2026 report
State of Physical AI Jobs 2026
Physical AI has moved from research vocabulary into hiring reality. The companies in this market are looking for people who can make intelligent systems work outside the lab: on robots, vehicles, drones, factory floors, warehouses, and simulation pipelines that support real deployment.
This report looks at active public listings tracked by Physical AI Jobs to understand where the market is hiring, which roles are most common, where the jobs are located, and what compensation looks like when companies publish salary data.
Last updated: October 7, 2026
Open roles
4,729
Companies hiring
110
Hiring cities
312
Median salary
$197,500
Executive summary
The Physical AI labor market is still engineering-led. Hiring is concentrated around the people needed to turn models, sensors, actuators, and simulation environments into deployed systems: robotics software engineers, autonomy engineers, perception teams, controls specialists, SLAM engineers, and simulation infrastructure builders.
The shape of the market is also practical. These are not only frontier research roles. Many listings point to production work: debugging robots, scaling autonomy stacks, building data and simulation tooling, validating behavior, and integrating software with hardware constraints.
Key findings
- The market is broad enough to support specialized job searches by role area, location, company, and remote status.
- Robotics and autonomy remain the center of gravity, but simulation, controls, perception, ROS, and motion planning are visible enough to deserve their own career tracks.
- Hiring is geographically concentrated, but not limited to one hub. The strongest cities matter, yet the company list shows demand across several Physical AI ecosystems.
- Remote work exists, but the field still has a strong onsite and hybrid bias because many teams need access to hardware, labs, vehicles, or test facilities.
- Published salary data suggests senior Physical AI engineering work is a high-compensation market, but salary coverage is incomplete because many employers do not disclose pay.
How to use this Physical AI hiring report
Treat the report as a map of the market, then move into the live pages when you need current openings. The report is a snapshot of tracked public hiring; role pages, company pages, salary tables, and live trends let you inspect the underlying market from different angles.
Choose a technical track
Compare broad categories with specialist role pages instead of relying on one job title.
Compare employers
Use company pages to compare active role mix and move from aggregate market data to a specific hiring team.
Browse companies hiringCheck the live market
Use live trends for the current inventory snapshot and salary tables when compensation is disclosed.
Market size snapshot
The simplest way to read the market is this: Physical AI is no longer a single job category. It is a cluster of engineering markets that overlap around robotics, autonomy, simulation, and machine perception. The open-role count shows enough depth for candidates to specialize instead of searching only for broad AI roles.
Unspecified
2,689
57% of open roles
Onsite
1,444
31% of open roles
Hybrid
441
9% of open roles
Remote
155
3% of open roles
Hiring categories
The category mix shows where companies are investing. Some teams are hiring for complete robotics systems, while others are hiring narrow specialists in controls, perception, simulation, mapping, and planning. For candidates, that means both generalist robotics engineers and deep subsystem specialists have a path into the market.
| Category | Open jobs | Share |
|---|---|---|
| Physical AI Jobs | 3,549 | 75% |
| Autonomy Jobs | 3,325 | 70% |
| Robotics Jobs | 3,097 | 65% |
| Simulation Jobs | 1,996 | 42% |
| Robot Perception Engineer Jobs | 1,540 | 33% |
| Controls Engineer Jobs | 1,204 | 25% |
| Embodied AI Jobs | 927 | 20% |
| SLAM Engineer Jobs | 595 | 13% |
| SLAM Jobs | 595 | 13% |
| Motion Planning Jobs | 581 | 12% |
| ROS Jobs | 436 | 9% |
| Robotics Engineer Jobs | 301 | 6% |
A useful pattern for job seekers is to search both broad and narrow terms. A robotics engineer may find relevant roles under robotics, autonomy, ROS, controls, or simulation depending on how each company describes the work.
Top companies hiring
The company list is a good proxy for where Physical AI is being operationalized. Some employers are building robot products directly. Others provide platforms, compute, simulation, vehicles, sensors, or infrastructure that make physical AI systems possible.
For candidates, the practical takeaway is to look beyond companies that describe themselves as robotics companies. Physical AI hiring also appears inside automotive, logistics, defense, simulation, semiconductor, manufacturing, and AI infrastructure teams.
Geography
Physical AI has a stronger location component than pure software. Hardware access, test sites, labs, factories, fleet operations, and vehicle programs all shape where companies hire. Remote roles exist, but dense local ecosystems still matter.
Top countries
The strongest locations are useful signals for career planning. If someone wants to break into the field, being near a dense robotics or autonomy cluster can increase the number of relevant interviews, even if some companies support hybrid work.
Compensation
Among listings with usable annual USD compensation, the median salary is $197,500 and the 75th percentile is $234,500. Salary coverage is partial because many companies do not publish compensation.
The salary data fits the technical demands of the field. Companies are paying for engineers who can combine software depth with real-world constraints: latency, calibration, safety, sensor quality, controls stability, simulation fidelity, and field debugging. Those skills are harder to substitute than general software experience.
Explore salary tablesSkills in demand
Skills in Physical AI tend to cluster around systems that touch the real world. The most valuable candidates usually connect multiple layers: software, sensors, controls, perception, simulation, data, and deployment tooling.
What this means for candidates
The strongest applications will usually show evidence of working systems, not only model knowledge. Projects that involve robot middleware, sensor integration, simulation, controls, perception pipelines, or field testing are especially relevant.
Candidates should search across adjacent titles. A role called robotics software engineer, autonomy engineer, perception engineer, simulation engineer, controls engineer, or SLAM engineer may all sit inside the same Physical AI hiring market.
For people coming from machine learning, the biggest gap is often deployment context. Employers want to know whether the candidate can handle noisy sensors, incomplete data, physical constraints, and systems that fail in ways pure software products do not.
What this means for employers
The talent market is competitive because many companies are hiring for overlapping skill sets. Robotics software, perception, simulation, controls, autonomy, and ML infrastructure candidates can often interview across several industries at once.
Clear job descriptions matter. The best Physical AI candidates need to understand the hardware context, autonomy stack, simulation environment, languages used, deployment expectations, and whether the role is research-heavy or production-heavy.
Compensation transparency also matters. Salary-bearing postings give candidates useful signal and make roles easier to compare in a market where job titles are not standardized.
Methodology
The report uses approved public jobs currently listed on Physical AI Jobs. Rejected, expired, and non-public jobs are excluded. Topic groupings come from the site’s topic-matching layer. Location counts use normalized job-location data.
Salary analysis includes annual USD salary-bearing listings only, uses midpoint values for ranges, and filters obvious outliers. This is not a complete labor-market census. It is a live snapshot of the public Physical AI roles tracked here, intended to show directionally where hiring activity is concentrated.
The report will become more useful as it is refreshed over time. Historical snapshots can add true trend lines for open roles, category growth, remote work, and compensation movement.