Figure AI — Complete Profile & Analysis 2026

Company Overview

Figure AI has moved faster than any humanoid robotics company in history. Founded by Brett Adcock in January 2022, the company went from a blank sheet of paper to a $2.6 billion valuation, a working humanoid robot, a BMW manufacturing partnership, and one of the largest funding rounds in robotics — all in under three years.

Brett Adcock's background explains the pace. Before Figure, he founded and ran Archer Aviation (eVTOL/air taxi company, NYSE: ACHR), where he demonstrated the ability to move a deep-tech hardware company from concept to public markets at speed. Adcock brings a founder archetype that is rare in robotics: someone who has already navigated the intersection of hardware complexity, regulatory challenge, and venture-scale fundraising.

Figure's strategy is to build a general-purpose humanoid robot and deploy it first in manufacturing environments where the ROI case is clearest. The BMW partnership — deploying Figure robots in BMW's Spartanburg, South Carolina plant — is the signature expression of this strategy. Rather than trying to build a humanoid that does everything, Figure is focusing on proving commercial value in one of the world's most demanding industrial environments.

In 2025, Figure announced Helix — their proprietary Vision-Language-Action model — signaling a strategic shift from dependence on the OpenAI partnership toward full-stack ownership of both hardware and AI. This gives Figure end-to-end control of the robot experience, from language understanding through motor execution.

Product Portfolio

Source: Figure AI disclosures, SVRC Research assessment

Figure 02: The Manufacturing-Ready Humanoid

Figure 02, unveiled in August 2024, represents a substantial leap from the original Figure 01 prototype. The most significant upgrades are in the hands (16 degrees of freedom enabling dexterous manipulation), onboard compute (local inference without cloud dependency), and battery life (estimated 5 hours vs. 2 hours for Figure 01). These improvements directly address the requirements of the BMW manufacturing deployment: Figure 02 needs to manipulate parts, operate autonomously for full shifts, and make decisions locally without network latency.

Helix VLA Model

Announced in 2025, Helix is Figure's proprietary Vision-Language-Action model. While Figure's earlier demos relied on OpenAI's language models for reasoning, Helix gives Figure end-to-end ownership of the AI stack. Helix is optimized for the specific perception and manipulation tasks that Figure 02 encounters in manufacturing environments — a narrower but deeper focus than general-purpose VLA models like Pi0. Running on Figure 02's onboard compute, Helix enables real-time language-conditioned manipulation without cloud dependency.

Funding & Valuation Timeline

Source: Crunchbase, PitchBook, company announcements

Source: Crunchbase, PitchBook, SVRC Research

Technology Deep-Dive

Hardware Design Philosophy

Figure's hardware design prioritizes commercial viability over research-grade performance. The robot is sized and proportioned to work in existing human workspaces without modification. At ~170 cm and ~60 kg, Figure 02 fits through standard doorways, reaches standard workbenches, and operates alongside human workers without significant facility changes. This pragmatic design philosophy reflects Brett Adcock's experience with Archer Aviation — where certifiability and deployability mattered more than peak performance.

Dexterous Hands

Figure 02's hands are one of its key differentiators. With 16 degrees of freedom per hand, they can perform manipulations that simple parallel grippers cannot — grasping irregular objects, operating human tools, and handling delicate components. The hands are designed for the specific manipulation tasks encountered in automotive manufacturing: handling body panels, manipulating wiring harnesses, and placing components in assemblies.

OpenAI Partnership & Helix

Figure's early demos were powered by an integration with OpenAI's language models, enabling the robot to understand and respond to natural language instructions. The partnership demonstrated that language-conditioned robot control was possible at a compelling level. However, dependence on a third-party AI provider introduces latency, cost, and strategic risk. The development of Helix — Figure's own VLA model — addresses all three concerns and gives Figure full ownership of the AI stack.

Neural Network Motor Control

Figure's proprietary neural network controls the robot's actuators in real-time, translating high-level task descriptions (from Helix or OpenAI) into precise motor commands. The controller handles balance, gait, and manipulation simultaneously, using reinforcement learning policies trained in simulation and refined on real hardware during deployment.

Source: Public announcements, SVRC Research timeline analysis

Market Position & Competitive Landscape

Competitive Strengths

• Execution speed: Fastest from founding to billion-dollar valuation in humanoid robotics history. Brett Adcock's serial founder experience is a genuine competitive advantage.
• BMW partnership: The most credible industrial humanoid deployment in the US. BMW's manufacturing standards provide external validation that no amount of demo videos can match.
• Full-stack ownership: With Helix, Figure controls both hardware and AI — a position that only Figure and Tesla occupy among major humanoid companies.
• Investor syndicate: Microsoft, OpenAI, Nvidia, Bezos, Intel — this investor base provides strategic partnership value beyond capital.
• Dexterous hands: 16-DoF hands are among the most capable on any commercial humanoid platform.

Competitive Vulnerabilities

• Unproven at scale: BMW is a pilot, not a fleet deployment. The gap between pilot success and scaled production is enormous in manufacturing robotics.
• Pricing unknown: Figure has not publicly disclosed pricing. At current production volumes, per-unit costs are almost certainly too high for broad commercial deployment.
• Cash burn: Humanoid hardware development is capital-intensive. $675M is substantial but not unlimited — the company needs to show commercial traction before the next raise.
• Competition from China: Unitree's G1 at $16K and H1 at $90K represent price points that Figure cannot match in the near term.

SVRC's Assessment

Figure's greatest strength is speed and conviction. In a field where most companies spend years in stealth before showing anything, Figure moved from founding to a working humanoid, a BMW deployment, and a proprietary AI model in roughly 36 months. This pace is a direct reflection of Brett Adcock's operating style — move fast, raise aggressively, secure the most credible possible deployment partner early.

The BMW partnership is genuinely important. BMW's Spartanburg plant is one of the most demanding manufacturing environments in the world. If Figure 02 can perform useful work there — even at a small scale — it validates the humanoid form factor for industrial use in a way that no research demo or controlled environment test can. This is a signal that the broader automotive and manufacturing industry is watching closely.

The development of Helix is strategically sound. Dependence on OpenAI for the core AI capability introduces risk that no hardware company should accept long-term. By building their own VLA model, Figure gains full-stack control and the ability to optimize the entire system — from perception through planning through motor execution — as a unified architecture. This is the same vertical integration logic that has driven Tesla's AI strategy.

The risk is that Figure has raised at a valuation that requires rapid commercial scaling. $2.6B demands either a clear path to hundreds of millions in revenue or continued fundraising at ever-higher valuations. The humanoid market is not yet large enough to support these valuations on revenue alone, which means Figure needs to keep executing at a pace that maintains investor confidence while simultaneously proving manufacturing-grade reliability.

Frequently Asked Questions