Put 211 openings on one page and the humanoid company looks different

As of Aug. 26, 2026, Figure's official jobs board shows 127 openings. 1X's careers page lists 84 public openings. Together, just these two humanoid companies account for 211 current postings.

The mix matters more than the headline number. Alongside AI researchers and robotics engineers are manufacturing, quality, test, supply chain, fleet operations, field service, deployment and technician roles. As Physical AI moves from research into products and customer sites, the labor structure around it is widening.

At 1X, Manufacturing Operations is the largest listed hiring cluster

Across 1X's 84 public openings, Manufacturing Operations accounts for 24 roles. Hardware Engineering and Software Engineering have 14 each, while Artificial Intelligence has 6. Fleet Operations has 3 and Supply Chain Operations has 6.

Robot Assembly Technicians, Manufacturing Quality Engineers, Manufacturing Engineers, Test and Validation Supervisors, Production Supervisors and Supplier Development Engineers sit on the same careers page as AI Researchers. Building robot intelligence and repeatedly building reliable robots are becoming separate organizational problems.

Figure's 127 openings show AI scaling alongside manufacturing and commercial operations

Figure's board currently displays 127 jobs. Among the visible major groups are 15 Helix AI roles, 16 BotQ Manufacturing roles and 13 Commercial Operations roles, with separate Actuation Systems and Controls teams as well.

BotQ includes CNC Machinists, Production Associates, Manufacturing Test Engineers, Quality Engineering Technicians and Supplier Quality Engineers. Commercial Operations includes Deployment Engineers, Field Service Technicians, Fleet Coordinators, Humanoid Robot Operators and Site Leads. The model team and the industrialization layer are expanding together.

The more interesting signal is the job that does not require robotics experience

Some Figure mechanical-engineering roles for robot hands explicitly say robotics experience is not required. Instead they ask for injection molding, GD&T, validation, failure analysis, NPI and high-volume production experience. Another robot-hands role points to consumer electronics, EV and aerospace as relevant high-tech backgrounds.

That does not mean an automotive or electronics engineer automatically qualifies for robotics. It does show, in the company's own hiring language, that proven capability in complex hardware design, validation and production can matter more than having 'robotics' in a previous job title.

Physical Intelligence separately hires people who make models work in the physical world

Physical Intelligence's Robotics Software Engineer posting is unusually explicit: the job is not about designing ML models. It is about low-latency runtime, sensor pipelines, networking and real-time systems that make models work reliably on physical robots.

Its Deployments Software Engineer builds safeguards, teleoperation interfaces, observability and reliability for customer-floor deployment. Its Robot Integration Engineer owns fleet readiness across electrical, mechanical and software interfaces. Research and production deployment are becoming distinct career layers inside Physical AI.

That creates adjacent paths for automotive, electronics and industrial automation talent

The recurring language points toward adjacent pools: automotive controls, validation, battery and supplier quality; consumer-electronics NPI, mechanical design and test; industrial-automation commissioning and field service; semiconductor-equipment precision mechatronics, reliability and customer-site support.

The useful translation is not industry-to-industry but capability-to-problem. 'Eight years in automotive' is less specific than the actuator, failure modes, validation environment, production scale and quality thresholds someone has actually owned. Public JDs show the direction of adjacency, not a guaranteed hiring outcome.

211 openings are not the whole market, but they show where the organization is spreading

This is a snapshot of public openings at Figure and 1X on Aug. 26, 2026. One posting does not necessarily equal one hire, private hiring is not visible, and requirements vary by hiring unit. A public JD is not the full internal requirement, and not found does not mean absent.

Even with those limits, the organizational signal is clear: Physical AI talent competition is expanding from a small set of AI scientists into a chain that runs through hardware, manufacturing, quality, fleet operations and deployment.

Physical AI Hiring Watch #1 — next time, the change matters more than the snapshot

This is the first Banseog External Intelligence Physical AI Hiring Watch. A single job-board snapshot is useful; the more valuable question is what appears, disappears or gets reorganized next month.

The next edition will compare new and closed roles, growing teams, new job titles and changing capabilities across the same company set. Subscribe to Intelligence if you want the next Physical AI hiring update delivered to you.

BANSEOG VIEW

Physical AI talent competition no longer stops at the model. Industrialization is widening the job map.

The 211 postings are a point-in-time snapshot of two companies, not the whole market. But the separation of manufacturing, quality, hardware, fleet and deployment into dedicated teams is a useful commercialization signal.

Banseog will track the same company set over time to compare new roles, closed roles and changing capability demand rather than treating any single posting as a market trend.

SOURCES

Primary sources and references

This article is a snapshot of official Figure and 1X career pages and selected official Physical Intelligence postings checked on Aug. 26, 2026. Figure displayed 127 jobs; 1X's current public listings were counted at 84. Posting counts do not equal hires and can change at any time. Same company does not mean same hiring unit; a public JD is not the full internal requirement; not found does not mean absent. No private Banseog candidate, client or search evidence is used.