Minerva's Roger: $10M Behind Remote-Controlled Humanoids for Dangerous Jobs

Minerva Humanoids came out of stealth on 6 October 2026 with about $10M in pre-seed funding to build Roger, a humanoid that a trained specialist drives through a VR headset. The target jobs are bomb disposal, hazardous material response and oil and gas work. Roger is a prototype: Minerva has published no height, weight, payload, battery or price figures.

What is Roger?

Roger is a semi-autonomous humanoid robot. According to The Next Web, a specialist wears a VR headset, sees through the robot's eyes and does fine manipulation from a safe distance. Minerva's own AI control system handles balance, fall recovery and navigation, so the operator does not have to steer every step.

In plain terms: a person's hands and judgment go to the hazard, and their body stays behind. Interesting Engineering reports that the company's control software is called Minerva Intelligence and that low-latency control loops translate the operator's movements into robot actions, according to the company.

Reports differ on the operator interface. Robotics & Automation News says it uses a VR headset and force-feedback haptic gloves. Tectonic Defense says gloves are coming "pretty soon."

The company says operators can learn full-body control in minutes. That is a company claim; no training study has been published.

What is Minerva's thesis?

Minerva's bet is that a humanoid body is the right shape for places built for humans, and that a human in the loop is the right brain for now. CEO Sandor Felber told General Catalyst: "Literally every single gauge, every single pipe, every single valve and lever has been designed with human height and force in mind."

The company frames Roger as a safety tool, not a labour replacement. In the Robotics & Automation News coverage, Felber says: "We didn't set out to replace the people doing them." Tectonic Defense paraphrases the robot as designed to replace humans in "the world's most dangerous work", but "replace humans" there is the reporter's wording, not a direct quote.

On bomb disposal, Dominic Asquith, a British Special Forces veteran who is Minerva's head of hazardous operations, describes the problem on the General Catalyst page. Tracked, single-arm robots send people on what EOD operators call "the long walk" because many tasks need two hands and real dexterity.

The long-term plan is more autonomy. Felber says that "at the moment, we do all of this with human supervision in place," and that he sees one human supervisor eventually overseeing a fleet of robots before full autonomy.

How is Minerva funded?

Minerva announced about $10M in pre-seed funding on 6 October 2026. General Catalyst led the round. Long Journey Ventures and Credo Ventures are co-leads. Valuation has not been disclosed.

Robotics & Automation News and The Next Web both name Hugging Face, Codecademy and Compound among the other backers.

One investor-side data point is worth noting. General Catalyst's Robin Dechant says that "a lot of people on our team were quite skeptical about the humanoid form factor" when the deal was considered. The firm backed it anyway.

What are Roger's specifications?

Almost none are published. Interesting Engineering reports that Minerva has not disclosed payload, battery life, operating speed or manipulation accuracy. The table below lists what the company has and has not said.

Felber did say on the record, per Tectonic Defense, that the robot costs more than consumer humanoids but "a lot less than human life." No number accompanied that.

How does Roger work?

Roger splits the job between a person and software. The person supplies intent and fine manipulation. The software supplies the parts that humans do badly through a headset: staying upright, recovering from a fall and walking.

Tectonic Defense quotes Felber describing the mapping: if the operator crouches, the robot crouches the same way. He calls the arrangement "shared autonomy." Tectonic also reports a prototype history of three builds, Roger 1, 2 and 3, moving from a mechanical prototype to an electromechanical one. That detail comes from one outlet.

The company says it plans a supply chain that does not depend on single-source Chinese components. Training data and model architecture for Minerva Intelligence are not disclosed.

How does Roger compare with other humanoids?

Minerva has published too little to compare Roger numerically with other humanoids. The pitch is a remote-presence tool for hazardous sites: a specialist, a robot body and a hazard. Tectonic Defense also reports that Minerva pitches an entirely Western supply chain, and Felber says the robot costs more than consumer humanoids (see above). Our posts on the Unitree G1+, the Galbot ET1 and Agility's Digit 5 cover other humanoids with published details.

Where will people actually use this?

Minerva names four settings, all described by the company, none yet in paid service.

1. Bomb disposal. An operator guides Roger to a suspicious device and handles it with two hands, instead of walking up in a protective suit. Minerva says it showed its technology at the NATO Explosive Ordnance Disposal Centre of Excellence in Slovakia before launch. No outcome was reported.
2. Hazmat response. Responding to hazardous material incidents from a distance.
3. Oil and gas. Offshore rigs and onshore plants, where Felber says valves, gauges and levers are sized for humans. He describes inspection and mapping rounds as things robots may eventually do autonomously.
4. Public safety at airports. The company says it is exploring deployments at two of the world's 10 busiest airports. It has not named them. Interesting Engineering reports that on-site testing for public safety applications was done at one of them.

Research partners are the Mohamed bin Zayed University of Artificial Intelligence, working on energy-sector use, and the University of Waterloo, working on robot vision in smoke and darkness.

What are the limitations of remote-controlled humanoids?

The main limits are delay, dexterity and operator load. A 2026 review in Complex Engineering Systems cites human-subject studies finding that round-trip delays of about 300 ms already degrade task execution, and that delays beyond 700 ms may make remote operation nearly infeasible. Delayed or intermittent video also lowers situational awareness.

That matters for an offshore rig or a bomb site with poor links. It is also why Minerva keeps balance and navigation on board the robot rather than in the operator's hands.

The General Catalyst page adds a human-factors argument. It cites Carlson and Murphy (2004) on field robots, with human failures roughly once every 17 minutes of robot use against mechanical failures every 24 hours. That is the investor's framing of an older study, so read it as motivation, not proof that Roger fixes the problem.

What's still unclear?

• Performance. There are no published specs, no independent test and no footage-based measurements of task speed. Every capability is company-reported.
• Customers. The first paid pilot is described as launching in autumn 2026, but no customer or airport is named.
• The 7,000 figure. Minerva estimates that about 7,000 energy workers die on the job each year, equivalent to roughly one every 75 minutes, per Robotics & Automation News. The company built it by applying the ILO's global fatal accident rate to the IEA's 76 million energy workers. It is an extrapolation, not a count.
• Source conflicts. Outlets disagree on headquarters (San Francisco versus Germany), on whether haptic gloves are in use now, and on the investor list. Tectonic Defense also reports FBI and NATO DIANA testing and a pending Ukrainian defence ministry pilot, which no other outlet repeats, so treat them as unconfirmed.
• The weakest point. A human-operated robot is only as good as the link and the operator. If autonomy does not improve, the service looks like an expensive version of a person at a console. If it does, the safety argument and the "not replacing workers" framing will need revisiting.

What does this mean for the industry?

Minerva is a test of whether the first paying humanoid market is danger, not convenience. If the pilots work, teleoperation with shared autonomy could become the practical route to humanoids in high-risk jobs, while fully autonomous machines keep maturing.

The funding also shows investor interest in Western-built hardware. General Catalyst's Dechant told Tectonic Defense that China is becoming the "center of gravity" of robots and that "we need a sovereign robot." Whether a $10M pre-seed company can ship a certified product for explosive and offshore environments is the open question. Watch for named pilots, a spec sheet and any third-party test.