In the long run, humanoid robots could emerge as a significant new source of magnet rare earth demand, adding pressure to a market already expected to move into deficit as EV and wind-turbine deployment accelerates. A constrained supply base and intensifying geopolitical competition may further increase supply risk. To build resilience, OEMs should reduce material intensity, secure offtake and recycled supply.
The first insight in CRU’s AI-Tech series sets out a central theme: AI will reshape commodity demand through the buildout of physical systems. That buildout has two channels. The first is digital AI infrastructure, including data centres, electricity networks and supporting equipment. The second is embodied AI, where intelligence is embedded in machines.
This second channel is where rare earths and magnets become especially important. Humanoid robots represent one of the clearest examples of AI moving into physical deployment at scale, and their growth would create new demand for the motors and magnetic materials that enable movement and control. In this piece, we explore how that shift could materially change the outlook for magnet rare earths, potentially deepening supply deficits and elevating the strategic importance of these materials in the wider AI race.
By 2040, CRU forecasts humanoid robot demand to be 151 M units – equivalent to 363 kt NdFeB magnets or 157% of the current NdFeB magnet demand. Regardless of demand from humanoid robots, magnet rare earths are forecast to be in deficit over the long run due to accelerated deployment of wind turbines and EVs.
As AI enters its next phase, we expect it to move from digital form (Claude, ChatGPT, Claude, etc.) to physical (i.e. humanoid robots), new resource battlegrounds, one of which which is NdFeB magnet and associated rare earths, will emerge.
Increasingly, by imposing trade restrictions, China is also using rare earths as a tool to navigate geopolitical tensions. Over the long term, growing demand for humanoid robots is likely to contribute to heightened restrictions.
Magnet rare earth demand from humanoid robots to be 449 kt under a high case scenario
Humanoid robots require up to 70 degrees of freedom (DOF) in their joints to achieve human like mobility. To achieve such DOF, humanoid robots require permanent magnet based motors. These permanent magnet motors are used in joints. While for lower powered joints (such as fingers) non magnet-based motors can be used, they are not usually preferred.
CRU estimates that a humanoid robot on average will require ~2-3 kg of NdFeB magnets. Based on CRU’s initial forecast for uptake of humanoid demand, we have developed three scenarios:
- REE demand for humanoid robots – low: Based on low case uptake scenario of humanoid robots, demand of magnets from humanoid robots will be 271kt NdFeB (~120% of 2026 NdFeB demand).
- REE demand from humanoid robots – base: Based on base case uptake scenario of humanoid robots, demand of magnets from humanoid robots will be 363kt NdFeB (~160% of 2026 NdFeB demand).
- REE demand from humanoid robots – high: Based on high case uptake scenario of humanoid robots, demand of magnets from humanoid robots will be 449kt NdFeB (~200% of 2026 NdFeB demand).
Significant demand from humanoid robots is unlikely to materialise before 2035, emerging only as the technology matures. Even in a scenario with no humanoid robot uptake, the rare earth magnet market is forecast to fall into deficit. With the additional surge in rare earth demand from 2035 onwards, the market is set to expand dramatically. Such an increase in demand for magnet rare earths will create further pressure on a market which is already poised to be in a deficit.
Magnet rare earth deficit to be ~160% of the total magnet supply base case under humanoid robot adoption
Between 2020–2025, CRU estimates demand for magnet rare earths grew by 66%. By 2040, even under a no humanoid robot adoption scenario, demand for magnet rare earths will be 89% higher. The market for magnet rare earths has surged in recent years and will grow rapidly regardless of deployment of humanoid robots, driven by accelerated roll out of EVs and wind turbines. This is increasingly placing pressure on the magnet rare earth market.
Across different humanoid adoption scenarios, the rare earth market is headed for a deficit. Humanoid robots further the scale of such a deficit, magnifying impact across all industries and major OEMs.
- Magnet REE supply deficit – no humanoid robot adoption: By 2040, the deficit will only be 41% of the total supply.
- Magnet REE supply deficit, including humanoid robots – low adoption: By 2040, the deficit rises to 126% of total supply, 86 percentage points higher than in the no-humanoid-robot-adoption scenario.
- Magnet REE supply deficit including humanoid robots – base adoption: By 2040, the deficit rises to 156% of total supply, 115 percentage points higher than in the no-humanoid-robot-adoption scenario.
- Magnet REE supply deficit, including humanoid robots – high adoption: By 2040, the deficit rises to 183% of total supply, 142 percentage points higher than in the no-humanoid-robot-adoption scenario.
Given the strategic importance of humanoid robots as a next generation AI technology, their widespread adoption is likely to intensify geopolitical tensions surrounding rare earths. At the same time, forecast deficits in magnet rare earths market, driven by both energy transition and emergence of humanoid robots, could also lead to demand destruction. Together, these factors place the rare earths market in an increasingly delicate and vulnerable position.
- Further geopolitical strain on rare earths: Artificial Intelligence related technologies are highly politicised. For instance, the US in past has blocked access to strategically important technologies, such as Anthropic’s Fable 5, to non-US citizens. Rare earths are similarly highly politicised, with China using trade restrictions as a tool to navigate uncertainty in its relationship with the US. As humanoid robots emerge as a strategically important next-generation artificial intelligence technology, counties are likely to intensify efforts to secure competitive advantage. This could also leave countries without stockpiles or domestic rare earth production routes in a particularly vulnerable position.
- Demand destruction:If growth in supply fails to keep up with demand, companies may accelerate thrifting and substitution efforts, reducing overall rare earth demand. If substitution efforts fail to materialise, demand for rare earth intensive products such as humanoid robots and EVs could also weaken given OEMs will then struggle to secure sufficient material, creating a bottleneck across the market.
Considering the criticality of rare earths, OEMs need to hedge against this risk
Demand for magnet rare earths is already rising rapidly because of energy transition, and will likely increase further with the emergence of physical AI. As the US and China compete for leadership in the AI race, humanoid robots represent a new strategic frontier. There is already precedent for governments restricting access to strategically important AI-related technologies. Given the importance of rare earth elements in humanoid robot manufacturing, it may become increasingly difficult to gain access to these metals. Example of past restrictions on AI-related technologies include:
- Nvidia chips:Citing national and economic security concerns, the US has previously restricted sales of its most advanced Nvidia chips to China.
- Claude Fable 5 and Claude Mythos 5:The US government has also previously restricted access to Anthropic’s flagship AI models for non-US citizens, whether inside or outside the United States.
Against this backdrop, and with the rare earths market already forecast to move into deficit, OEMs will need to strengthen their strategies to protect themselves against potential supply constraints in rare earths and other critical materials. Broadly, there are two main ways OEMs can respond:
- Thrifting:Develop products and manufacturing processes that reduce rare earth intensity or eliminate it altogether. This could include changing motor types, or redesigning components to reduce dependence on magnet rare earths.
- Offtake agreements:Secure long-term supply through offtake agreements with emerging mining companies or recyclers. This can help OEMs reduce exposure to supply shocks, improve visibility over future material availability, and diversify sourcing away from highly concentrated supply chains.
Over the next decade, critical metals markets are likely to come under even greater pressure. OEMs and governments alike will need to take proactive steps to manage this risk. This may include securing supply through recycling, investing upstream, pursuing strategic partnerships or advancing thrifting strategies. CRU can help you navigate these changes and uncertainties across critical materials markets. To learn more, contact us.
