop, acknowledging uncertainty where it exists, and updating analysis as new data emerges. Intellectual honesty over the appearance of authority.
Every piece of analysis passes through a single filter: would this be useful to someone actually building in this space? Not just interesting , useful. That means actionable framing, honest uncertainty, and no hype padding.
The actuator, sensor, and controller decisions made today will determine which humanoid platforms are viable at scale in five years. ActuatorHQ exists to make that layer legible , to engineers who need specs and to investors who need to understand what those specs mean competitively.
A student of Physical AI, learning in public by studying the actuator technology that powers humanoid robots.
If you're an engineer, investor, founder, or analyst working in or adjacent to Physical AI , and you want analysis that actually connects component specs to market reality , I'd like to hear from you. Whether it's a specific actuator question, a market dynamics discussion, or a collaboration on research, reach out directly. No gatekeeping, no sales process.