Mundane Systems

Mundane Systems Inc. (styled mundane) is an American robotics startup based in Palo Alto, California, that builds teleoperated robots for hazardous skilled-trades work, starting with energized electrical work. A tradesperson wears a force-feedback exoskeleton and drives the robot with their own arms, seeing through the robot's cameras and feeling what it touches, so the robot rather than the worker is exposed to the hazard ^[1][2]. The company describes its product as "a second body for the skilled trades" and says its mission is "to end injuries on the job and open a path to unlimited labor" ^[1][3]. Its chief executive is co-founder Theo Nash ^[12][13].

Mundane positions itself against robotics efforts that aim to replace workers. Its position is that a skilled operator should stay in control of the robot, and that the recorded work can later train AI that handles routine parts of a task under the operator's direction ^[2][6].

Overview

History

Theo Nash, a co-founder, grew up around auto garages in London and studied at Stanford, according to the show notes of the Machine Minds podcast, which interviewed him in March 2026 (episode 128, hosted by Greg Toroosian). The same notes say he had built teams in Palo Alto, Shenzhen and Vancouver ^[12]. Mundane's website identifies its co-founders as Theo Nash, Maximilian Drach and Cameron Hilman, and its news page names Eoin O'Doherty as chief operating officer ^[1][4][8]. Purdue University's College of Engineering listed Drach as "Vice President, Embodied Systems and Co-Founder" of mundane ^[14].

The company's public appearances during 2026 included:

At the Purdue summit, whose 2026 program addressed artificial intelligence as it relates to physical AI, nuclear energy and space, Drach spoke about moving robots out of demonstrations and into production environments, according to Mundane's account of the event ^[7][14]. Mundane says that at the Plug and Play event it pitched in the Deep Tech session, where the program introduced it as a company "building humanoids that are useful beyond just lab demos, but in actual real-world deployments" ^[8].

On 28 October 2026 the company's X account posted a film of about 82 seconds with the text "mundane gives tradespeople a second body that can take the shock or blast instead of them" and "Our mission is to amplify the skilled labor force, not replace it" ^[10].

Technology

Mundane's technology page is adapted from a company document titled "Dissolving the Embodiment Gap Between Human Operator and Robot" ^[2]. The company uses the term "embodiment gap" for the distance between working with one's own hands and working through a machine, and says that closing it is "the whole project" ^[2]. In an October 2026 interview, Nash defined it as "the difference in geometry, feeling, and proprioception between the human and the robot" ^[11]. (The same phrase is used more broadly in robot learning for the mismatch between a human demonstrator's body and a robot's; see Teleoperation.)

The company's critique of existing interfaces

Mundane argues that conventional teleoperation fails operators in three ways: misaligned or lagging stereo video causes eye strain and motion sickness; controllers that touch the operator only at the hand cannot convey contacts along the robot's arm, and vibration "can hint that you touched something, but it cannot stop or guide your hand"; and when only the hand is tracked, software must guess the robot's elbow position ^[2]. The company presents the following comparison of approaches (its own framing) ^[2]:

Sight, touch and proprioception

The company organizes its design around three senses ^[2]:

• Sight. The operator sees through the robot's cameras in stereo, with depth that can be judged at arm's length. The robot's head follows the operator's head, and the headset keeps the view responsive while the robot catches up ^[2].
• Touch. The exoskeleton contacts the operator along the arm rather than only at the hand, so a bump on the robot's forearm is felt on the operator's forearm, and a solid object physically stops the operator's arm. According to Mundane, a physical model of each arm predicts the forces produced by the machine itself, and only the remainder, attributed to the outside world, is passed to the operator. The company says keeping this feedback stable when the operator is far away is "the hardest part, and the part we are working on now" ^[2].
• Proprioception. The operator's arm, the exoskeleton and the robot arm share the same joints in the same order on the same axes, so each human joint drives the matching robot joint directly with no inverse-kinematics step. Mundane says this lets an operator move the robot's elbow independently of its hand, which it describes as a safety matter in tight energized spaces, and that most operators stop thinking about the robot within minutes. The design is marked "Patent pending" ^[2].

Nash told the with robots newsletter that the exoskeletal controller "links your fingers, hands, elbows, and shoulders directly to the robot," and said of the system, "We've been testing it for eight or nine hours on people, and it's been working pretty well" ^[11]. He also said Mundane had collected data in-house with an ARX arm in a leader-follower setup similar to ALOHA, and argued that such rigs, and head-camera "egocentric" capture, slow workers down compared with doing a task by hand ^[11].

Modular robot

Mundane treats the torso (two arms, two hands and a head with cameras) as the part of a humanoid robot that does skilled work, and builds that part to match the operator joint for joint. The torso can be mounted on a wheeled base, a mast, legs or the end of a boom ^[2]. Nash said in October 2026 that the company's robots run on wheels, using an off-the-shelf autonomous mobile robot (AMR) base bought from an existing manufacturer, and that the same chest, arm and head module can be fitted to a cherry picker for elevated work. He said the company chose wheels over legs because a robot falling over near customers' equipment is a risk customers will not accept ^[11].

Applications

Mundane lists six target industries on its website ^[1]:

Nash said the company was "particularly focused today on the electrical space," named nuclear power stations and decommissioning as environments the company could address "in time," and said he was applying the system to "the most dangerous, high-risk, arc-flash electrical environments" ^[11]. Under the company's commercial model, customers lease the robot and exoskeleton and their own qualified workers operate it under their own procedures; Mundane says the system does not depend on autonomy to be useful ^[2].

NECA 2026 demonstration

Mundane Systems Inc. is listed among exhibitors in the official show guide for NECA 2026, the National Electrical Contractors Association convention and trade show held 4-7 October 2026 at the Mandalay Bay Convention Center in Las Vegas, with show-floor hours on 5-7 October ^[15][16]. According to Mundane, more than 100 electricians, linemen, contractors and safety leaders operated its robot at the booth over the three show days, and the booth received a Best Booth award ^[5]. Lineman Jake Holsthaus of Written in Red Training joined the company at the booth; Mundane quotes him as saying, "If I can do that same job standing 50 feet away instead of right next to it, why wouldn't I?" ^[5]. The same show guide lists another teleoperated product aimed at electrical work, Hardhat Robotics' SparkE-B, described as a robot for energizing and de-energizing high-voltage panels ^[15].

Path to autonomy

In a white paper published on 17 June 2026, Mundane set out three phases, each of which it says is economically useful on its own ^[6]:

The white paper argues that the main shortage in physical AI is force-rich, recovery-aware data from real industrial work, and says Mundane's systems are designed to record, where available, joint positions and velocities, motor currents and estimated torques, end-effector pose, gripper state and force, force/torque sensor readings, stereo and multiview video, operator commands, contact events, interventions and recoveries ^[6]. It cites published research as support for the mechanism (while stating that none of it proves Mundane's own deployment thesis), including the Universal Manipulation Exoskeleton paper from Ant Group and Stanford, Mobile ALOHA, Stanford's BEHAVIOR Robot Suite and "Just Add Force" ^[6][17]. The white paper discloses that Mark Cutkosky of Stanford University, a co-author of the Universal Manipulation Exoskeleton paper, serves on Mundane's advisory board ^[6][17]. It also invokes the Jevons paradox to argue that cheaper skilled work would increase demand for it rather than reduce employment ^[6].

The technology page says the company's long-term mission is "indistinguishable telepresence" and that, where customers allow it, work done through the system becomes data that makes the robot more capable, with the operator's role shifting from performing every motion to directing and checking the work ^[2]. In the October 2026 interview Nash said that "when AI gets good, they'll be supervising multiple robots in the field" ^[11]. Mundane's careers page in October 2026 listed in-person roles in Palo Alto including a robot learning engineer, a haptics and teleoperation controls engineer, and a mechanical engineer to take its arms "from prototype to the first 100 units in the field" ^[9].

Continue reading

• EExoskeleton (Wearable Robotics)An exoskeleton, sometimes called a powered exoskeleton, exosuit, or wearable robot, is a powered or passive mechanical structure worn on the body to augment, assist, or restore human movement.Mentioned 9 times above
• TTeleoperationTeleoperation is the operation of a machine by a human at a distance, with commands carried over a communication link and the machine's cameras, microphones and force sensors standing in for the operator's own…Mentioned 7 times above
• PPhysical AIPhysical AI is a contested umbrella term for artificial intelligence systems whose decisions are coupled to sensing and action in the physical world.Mentioned twice above
• Humanoid Robot"Humanoid" redirects here. For the London-based robotics company, see Humanoid (company).Listed in See also
• IImitation LearningImitation learning is a family of methods for learning sequential behavior from demonstrations.Listed in See also
• RRobot foundation modelA robot foundation model is a large-scale machine learning model, typically based on the transformer architecture, that is pre-trained on broad, diverse datasets of robot interactions and then adapted to a…Listed in See also

More from See also

• Embodied AI

More to read

• Robot learning
• Mobile ALOHA
• Mimic Robotics
• Machenta
• Simulation (in AI and robotics)

Topics:Embodied AIPhysical AIRoboticsRobotics Companies

References

1. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9Mundane Systems Inc. "Mundane" (company homepage). mundane.co (accessed 30 October 2026).
2. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12 ^13 ^14 ^15Mundane Systems Inc. "Technology: The interface for the future of hands-on work" (adapted from "Dissolving the Embodiment Gap Between Human Operator and Robot"). mundane.co/technology (accessed 30 October 2026).
3. ^Mundane Systems Inc. "Our story." mundane.co/our-story (accessed 30 October 2026).
4. ^1 ^2Mundane Systems Inc. "News and press." mundane.co/news (accessed 30 October 2026).
5. ^1 ^2 ^3Mundane Systems Inc. "100+ tradespeople operate the robot at NECA 2026." 7 October 2026. mundane.co/...neca-2026
6. ^1 ^2 ^3 ^4 ^5 ^6 ^7Mundane Systems Inc. "Most of robotics is racing to delete the worker. What if we amplified the worker instead?" White paper, 17 June 2026. mundane.co/...amplify-the-worker
7. ^1 ^2Mundane Systems Inc. "Mundane cofounder Maximilian Drach speaks at the Purdue Engineering Tech Summit." 18 June 2026. mundane.co/...purdue-tech-summit
8. ^1 ^2 ^3Mundane Systems Inc. "Mundane COO Eoin O'Doherty pitches Mundane at Plug and Play's AI for Physical Industries." 11 August 2026. mundane.co/...plug-and-play
9. ^Mundane Systems Inc. "Careers." mundane.co/careers (accessed 30 October 2026).
10. ^1 ^2mundane (@mundanebot). Post on X, 28 October 2026. x.com/...2115551780805472263
11. ^1 ^2 ^3 ^4 ^5 ^6 ^7"The Robot Takes the Blast" (interview with Theo Nash). with robots (Substack newsletter), 29 October 2026. withrobots.substack.com/...the-robot-takes-the-blast
12. ^1 ^2 ^3Toroosian, Greg. "Automating the Mundane: The Team Rewriting Everyday Work with Theo Nash." Machine Minds, episode 128, 11 March 2026. machineminds.buzzsprout.com/...work-with-theo-nash
13. ^1 ^2 ^3Humanoids Summit. "Agenda: Tokyo 2026" (Day 1, 28 May 2026). humanoidssummit.com/agenda-2
14. ^1 ^2 ^3Purdue University College of Engineering. "Tech Summit" (18 June 2026, Computer History Museum). engineering.purdue.edu/...tech-summit
15. ^1 ^2 ^3National Electrical Contractors Association. "The Official NECA 2026 Las Vegas Show Guide" (exhibitor list as of 2 September 2026). ecmag.com/...al-neca-2026-las-vegas-show-guide.pdf
16. ^NECA Convention and Trade Show 2026. Official show site. necashow.org/...Enter
17. ^1 ^2Liang, L.; Xu, J.; Qi, X.; Cai, Y.; Ding, H.; Wang, L.; Sun, Z.; Chow, J.-H.; Yang, M.; Cutkosky, M. "Universal Manipulation Exoskeleton: Learning Compliant Whole-body Policies with Real-time Torque Feedback." arXiv:2606.14218, June 2026. arxiv.org/...2606.14218

Show all 17 references