What is Fourier Intelligence and the GR-3 humanoid?

Fourier Intelligence: institutional facts

Fourier Intelligence is a humanoid robotics company headquartered in Shanghai, China. The company's technical heritage runs through lower-limb exoskeleton and rehabilitation robotics rather than the warehouse-and-factory engineering most US and Chinese humanoid cohort members trace back to. This is editorially distinctive: Fourier's research lineage produces a different set of priors about how humanoid platforms should be engineered (rehabilitation context emphasizes safety-around-humans, fine-grained force control, and human-environment compatibility).

Fourier is privately held. For broader investor-disambiguation context across the humanoid cohort, see best humanoid manufacturer to invest in. For the Chinese humanoid cohort context (Fourier alongside Unitree, UBTech, AgiBot, and several others), see which humanoid robots are American vs Chinese.

GR-3 as third-generation Fourier humanoid

The GR-3 is Fourier's flagship general-purpose humanoid platform, evolved from GR-1 and GR-2 predecessors. The generation-to-generation lineage shapes the verification posture: each generation reflects cumulative engineering refinement rather than a green-field rewrite. Fourier's commercial positioning emphasizes research-platform availability with rehabilitation and assistive-technology context, distinct from peer Chinese humanoid manufacturer positioning:

• Unitree G1 + R1: research-tools pricing strategy (low-cost catalog availability for research and developer customers).
• UBTech Walker S2: factory-and-industrial-pilot positioning (HKEX-listed; BYD / Geely / Foxconn customers).
• AgiBot: emerging-flagship positioning (general-purpose humanoid with consumer + industrial scope).
• Fourier GR-3: rehabilitation-and-medical-context positioning evolving toward general-purpose humanoid.

The four positions illustrate that the "Chinese humanoid cohort" is not a single editorial bloc. Each major Chinese manufacturer operates a structurally distinct commercial thesis; framework readings should preserve this differentiation rather than collapsing the cohort into a generic geographic category.

What the framework verifies and what it does not

Applying DEPLOY's verified-vs-claimed framework to Fourier Intelligence GR-3:

• Company verified: Fourier is a real Shanghai-headquartered company with multi-year operational history in rehabilitation robotics. The transition to general-purpose humanoid is verifiable through product-launch communications and research publications.
• GR-3 product verified at research-platform scale: the platform ships to research and rehabilitation customers; the verification surface is research-customer relationships rather than scaled-throughput commercial deployment.
• Five-tier availability: GR-3 sits at the emerging-manufacturer tier with research-platform availability context. The platform is purchasable via Fourier's commercial channels at research-and-institutional pricing tier; see DEPLOY's pricing page for Fourier GR-3 for consumer-evaluation context.
• Four-tier capability: GR-3 operates primarily at the research-and-demonstration tier (with rehabilitation-and-assistive context as the distinctive demonstration domain). Verified consumer-deployment or scaled enterprise-deployment is not the current state.
• Cap-flag: scaled commercial throughput at the depth of Apptronik Apollo's three-customer enterprise breadth, Agility Digit's GXO Flowery Branch single-customer 100,000-tote depth, or Figure 02 at BMW Spartanburg's automotive-OEM acceptance has not landed for GR-3. The cap-flag is the editorial truth, not a gap.

Rehabilitation positioning matters editorially

Fourier's rehabilitation heritage is not a marketing footnote. The research lineage produces engineering priors that distinguish GR-3's product positioning:

• Safety around humans is the design starting point, not an afterthought. Rehabilitation context requires force-control, fall-prevention, and human-environment compatibility that warehouse-and-factory engineering does not always foreground.
• Assistive-task context shapes capability framing. The GR-3 product hypothesis tilts toward human-assistance rather than human-replacement, which matters for how the platform's task envelope is articulated.
• Generation-to-generation iteration emphasizes engineering refinement over green-field rewrites. This is a different commercial trajectory from US peers that founded the company specifically to ship the current product (Figure AI, Apptronik) versus Fourier's prior product line evolved into humanoid context.

Whether the rehabilitation heritage produces commercial advantage at deployment scale is a forward operator question. The framework reads Fourier's positioning as a distinct editorial posture within the Chinese humanoid cohort, with verification surfaces accumulating per deployment customer rather than at aggregate-throughput depth.

Where to go for context

For consumer-evaluation context on Fourier GR-3, see DEPLOY's pricing page for GR-3. For canonical institutional depth at the registry layer (founding history, leadership, deployment record, source-depth verification across available sources), see Fourier Intelligence's registry record.

For Chinese humanoid cohort comparison context, see Unitree G1 + R1 pricing, UBTech Walker S2, and which humanoid robots are American vs Chinese.

For the broader humanoid availability and capability frameworks, see can I buy a humanoid robot in 2026 and what can humanoid robots actually do today. For methodology canonical references applicable to Fourier GR-3: the 4-way autonomy-boundary taxonomy (Chinese cohort consumer-pricing tier) + the 9-tier source-quality rubric.