Working at Figure AI in 2026: Building Humanoid Robots at $39 Billion

Figure AI is not a typical AI company. While most of the companies in our Culture Directory are building software — models, APIs, developer tools — Figure is building physical machines that walk, grasp, and work alongside humans in real factories. Founded in 2022 by serial entrepreneur Brett Adcock (who previously built Archer Aviation and Vettery), Figure has gone from a three-person team to 619 employees and a $39 billion valuation in under four years.

That velocity is staggering. But velocity in hardware — especially humanoid robotics — carries a different kind of intensity than scaling a SaaS product. The hours are longer, the failure modes are physical, and the margin for error is measured in centimeters and milliseconds. This article digs into what it’s actually like to work inside that pressure cooker, drawing from employee reviews, verified compensation data, and Figure AI’s own public disclosures.

Figure AI at a Glance

Two numbers jump out of that table. The $39 billion valuation makes Figure AI worth more than many publicly traded tech companies — while still being a three-year-old startup with under 700 employees. And the 3.0 work-life balance score, while based on limited data, signals something important: this is a company optimized for speed, not comfort.

The Mission: Why Figure AI Exists

Brett Adcock’s thesis is straightforward: the global economy faces a structural labor shortage that technology hasn’t solved. Warehouses, factories, and logistics networks need millions of workers they can’t hire. Software automation handles information work, but physical tasks — picking, placing, assembling, transporting — still require human hands. Adcock’s bet is that general-purpose humanoid robots will fill that gap, and that Figure AI will be the company that makes humanoid robots commercially viable before anyone else.

That bet attracted backing from some of the most consequential names in tech: NVIDIA, Microsoft, OpenAI, Jeff Bezos, Intel, Samsung, and Parkway Venture Capital all participated in Figure’s funding rounds. The Series C alone exceeded $1 billion in committed capital — the largest single round ever raised by a pure-play humanoid robotics company.

The mission is ambitious enough to attract the kind of talent that doesn’t get excited by incremental improvements. If you work at Figure, you’re trying to build something that has never existed at commercial scale. That matters for culture: mission-driven companies at this intensity level tend to attract high performers who are willing to sacrifice work-life balance for the chance to work on something genuinely historic.

What Figure Has Actually Built

Unlike many AI startups that remain in research mode for years, Figure has moved from concept to commercial deployment with remarkable speed.

• Figure 01 — The proof-of-concept. Demonstrated basic humanoid locomotion and manipulation. Served as the research platform that validated the core approach.
• Figure 02 — The breakthrough. Deployed at BMW’s Spartanburg, South Carolina plant, Figure 02 helped build over 30,000 BMW X3 vehicles. The robot worked 10-hour shifts, five days a week, loading more than 90,000 sheet metal parts with 99% accuracy. This wasn’t a demo — it was a paying customer deployment in a real automotive production line.
• Figure 03 — The production model. Unveiled in October 2025, Figure 03 was redesigned from the ground up for mass manufacturing and home environments. At 5’8” and 61 kg with a 20 kg payload capacity, it’s smaller and lighter than its predecessor. The design includes mesh fabric covering with foam padding (reducing injury risk for human coworkers) and wireless charging integrated into the feet — the robot refuels by standing on a charging pad.
• Helix — Figure’s proprietary vision-language-action model for humanoid control. This is the AI brain that enables Figure robots to understand natural language commands, perceive their environment through six cameras, and execute complex multi-step physical tasks.
• BotQ — Figure’s dedicated manufacturing facility for humanoid robots. Building your own factory to build robots is a massive capital commitment that signals Figure is serious about scale, not just research papers.

BMW has already expanded the partnership. After the Spartanburg success, BMW Group Plant Leipzig in Germany is running a pilot with Figure 03, evaluating where the next-generation robot can create additional value in European production lines. This is the kind of real-world traction that separates Figure from the dozens of humanoid robotics companies that exist primarily in presentation decks.

Glassdoor Ratings Breakdown

Figure AI’s Glassdoor profile has a limited number of reviews, which means these ratings should be interpreted with caution. But even with small sample sizes, patterns emerge.

The pattern is striking. Culture & Values, Compensation, Diversity, and Senior Management all cluster at 4.0 — strong scores that suggest employees generally respect the mission, the leadership, and the pay. But the overall 3.5 and the 3.0 WLB drag the picture down. This is the classic profile of a high-intensity startup: people believe in what they’re building and feel compensated fairly, but the hours take a toll.

For comparison, among AI companies ranked by work-life balance, a 3.0 WLB would place Figure near the bottom. Companies like Notion (4.2 WLB), Linear (4.4 WLB), and HubSpot (4.1 WLB) offer dramatically better balance. But those companies are building software. Building physical robots that need to survive factory floors is a fundamentally different proposition.

Important caveat: These ratings are based on a small number of reviews. We’ll update this analysis as more data becomes available. For the latest verified data, see Figure AI’s full culture profile.

What Employees Actually Say

What employees love

The positive reviews share a common thread: conviction. People at Figure AI believe they’re working on something that will reshape the physical world. The management praise is notable for a startup this young — employees specifically call out data-driven decision-making and lack of micromanagement, which suggests that Brett Adcock’s prior startup experience has translated into a more mature operational culture than most three-year-old companies.

What could be better

The negatives center on two things: hours and consistency. The hours are unavoidable at a company trying to out-execute Boston Dynamics, Tesla Optimus, and a dozen other well-funded competitors. The team-dependent experience is more concerning — it’s a growth problem. When a company scales from 100 to 600 people in two years, not every manager hired will be equally effective. This is worth asking about during the interview process: which team would you join, and who manages it?

Compensation & Equity

Figure AI pays aggressively. In the race for robotics and AI talent, the company competes directly with OpenAI, Anthropic, Tesla, and Google DeepMind for many of the same engineers. The compensation reflects that competition.

Verified compensation data shows median software engineer total compensation around $375K, with Staff-level engineers earning $1M+ in total comp including equity. These numbers put Figure in the top tier of AI company compensation — comparable to frontier AI labs and well above the median for our 118 profiled companies.

The equity story is particularly interesting. At a $39B valuation, Figure AI’s equity is already priced at a premium — but the company’s ambitions extend far beyond the current valuation. If humanoid robots achieve even a fraction of their projected market potential (some analysts estimate a $150B+ market by 2035), early equity holders could see significant upside. The risk, of course, is that hardware companies can burn capital faster than software companies, and the path from $39B private valuation to a successful public offering is not guaranteed.

For a detailed comparison of how Figure AI’s comp stacks up, see our highest-paying AI companies ranking and the AI engineer salary guide by level.

Engineering Culture & Tech Stack

Figure AI’s engineering challenge is uniquely multidisciplinary. Unlike pure software companies, Figure needs engineers who can work across the full stack of physical AI — from low-level motor control firmware to high-level vision-language models.

Core Tech Areas

The engineering organization spans several distinct disciplines:

• Embodied AI. Figure’s Helix model is a vision-language-action system that enables robots to understand natural language, perceive their environment, and plan multi-step physical tasks. This is the frontier of AI research — combining LLM-style language understanding with real-world perception and motor control.
• Hardware engineering. Mechanical design, actuator development, sensor integration, thermal management. Building a humanoid that’s safe to work alongside humans while carrying 20 kg payloads requires solving dozens of simultaneous engineering constraints.
• Manufacturing engineering. BotQ, Figure’s robot factory, is itself an engineering achievement. Designing production processes for humanoid robots at scale is a problem nobody has solved before.
• Simulation & testing. Training robots in simulation before deploying them in the physical world. This includes physics simulation, digital twins of factory environments, and sim-to-real transfer techniques.
• Infrastructure & data. Managing the massive datasets generated by robots operating in real environments — video, sensor telemetry, interaction logs — and building the data pipelines that feed back into model training.

For engineers coming from pure software backgrounds, Figure AI represents a different kind of challenge. The feedback loops are longer (you can’t A/B test a robot the way you A/B test a web page), the failure modes are more dramatic (a bug in a robot can cause physical damage), and the interdependencies between teams are deeper. A change in the AI model might require a change in the sensor configuration, which might require a change in the mechanical housing. This creates a collaborative culture by necessity.

Who Thrives at Figure AI

Based on the culture signals, employee reviews, and the nature of the work, here’s who tends to do well at Figure:

• Mission-driven engineers. If building the world’s first commercially viable humanoid robot doesn’t excite you at a deep level, the hours won’t be worth it. The people who thrive here are genuinely obsessed with the problem, not just collecting a paycheck (even though the paychecks are generous).
• Hardware-software hybrids. The most impactful engineers at Figure are those who can think across the stack — understanding both the AI models and the physical constraints of the robots they control. Pure software engineers can contribute, but the highest-leverage roles require some comfort with the physical world.
• People comfortable with ambiguity. Figure is still defining many of its processes, organizational structures, and technical approaches. If you need a well-documented playbook for everything, a larger company like Databricks or Cloudflare might be a better fit. Figure rewards people who can navigate uncertainty and make decisions with incomplete information.
• Marathon runners, not sprinters. The startup intensity at Figure isn’t a temporary push toward a launch — it’s the sustained pace of a company trying to build something that has never existed. You need to be able to sustain high output over years, not just weeks.

Figure AI is not ideal for people who prioritize work-life balance above all else. The 3.0 WLB score and consistent “long hours” feedback make that clear. It’s also not ideal for people who want a fully remote role — hardware development requires physical presence, and most Figure roles are on-site in Sunnyvale. If remote work is a priority, explore remote-friendly companies in our directory instead.

Figure AI vs. the Competition

Figure AI doesn’t operate in a vacuum. The humanoid robotics space has become intensely competitive, with well-funded players attacking the problem from different angles:

• Tesla Optimus — Elon Musk’s humanoid robot project has Tesla’s manufacturing expertise and financial resources behind it. But Optimus is one division of a massive company, not the entire company’s focus. Figure is a pure-play bet on humanoid robotics.
• Boston Dynamics — The godfather of advanced robotics. Atlas is the most agile humanoid ever built, but Boston Dynamics has historically struggled with commercialization. Figure has the advantage of being built for commercial deployment from day one.
• 1X Technologies — Norwegian humanoid robotics company backed by OpenAI. Strong on the AI side but less advanced in commercial deployment than Figure.
• Agility Robotics (Digit) — Already deployed in Amazon warehouses. A serious competitor, though Digit uses a different form factor (it doesn’t look like a human).

What sets Figure apart is execution speed combined with commercial traction. Building 30,000+ cars at BMW while simultaneously developing a next-generation robot is something none of the competitors have matched. The question is whether that speed can be sustained as the company scales.

Open Positions at Figure AI

Figure AI is actively hiring across AI, robotics, manufacturing, and operations roles. Given the hardware-intensive nature of the work, most positions are based at the Sunnyvale headquarters or the BotQ manufacturing facility. The company hires across experience levels but particularly values candidates with backgrounds in robotics, computer vision, reinforcement learning, mechanical engineering, and manufacturing systems.

Browse Figure AI’s open roles and full culture breakdown on our Figure AI culture profile page, or explore similar engineering-driven companies in our directory.