TechFEATURE

Humanoid Robots: Trillion-Dollar Bank Forecasts Meet Factory-Floor Reality

林紀旭 James LinEditor-in-Chief
Published · Updated
Tesla's Optimus has scaled its 2025 production target from 5,000 to roughly 2,000 units and paused assembly in October, while independent estimates put per-unit manufacturing cost at $50,000–$100,000 against a $20,000–$30,000 goal. Meanwhile Figure AI's valuation rose from $2.6 billion to $39 billion in about 18 months, and Agility Robotics' Digit has moved over 100,000 totes at a GXO warehouse. Goldman Sachs projects a $38 billion market by 2035 versus roughly 20,000 units shipped in 2025.

Why does Tesla's Optimus keep slipping on production timelines?

Tesla has repeatedly cut back its own Optimus production targets and paused assembly rather than ramping output as plannedCITE:E1. Elon Musk originally targeted at least 5,000 units built by the end of 2025, a figure later revised down to roughly 2,000CITE:E1. According to TrendForce, Tesla reportedly halted assembly in October 2025 after design problems with the robot's humanlike hands and forearms proved difficult to manufacture at scale, leaving a stockpile of hand-less half-built units with no stated delivery dateCITE:E1.

Why does the cost gap block mass-market humanoids?

Tesla's stated long-term ambition is a $20,000–$30,000 retail price and annual output in the millions, but that figure is Musk's spoken target rather than an official price, and independent estimates put current per-unit manufacturing cost at $50,000 to $100,000CITE:E2. That gap sits against a starkly different cost structure in China: Unitree's G1 launched in 2024 at $16,000CITE:E5, and its officially sold price has since dropped to about $13,500CITE:E5. Unitree's structural advantage traces to actuator, battery, and assembly supply chains that overlap with the electric-vehicle industryCITE:E5.

MetricTesla OptimusUnitree G1
Stated target price$20,000–$30,000$16,000 (2024 launch)
Current price/cost$50,000–$100,000 (estimated manufacturing cost)~$13,500 (2025)

What does Figure AI's valuation surge signal about capital flows?

Figure AI's valuation rose roughly 15-fold in about 18 months as investors bet heavily on humanoid roboticsCITE:E3. The company closed a Series C in September 2025 that raised more than $1 billion, lifting its post-money valuation to $39 billion, up from $2.6 billion at its Series B in February 2024CITE:E3. Participants in the round included NVIDIA, Brookfield, and Intel CapitalCITE:E3.

Where are humanoid robots already working in real factories and warehouses?

Two humanoid robots have moved from demonstration to sustained, paid deployment inside specific work stationsCITE:E4CITE:E6. Figure's 02 unit ran a pilot at BMW's Spartanburg, South Carolina plant, where over roughly 10–11 months in 2025 it helped produce more than 30,000 BMW X3 vehiclesCITE:E4. BMW and Figure followed up in October 2025 with the third-generation Figure 03, which runs Figure's self-developed Helix vision-language-action model and has taken on logistics picking tasksCITE:E4. Separately, Agility Robotics' Digit has moved more than 100,000 totes at a GXO logistics center in Flowery Branch as of November 2025, marking the first humanoid robot in paid, commercial robots-as-a-service deploymentCITE:E6. Both deployments remain confined to structured tasks — vehicle-assembly support and tote handling — rather than general-purpose workCITE:E4CITE:E6.

How do bank forecasts compare with today's pilot scale?

Goldman Sachs and Morgan Stanley both project markets far larger than any current deployment, while acknowledging these are forecasts rather than realized salesCITE:E7. Goldman Sachs estimates the global humanoid robot market could reach $38 billion by 2035 with annual shipments of about 1.4 million units, compared with an estimated 20,000 units shipped in 2025CITE:E7. Morgan Stanley's longer-range estimate puts humanoid robot revenue at roughly $5 trillion by 2050CITE:E7.

ForecastValueTimeframe
Estimated 2025 shipments~20,000 units2025
Goldman Sachs market size forecast$38 billion / ~1.4 million units/year2035
Morgan Stanley revenue forecast~$5 trillion2050

What this means: Tesla's stalled assembly line and $50,000–$100,000 manufacturing cost sit alongside Figure AI's $39 billion valuation and Unitree's sub-$14,000 pricing, showing capital markets and cost realities are moving on different timelines. The confirmed commercial activity to date — 30,000 BMW X3s assisted at one plant and 100,000 totes moved at one warehouse — is orders of magnitude smaller than the 1.4-million-unit annual shipment figure Goldman Sachs projects for 2035.

📊 Evidence

FAQ

Why does Tesla's Optimus keep slipping on production timelines?

Tesla has repeatedly cut back its own Optimus production targets and paused assembly rather than ramping output as plannedCITE:E1.

Why does the cost gap block mass-market humanoids?

Tesla's stated long-term ambition is a $20,000–$30,000 retail price and annual output in the millions, but that figure is Musk's spoken target rather than an of…

What does Figure AI's valuation surge signal about capital flows?

Figure AI's valuation rose roughly 15-fold in about 18 months as investors bet heavily on humanoid roboticsCITE:E3.

Where are humanoid robots already working in real factories and warehouses?

Two humanoid robots have moved from demonstration to sustained, paid deployment inside specific work stationsCITE:E4CITE:E6.

📎 Sources

  1. trendforce.com
  2. roboticscenter.ai
  3. figure.ai
  4. press.bmwgroup.com
  5. robotics247.com
  6. agilityrobotics.com
  7. goldmansachs.com

Related data

Author's Take林紀旭 James Lin

The numbers in this story pull in opposite directions at the same moment: Tesla's own manufacturing cost estimate of $50,000–$100,000 sits nearly four times above Unitree's $13,500 retail price, yet Figure AI's valuation still climbed to $39 billion on the strength of a 30,000-unit BMW pilot and a 100,000-tote GXO deployment — real but narrow, single-task wins. That mismatch suggests capital is pricing in the 1.4-million-unit 2035 scenario Goldman Sachs describes, not the roughly 20,000 units actually shipped in 2025. The metric worth watching next is whether Tesla resumes Optimus assembly with a hand design that ships at volume, since a stalled assembly line at the highest-profile program is a direct constraint on how fast the pilot-to-mass-production gap actually closes.

林紀旭 James LinEditor-in-Chief

Related

FEATURE

AI Inference Prices Are Falling About 10x a Year — What the Token-Cost Data Actually Shows

Equivalent-capability AI inference prices have fallen roughly 10x per year across multiple benchmarks, with OpenAI's GPT-4o mini priced about 99% below its 2022 predecessor and DeepSeek-R1 undercutting OpenAI's o1 by roughly 27x. Yet Google now processes about 50 times more tokens monthly than a year earlier, and Gartner projects 2025 global generative-AI spending at $644 billion, up 76.4% — proof that cheaper tokens are fueling more total usage and spend, not less.

Nathan ·
FEATURE

China's AI Playbook Under Export Controls: Ahead on Models, a Generation Behind on Chips

China's AI industry shows a split scorecard: on models, DeepSeek-V3's open-weight, MIT-licensed mixture-of-experts architecture and DeepSeek-R1's benchmark parity with OpenAI's o1 preceded a $593 billion single-day NVIDIA sell-off, while the widely cited $5.576 million cost figure covers only the final training run. On hardware, Huawei's Ascend 910C still delivers about 60% of NVIDIA H100's per-chip performance, and HBM supply constraints cap actual 2025 shipments far below estimated capacity, even as US export controls kept tightening through January 2026.

Nathan ·
FEATURE

Why Are Tech Giants Racing to Buy Nuclear Power? Can SMRs Solve AI Data Centers' Electricity Crisis?

Facing a projected jump in global data center electricity demand from about 415 TWh in 2024 to about 945 TWh in 2030, Microsoft, Google, Amazon, and Meta have all signed nuclear power deals — but most rely on small modular reactors (SMRs) that won't deliver meaningful capacity until 2030 or later, and NuScale's 2023 project cancellation shows the economics are still unproven.

Nathan ·