Figure Robot Hand: Palm Cameras, Tactile Sensing and Helix

A source-checked guide to Figure robot hand, covering how it works, verified evidence, comparison methods, failure modes, practical uses and missing data.

Introduction

Figure 03 places cameras in the palms and adds tactile sensing to the hand, changing what the robot can observe during close manipulation. Those features are documented, but public task videos still require control-mode and repeatability analysis. The Figure robot hand is a five-finger end-effector integrated with the Figure 03 humanoid and its Helix control system. Palm cameras provide viewpoints that remain close to the object during reaching and grasping. Tactile sensors provide contact information. Neither feature alone proves reliable autonomous manipulation. This article explains the mechanisms behind Figure robot hand, compares documented systems, separates real-robot evidence from claims and identifies the measurements that remain missing. The analysis treats kinematics, sensing, actuation and demonstrated task performance as separate layers. It avoids ranking hands by appearance or joint count alone.

Key findings

  • The official product announcement documents palm cameras and a first-generation tactile system.
  • Use head cameras for scene context and palm cameras for near-field geometry.
  • Palm images can be blurred by motion, occluded by the object or affected by lighting.
  • Industrial part placement and material handling.
  • Exact hand DoF, actuator layout, tactile taxel count and independent durability data are not public.

Figure Robot Hand: Palm Cameras, Tactile Sensing and Helix — evidence comparison

The table records what each source establishes and keeps missing data visible.

System or methodWhat the evidence establishesEvidence classMain unresolved point
Figure 03 hardwareThe official product announcement documents palm cameras and a first-generation tactile system.Officially documentedExact hand DoF, actuator layout, tactile taxel count and independent durability data are not public.
Helix demonstrationsFigure has published autonomous manipulation demonstrations tied to its vision-language-action system; conditions and edited segments must be read carefully.Company demonstrationFigure's production and sensitivity figures are company-reported.
BMW workFigure reports production work at BMW, including part handling and thousands of operating hours, as company-reported deployment evidence.Company-reported industrial evidencePublic videos do not expose every intervention, reset or failed trial.

Definition and design boundary

The Figure robot hand is a five-finger end-effector integrated with the Figure 03 humanoid and its Helix control system. Palm cameras provide viewpoints that remain close to the object during reaching and grasping. Tactile sensors provide contact information. Neither feature alone proves reliable autonomous manipulation. The scope used here excludes adjacent systems that share vocabulary with Figure robot hand but do not perform the same function. The boundary prevents a perception model, simulation result, component price, historical prototype or edited demonstration from being presented as evidence for a complete deployed system.

How the hand architecture works

Use head cameras for scene context and palm cameras for near-field geometry. Fuse visual observations with joint state and tactile contact. Generate hand, wrist and arm actions through the robot policy. Maintain closed-loop correction as objects become occluded by the fingers. Measure recovery after missed grasps rather than only successful clips. The pipeline remains closed loop: sensing updates the state estimate, the controller selects or constrains an action, the robot executes it and new observations determine whether to continue, correct or stop. Latency, calibration and safety limits can change the result even when the high-level model remains the same.

What public evidence shows

Figure 03 hardware: The official product announcement documents palm cameras and a first-generation tactile system. This is classified as officially documented. The classification records what the source establishes and leaves unstated fields as not publicly disclosed. It should not be extended to different robot versions, sites or tasks without new evidence.

Helix demonstrations: Figure has published autonomous manipulation demonstrations tied to its vision-language-action system; conditions and edited segments must be read carefully. This is classified as company demonstration. The classification records what the source establishes and leaves unstated fields as not publicly disclosed. It should not be extended to different robot versions, sites or tasks without new evidence.

BMW work: Figure reports production work at BMW, including part handling and thousands of operating hours, as company-reported deployment evidence. This is classified as company-reported industrial evidence. The classification records what the source establishes and leaves unstated fields as not publicly disclosed. It should not be extended to different robot versions, sites or tasks without new evidence.

How to compare dexterity claims

The analysis treats kinematics, sensing, actuation and demonstrated task performance as separate layers. It avoids ranking hands by appearance or joint count alone. A defensible comparison records the exact system version, task, environment, control mode, trial count and source date. Published numbers are retained only when the source defines what was measured. Missing fields remain marked as not reported rather than estimated.

Failure modes during manipulation

The main failure modes are concrete: Palm images can be blurred by motion, occluded by the object or affected by lighting. Tactile sensitivity claims do not specify full skin coverage or standardized test methods. A policy can fail after an object shifts outside training distributions. Cycle-time evidence from one task does not transfer to fine domestic manipulation. A useful evaluation records the state before the failure, the intervention required, the recovery time and whether the same failure repeats after a reset.

Credible applications today

Credible applications include Industrial part placement and material handling, Bimanual household tasks demonstrated by Figure and Data collection for Helix and hand-eye coordination research. These applications should be described with the robot, task boundary, operator role and environmental constraints. Experimental capability, commercial availability and routine deployment are reported as separate statuses.

Questions buyers and researchers should ask

A buyer, developer or researcher should ask for the exact hardware and software version, raw trial counts, intervention logs, control frequency, safety limits, maintenance requirements and licensing terms. The answer should identify which results were obtained in simulation, on one physical robot, across several embodiments or in an operational site. A missing answer is itself useful evidence about maturity.

Limitations and missing information

  • Exact hand DoF, actuator layout, tactile taxel count and independent durability data are not public.
  • Figure's production and sensitivity figures are company-reported.
  • Public videos do not expose every intervention, reset or failed trial.
  • Specifications, prices, repositories and deployment status can change after publication.
  • Benchmarks from different robots or environments are not directly comparable.

Conclusion

The strongest conclusion about Figure robot hand comes from the evidence boundary, not the most impressive clip. The official product announcement documents palm cameras and a first-generation tactile system. At the same time, exact hand dof, actuator layout, tactile taxel count and independent durability data are not public. Practical value is clearest in industrial part placement and material handling, bimanual household tasks demonstrated by figure. Deployment or adoption should therefore depend on repeated task results, disclosed intervention, safe fallback behavior and a complete cost or maintenance model. Where sources omit a number, the article leaves it undisclosed rather than converting a claim, target or partial test into a precise fact.

Frequently asked questions

What does Figure robot hand mean?

The Figure robot hand is a five-finger end-effector integrated with the Figure 03 humanoid and its Helix control system. Palm cameras provide viewpoints that remain close to the object during reaching and grasping. Tactile sensors provide contact information. Neither feature alone proves reliable autonomous manipulation. The article uses this definition to exclude neighboring technologies or claims that do not meet the same evidence threshold.

How should Figure robot hand be evaluated?

It is evaluated by recording Use head cameras for scene context and palm cameras for near-field geometry, Fuse visual observations with joint state and tactile contact, Generate hand, wrist and arm actions through the robot policy. The system version, environment, control mode, trial count, intervention rate and failure recovery must be disclosed before results can be compared.

What real-world evidence is available?

Public evidence includes Figure 03 hardware, where the official product announcement documents palm cameras and a first-generation tactile system. It also includes Helix demonstrations, where figure has published autonomous manipulation demonstrations tied to its vision-language-action system; conditions and edited segments must be read carefully. Each result remains limited to the published robot, task and conditions.

What information is still missing?

The largest limitations are exact hand dof, actuator layout, tactile taxel count and independent durability data are not public, figure's production and sensitivity figures are company-reported, public videos do not expose every intervention, reset or failed trial. These gaps prevent a precise universal ranking and can change the engineering or commercial conclusion for a specific robot, country, task or workplace.

Is the technology ready for practical use?

Current credible uses include industrial part placement and material handling, bimanual household tasks demonstrated by figure, data collection for helix and hand-eye coordination research. Readiness depends on repeated real-world performance, safety controls, human intervention, maintenance and cost. A single successful demonstration is insufficient evidence of routine deployment.

Sources and methodology

The analysis treats kinematics, sensing, actuation and demonstrated task performance as separate layers. It avoids ranking hands by appearance or joint count alone.

Sources were checked on July 11, 2026. Official product pages, research papers, repositories, standards and customer documents were prioritized. Company metrics remain labeled as company-reported unless an independent source establishes the same result.

  1. Introducing Figure 03 — Figure AI · October 9, 2025
  2. F.02 Contributed to the Production of 30,000 Cars at BMW — Figure AI · November 19, 2025
  3. Figure 03 at BMW — Figure AI · June 30, 2026
  4. Helix — Figure AI · February 20, 2025 · accessed July 11, 2026
  5. Helix 02 — Figure AI · 2026
  6. NEO hands — 1X Technologies · July 9, 2026

Related TechniaHQ guides

Official image recommendations

Fact-check report

Verified: July 11, 2026

Confirmed

  • The official product announcement documents palm cameras and a first-generation tactile system.
  • Figure has published autonomous manipulation demonstrations tied to its vision-language-action system; conditions and edited segments must be read carefully.

Not confirmed or incomplete

  • Exact hand DoF, actuator layout, tactile taxel count and independent durability data are not public.
  • Figure's production and sensitivity figures are company-reported.
  • Public videos do not expose every intervention, reset or failed trial.

Fast-changing information

  • Commercial availability, prices, model versions and software access.
  • Deployment counts, company partnerships and repository maintenance status.