Skip to main content
robotics

South Korea Has 1,012 Robots per 10,000 Workers. Hyundai Wants Thousands More — and the Engineers Don't Exist Yet

By Priya Nair

SoftBank cashes out, Hyundai takes full control

Hyundai Motor Group is buying SoftBank's remaining 9.65% stake in Boston Dynamics for $325 million, turning the Waltham, Massachusetts robotics firm into a wholly owned subsidiary. The purchase, expected to gain board approval on June 22, closes out the put option SoftBank retained from the 2021 transaction in which Hyundai acquired an 80% controlling stake for roughly $880 million, a deal that valued Boston Dynamics at about $1.1 billion.

That 2021 valuation reads like a marker from a different era. SoftBank had acquired Boston Dynamics from Alphabet in 2017, after Google's own 2013 purchase of the lab. The company's robots were famous on YouTube long before they became commercial products. The $325 million price tag for the final stake is a rounding error against Hyundai's broader capital spend. The signal is the ownership structure itself: Hyundai wants no partner at the table when it decides what Boston Dynamics builds, where it deploys, and how fast it scales.

The timing is not accidental. At CES in January, Hyundai and Boston Dynamics showed the electric Atlas humanoid robot in public. Behind the stagecraft sits a deployment plan. A production version of Atlas is expected to begin work at Hyundai's electric vehicle plant near Savannah, Georgia, by 2028. Hyundai plans to start Atlas with parts sequencing at its Georgia Metaplant, then move toward heavier operations by 2030.

Hyundai's advantage is that it doesn't have to imagine the first customer. It owns the factories, the vehicle programs, and now the entire robotics company. Boston Dynamics CEO Robert Playter told Business Insider in January that Atlas would need to learn new factory tasks in a day or two and reach 99.9% reliability before it could be genuinely useful on the floor. That bar is high. It is also the right one. Hyundai Mobis, the group's components arm, has been tied to actuator production for Atlas, keeping one of the robot's most critical hardware systems inside Hyundai's own industrial base.

SoftBank's exit reflects its own priorities. SoftBank is forming Roze AI, a venture aimed at using AI and robotics to build physical infrastructure including data centers. Masayoshi Son is targeting a $100 billion valuation for Roze and a public listing as soon as this year. The $325 million from the Boston Dynamics sale is seed capital by comparison. Son is not walking away from robotics as a concept. He is moving toward robots as part of an AI infrastructure buildout, and Boston Dynamics, a product company with hard engineering problems and a slower revenue curve, no longer fits that thesis.

Hyundai wants the robot on the factory floor. That is a narrower bet, but it is easier to judge. By 2028, Atlas is supposed to be doing real work in Georgia, not walking across a stage in Las Vegas. If Hyundai can turn that into repeatable manufacturing value, the SoftBank exit will look less like a tidy financial cleanup and more like the moment Hyundai stopped borrowing a robotics future and decided to own it outright.

Inside Hyundai's Korean robot-factory buildout

Hyundai's robot-production buildout is anchored in South Korea, where the company is retooling existing manufacturing infrastructure to absorb Boston Dynamics' legged and humanoid robot lines. The centerpiece is the Ulsan complex (1,200 acres, five independent plants, over 34,000 workers, output of 5,600 vehicles daily), which Hyundai confirmed will cut new plant capacity by up to 37.5%, down to 250,000 units, and shift toward flexible multi-platform production spanning ICE, hybrid, and electric vehicles. That pivot frees floor space and labor allocation for adjacent programs, and Hyundai's own careers portal (talent.hyundai.com) lists open manufacturing and robotics-adjacent roles tied to the transition, though the company has not publicly broken out how many positions are specifically earmarked for Boston Dynamics robot production versus general EV retooling.

The Seongnam facility, long home to Hyundai's R&D center, is the other node. Industry reporting points to Seongnam as the site where Boston Dynamics' Spot and Atlas production would be localized for the Korean and Asian markets, with Hyundai leveraging the existing engineering workforce rather than building a greenfield plant. No public job posting count specific to Seongnam robot manufacturing is available, but Hyundai's talent portal shows active listings for manufacturing engineering roles in the broader Gyeonggi region, and the company's stated intent to integrate Boston Dynamics into its "robotics cluster" points to Seongnam as the primary production hub outside the Ulsan complex.

Hyundai has not disclosed a headcount target for its robot-production workforce. What is visible: the Ulsan plant's 34,000-person workforce is being restructured rather than expanded, the Seongnam R&D center is absorbing robotics production engineering, and Hyundai's careers portal shows a steady cadence of manufacturing and test roles that did not exist before the Boston Dynamics acquisition closed. The company's shift to flexible production at Ulsan, cutting 37.5% of new-plant output capacity, is the financial signal that robot manufacturing is being folded into existing lines rather than funded as a standalone capital project.

What Boston Dynamics' job postings reveal about industrializing physical AI

The job postings on Boston Dynamics' careers page read less like a robotics startup's wish list and more like an auto manufacturer's production playbook. That's the point. Hyundai didn't buy Boston Dynamics to keep building robots in a Waltham lab. It bought them to build tens of thousands. The open positions make that ambition concrete and expose the gap between a prototype that works on a demo floor and a product that survives a factory shift.

Start with the titles. A year ago, Boston Dynamics' hiring leaned heavily toward research: perception scientists, control theorists, motor designers. Those roles still exist. But the current listings are dominated by a different breed of engineer. Three separate test roles — Senior Manufacturing Test Engineer, Principal Manufacturing Test Engineer, and Robustness Test Technician III — all full-time, all at the Waltham POST facility. The company isn't just building robots anymore. It's building the systems to verify that every robot works the same way, every time, at volume.

The Senior Manufacturing Engineer posting for Atlas lays out what that actually means. The role calls for someone to develop assembly workflows for electric motors and actuators "from prototype through low-volume production," design custom fixtures and jigs for precision alignment, write standard operating procedures and training materials for assembly technicians, and implement first article inspections and in-process quality checks. The language is straight out of automotive manufacturing: DFM/DFA collaboration, GD&T inspection, capability studies, Lean and Six Sigma methodologies. The job description even explicitly states the role "will interface with our parent company, Hyundai Motor Group, leveraging their deep manufacturing expertise in the automotive sector."

That interface is the whole thesis of the acquisition. Hyundai makes roughly 5 million vehicles a year. Boston Dynamics currently makes robots in the hundreds. Bridging that gap requires people who understand not just how to build a complex electromechanical system, but how to build it repeatedly, with tolerances tight enough that a Spot rolling off a Hyundai line in Ulsan performs identically to one shipped to a refinery in Texas. The Associate Director, Technical Operations — MFG Test role signals this is a management-layer priority, not just a floor-level hiring spree.

Then there's the robustness angle. The Robustness Test Technician III position is the kind of role that doesn't exist at a pure research company. This person stress-tests robots to failure: running Spot's legs through thousands of gait cycles, submerging sensor housings, validating that a humanoid's actuators survive thermal cycling from a Georgia summer to a Minnesota winter. Industrial robots don't just need to work. They need to work 10,000 hours without a field failure. That's a fundamentally different engineering challenge than nailing a 30-second clip for a keynote.

The Senior Staff Optical Sensor Systems Engineer role for Atlas adds another layer. The person in this position calibrates and validates the vision systems that let Spot read an analog gauge at 98% accuracy using Google's Gemini Robotics-ER 1.6 model, up from 23% on the previous model. That statistic only matters if every unit ships with sensors aligned to the same spec. A perception model is only as good as the hardware it runs on, and at production scale, that means process controls, calibration fixtures, and inspection routines that don't yet exist at Boston Dynamics.

Boston Dynamics' careers page shows seven roles added in the past week, spanning manufacturing engineering, test, and software. The shift in hiring mix is the clearest signal yet of what Hyundai's ownership changes: the hardest problem in physical AI isn't the AI. It's building the robot around it, over and over, without the thing falling apart.

Who else is fighting for Korea's robotics engineers

Hyundai's hiring blitz doesn't exist in a vacuum. South Korea is the most automated country on earth, with 1,012 robots per 10,000 manufacturing workers, more than six times the global average, per the International Federation of Robotics. That density has created a labor market where robotics engineers are scarce, expensive, and fought over by every major conglomerate and a government spending serious money to stay ahead.

The Ministry of Trade, Industry and Energy (MOTIE) announced the Fourth Intelligent Robot Basic Plan in January 2024, committing over $2.24 billion across public and private sectors by 2030. The plan targets eight core technologies (servomotors, controllers, reducers, sensors, grippers, autonomous movement software, autonomous operation software, and human-robot interaction systems) with a goal of raising the local manufacturing rate of core robot parts from 44 percent to 80 percent. That build-out requires thousands of hardware and software engineers who can design, test, and produce those components at scale.

On top of that, the Ministry of Science and ICT selected Lotte Innovate to run a 7.25 billion won ($5.5 million) program training master's and doctoral-level AI specialists in physical AI and humanoid robotics, in partnership with KAIST, Yonsei University, and Inha University. The program runs from December 2026 through December 2029. KAIST handles visual, language, and navigation data interfaces. Yonsei covers AI safety, ethics, and edge reasoning. Inha focuses on path searching and anomaly detection. The explicit goal is an industry-specific foundation model built from university expertise and corporate field demand.

The government is also directly subsidizing robotics facility investment. As of May 2026, the Ministry of Trade, Industry and Resources expanded its advanced strategic industry grant program to include robotics and defense, covering 30 to 50 percent of new domestic site and facility investment costs for SMEs. The program backed roughly 120 billion won for 22 corporations in 2025 and increased total support to about 170 billion won in 2026.

Samsung and LG are competing for the same talent pool. Samsung lists AI security policy and AI agent security management roles on its internal careers board. LG Electronics led Bear Robotics' $60 million Series C round, a signal that the company views robotics integration as a strategic capability worth owning. Bear Robotics co-founder Juan Higueros said LG's CEO visited before CES and moved quickly after seeing the company's fleet management technology.

Category Source / Firm Value
Robustness Test Technician III hourly pay Boston Dynamics job posting $37–$41/hr
Bear Robotics Series C round LG Electronics $60 million
Korean robotics market (2024) Goover report ~$895 million
MOTIE Fourth Intelligent Robot Basic Plan (by 2030) Ministry of Trade, Industry and Energy $2.24 billion
Roze AI target valuation SoftBank $100 billion

The Korean robotics market was valued at roughly $895 million in 2024, according to the Goover report, with service robot deployment in restaurants alone jumping from 50 units in 2019 to over 3,000 in 2023. That growth is colliding with a shrinking workforce, as South Korea's fertility rate hit a record low of 0.72 in 2023, which means every major player in the Korean robotics ecosystem is hiring against the same demographic math.

For Hyundai, the Boston Dynamics acquisition and its production hiring surge are one move inside a much larger, already-running war for robotics talent. The question isn't whether Korean companies will keep hiring. It's whether the supply of engineers can keep pace with the government's targets and the conglomerates' ambitions, and whether the robots Hyundai is betting on will be ready before the labor pool runs dry.


Working in robotics? Zero G Talent tracks the openings: browse robotics jobs, openings at Boston Dynamics, and the people building the field.

Ready to Start Your Space Career?

Browse robotics jobs and find your next opportunity.

View robotics Jobs