The Tally at Figure
Hard counts now document the crossover at Figure AI. Late June 2026 brought roughly 740 humanoid robots on-site against about 660 employees. The company crossed into robot-majority in the second quarter, per icharles.com on June 22 and a chart CEO Brett Adcock shared via RoboHorizon showing over 700 machines and staff near 650. ExplainX had earlier logged only 200–250 people, but late-June trackers agree on the 650–660 band. The lower earlier count would have steepened the ratio; both sets confirm machines lead. To keep the fleet learning, Figure is hiring data-operations and global expansion leads rather than more roboticists.
Figure launched in 2022 with cash from Microsoft, Nvidia, Jeff Bezos, and OpenAI, and built almost no robots through 2023 and most of 2024. The line stayed near zero into early 2025, then turned straight up. BotQ, the manufacturing plant, and the October 2025 unveiling of the Figure 03 (a humanoid made for mass production, not show) flipped the curve. Figure ramped from one robot a day to one an hour in 120 days, 24 times faster.
Adcock said in a Sourcery with Molly O'Shea interview that the bottleneck isn't manufacturing but intelligence.
"In my mind, this is not a manufacturing problem. This is an intelligence problem."
He publicly targets fully robot-built robots within 24 months and a path to a million units a year. BotQ will build 12,000 units annually at first, ramping to 100,000 — the scale that pushed the on-site fleet past the payroll.
Figure's 740 machines look tiny beside the global industrial base. Deloitte forecasts in its TMT Predictions report that factories installed over 5 million industrial robots by 2025, heading to 5.5 million in 2026 — about one robot for every 1,600 people on Earth. Those are fixed arms, not humanoids. Deloitte sees only 5,000–7,000 AI humanoids shipped in 2025, rising to 15,000 in 2026 at $14,000–$18,000 each.
| Deloitte humanoid forecast | 2025 | 2026 |
|---|---|---|
| Shipments | 5,000–7,000 | 15,000 |
| Avg. price per unit | $14,000–$18,000 | $14,000–$18,000 |
BMW already runs 40 Figure 03 units at its Spartanburg plant, with phased expansion through 2027, handling parts movement and repetitive assembly beside people, not replacing them. The internal tally — more robots than employees at Figure itself — marks the symbolic flip. RoboHorizon called it "the first concrete step toward the long-theorized 'lights-out' factory, run entirely by machines." Tesla, Agility Robotics, and Apptronik race the same track, but Figure first confirmed the internal crossover.
Figure closed its Series C at $1 billion and a $39.5 billion valuation. Physical AI startups drew $6.4 billion in Q1 2026 alone. Capital bets the humanoid curve bends up now.
More Robots, More Human Demonstrators
Each humanoid leaving BotQ adds a permanent charge to Figure's human-data bill. As the fleet grows, the team that demonstrates tasks to the machines must grow too. A LinkedIn post for an AI Data Operations Lead positions human data as a core strategic function at Figure, with the role owning "a meaningful slice of how we produce human data for our humanoid robots."
Figure built over 350 third-generation humanoids at BotQ, reaching a rate of one per hour, per the company's figure.ai page. Every unit needs training examples before homes or commercial sites can use it. The lead post tasks its owner with hitting "volume, quality, and velocity targets for a specific category of human data" and managing "from sourcing and onboarding contributors to managing the daily operations that keep your pipeline running." More field robots mean more skills in play and more repetitions per skill for clean models.
Robots collect data once shipped, Filemarket AI noted after watching Figure's pre-ship tests. But field signal poorly substitutes for upfront human demonstration. A robot failing on a job generates failure data yet disrupts the customer and cuts useful training hours. Front-loading correct behavior from people costs less, requiring a standing contributor pool.
That pool is the ops hire's core. The post tasks the lead with "contributor acquisition, activation, and retention strategy and execution" and building "operational infrastructure for contributor recruitment, onboarding, daily session management, quality assurance, and incentive management." Those contributors physically show the robot a task. As the fleet expands, the contributor pipeline must expand or velocity targets slip.
The US base salary for the AI Data Operations Lead runs $175,000 to $200,000 annually, five days a week in San Jose. The company puts ops money behind human data, not just mechanical output.
The daily dashboard makes the link visible. The job says the lead will track "volume, quality, and throughput metrics daily, identifying bottlenecks, and shipping operational fixes fast." When robot counts climb, those dashboards demand higher human-data volume. A general-purpose humanoid built for "human-level intelligence" cannot borrow teaching from another machine. It needs people showing it how.
Who Will Open the Next Robot Data Site?
Figure's answer to a fleet that tops its worker count is a hiring packet aimed at geography, not code. The company posted a Lead, Data Operations Global Expansion role on July 11, 2026, matching its careers page. The job sits in North San Jose with a five-day in-office rule and pays $100,000 to $160,000 base. Zero G Talent's board lists Figure AI with 75 open roles; this position sits inside that count.
The role leads the global expansion team that stands up new data operations sites, owning launches from site selection through steady-state handoff to local operators. The hire will identify and launch in new geographies, managing entity setup, localization, facilities, partnerships, and community building.
Data is one of the largest and most strategic functions at Figure, and this role owns how we go from zero to a running site in a new market.
The post frames the work as build-out, not research.
The lead must build and refine the expansion playbook: repeatable systems, SOPs, and tooling that make each launch faster. They own a dashboard tracking volume, quality, and throughput across sites daily, spotting bottlenecks and shipping fixes fast. The post tasks the person with leading launches end to end — entity, facility, hiring, training, hardware, software, partnerships, community — solving zero-to-one problems and coordinating with engineering and product. Coordination, not coding, is the verb.
Figure asks candidates for 5–10 years in operations at startups and global corporations, plus fluency in Spanish and Arabic. Spanish maps to the listed Costa Rica (LATAM) site; Arabic signals an unnamed region. The lead works from North San Jose, then travels to stand up sites.
The pay band sits below engineering posts on the same board. Zero G Talent shows Figure's overall salary spread runs $52k–$400k with a $220k median. The table contrasts a sample of those listings with the ops role.
| Role | Location | Salary band (USD/yr) |
|---|---|---|
| Lead, Data Operations Global Expansion | San Jose / Costa Rica | 100,000–160,000 |
| Helix AI Engineer, Android | San Jose, CA | 150,000–400,000 |
| Helix AI Engineer, Backend Infrastructure | San Jose, CA | 150,000–400,000 |
| Security Engineer, Product and Device Security | San Jose, CA | 150,000–350,000 |
| Software Engineer, Privacy & Data Governance | San Jose, CA | 150,000–350,000 |
Why now? The fleet passed the employee count, and Figure needs local data-capture sites that feed training without shipping everything to Sunnyvale. The post ties the goal to the mission: ship humanoid robots with human-level intelligence. Data sites scale that mission.
Jobsbyculture's July 11 scrape shows competitive equity at Figure, but employee reviews rate work-life balance 3.2 out of 5 and warn of long startup hours. Polarized reviews suggest team-dependent experience. The expansion lead will feel that pace: opening foreign entities is a zero-to-one problem repeated until it becomes a script.
The location list (San Jose, United States, Costa Rica) is short but telling. Costa Rica gives a Spanish-speaking beachhead for Latin America; the Arabic requirement hints at Gulf or North Africa next. The hire owns the playbook so the second site costs less time than the first.
Figure added 11 roles in the past seven days on Zero G Talent's board, including Helix AI Engineers and security posts, but the operations lead multiplies sites. When the lead finishes a handoff in San José, Costa Rica, they board a flight to pitch the next entity. The robot count keeps rising. The ops hires follow.
Ops Hires, Not Coders, Carry the Scale-Up
Figure now fields more machines than staff. The company's answer is a hiring push for global data-operations expansion leads, not a surge of roboticists.
Why does this ops layer define the scale-up? The WEF's 2026 report on organizational transformation (https://reports.weforum.org/docs/WEF_Organizational_Transformation_in_the_Age_of_AI_How_Organizations_Maximize_AI's_Potential_2026.pdf) calls the shift to AI-augmented work "not a primarily technological challenge, but an organizational one." WEF data accompanying the report shows only about one in seven organizations use AI to fundamentally redesign work.
The differentiator is not model sophistication, but the ability to align people, processes and intelligent systems around shared outcomes.
Deloitte's December 2025 tech trends brief (https://www.deloitte.com/us/en/insights/topics/technology-management/tech-trends/2026/ai-future-it-function.html) says competitive edge comes from "building teams that can design, manage, and evolve the way humans and machines work together." Figure already built the machine. The bottleneck is the human-data pipeline that trains it. Standing up sites where local staff perform demonstrations is the organizational fix.
The same logic appears across the humanoid field. Apptronik hired executives from Boston Dynamics and Amazon to scale global robot services, Humanoids Daily reported. Humanoid added two senior leaders and passed 160 staff in engineering, product, and business, according to The Humanoid. Those moves stock the top of the org chart. Figure's posting seeks the field operator who opens the door in a new market.
This hire stays clear of robot control software: the Helix AI Engineer roles that write models moving robot joints stay in San Jose at up to $400k per the board. The expansion lead ships no firmware and racks no power units.
The precise robot-to-human ratio at Figure gets thin coverage beyond the symbolic milestone. The organizational signal is solid. Figure's creation of a hybrid site-launch role places it in that minority. Its LinkedIn post for an AI Data Operations Lead shows it building an operations org to produce human data for humanoid robots. The expansion hire extends that org across borders.
The next step is practical. The Lead, Data Operations Global Expansion will handle entity setup, facility selection, and local partnerships in target regions. Success renders as a running site with a trained local coordinator, not a merged pull request.
Working in robotics? Zero G Talent tracks the openings: see every open Figure AI role, browse robotics jobs, the companies hiring, and the people building the field.