Humanoid robot sales reach 7,000 units worldwide in 2025

Global sales of humanoid robots totaled roughly 7,000 units in 2025, a small early-stage market primarily used for research and data collection rather than productive deployment. The International Federation of Robotics has begun tracking humanoids separately, while automakers test small factory pilots. Bank of America forecasts shipments will rise sharply to 90,000 in 2026 and 1.2 million by 2030, signaling substantial long-term growth potential despite limited current commercial scale.

Analysis

The key investable signal is not unit volume but the gap between experimental deployments and repeatable factory economics. Near-term humanoid purchasing is likely to be funded from OEM R&D and automation capex budgets, making revenue immaterial for most large-cap beneficiaries through the next 1-3 quarters; the relevant earnings sensitivity is therefore concentrated in component suppliers with small revenue bases, where order announcements can move multiples well ahead of shipments. The market should discount aggressive 2030 shipment forecasts heavily until customers disclose cycle-time improvement, uptime, safety certification and labor-cost payback versus fixed automation.

The likely first-order winners are established industrial-automation vendors with installed factory relationships—ABB, FANUY and ROK—rather than pure humanoid narratives, because integration, controls, maintenance and safety systems capture recurring wallet share regardless of robot form factor. NVDA benefits from training and edge-inference demand, but the incremental hardware content per deployed unit remains too uncertain to alter near-term data-center estimates. Auto OEM pilots may ultimately favor incumbent robot suppliers and machine-vision providers over vertically integrated humanoid developers if factories prioritize reliability and serviceability.

Consensus risk is that the shipment curve is being extrapolated from a very low base while productive use cases remain unproven. A 6-18 month de-rating catalyst for speculative robotics exposure would be delayed commercial orders, weak disclosed utilization, or customer migration back to cheaper purpose-built robots; conversely, a named auto OEM converting a pilot into a multi-site order with disclosed payback below two years would validate the category and compress skepticism rapidly. BAC has no material direct earnings linkage; its relevance is limited to research-driven sentiment rather than a fundamental catalyst.

Key Decisions for Investors

• No directional BAC trade: monitor whether BofA publishes investable-company revenue estimates or changes sector ratings; the underlying market forecast alone is not an earnings catalyst for the bank.
• Build a 6-12 month watch-list long in ABB and FANUY, rather than chasing humanoid-themed private-market proxies; enter only after evidence of commercial order conversion or automation-capex guidance acceleration. Thesis is recurring integration/service revenue, with falsification if factory-automation orders weaken for two consecutive reporting periods.
• Use a relative-value expression if humanoid enthusiasm accelerates: long ABB versus short a broad speculative technology basket such as ARKQ, sized beta-neutral. ABB offers exposure to actual factory automation spend, while ARKQ is more vulnerable if pilots fail to become scaled deployments; reassess on a disclosed large multi-site humanoid contract.
• Treat NVDA as an alert, not a humanoid-specific recommendation: require management or customers to quantify incremental robotics compute demand before attributing material upside. A broad industrial robot order cycle would be supportive, but current category volumes are insufficient to change the core AI-infrastructure valuation debate.