Every Robot Needs Bearings

The market cannot yet tell you which robot form factor, OEM, or actuator architecture wins. It can tell you one thing with near certainty: whatever gets built will be full of bearings. Every motor and every joint in a robot requires at least one bearing to reduce friction and support rotating parts. That makes bearings a rare asset in the robotics and physical AI trade. They give exposure to unit growth across the whole sector without a bet on any particular design or builder.

@Morgan Stanley, whose research team projects the humanoid robot market alone reaching roughly $5 trillion by 2050 with over one billion units in operation, forecasts the robot bearings market growing approximately 300x through 2050. Bearings are the unglamorous, lower-multiple expression of the same view: robots are coming, and their skeletons are made of precision components that a handful of incumbent manufacturers already dominate.

Themes in this thesis

This portfolio expresses six connected investment themes along a single value chain:

1. Robotics and humanoid robot components. Precision bearings, strain-wave gears, actuators, and joint modules that go into every robot regardless of form factor. The purest expression of physical AI and embodied AI hardware.
2. Industrial automation and motion control. The makers of precision motion systems, linear guides, ball screws, servo components, and electric motors that power factory automation today and robot fleets tomorrow.
3. Precision manufacturing and engineered components. High-tolerance, aerospace-grade component manufacturers whose micron-level process knowledge is the moat robotics demand will monetize.
4. Specialty steel and advanced materials. The clean alloy steel and specialty metals that precision bearings, gears, and actuators are machined from.
5. Rare earth permanent magnets. Every robot motor needs them; magnet supply is the same architecture-agnostic bet one step upstream.
6. U.S. manufacturing reshoring and industrial distribution. Domestic producers benefiting from supply-chain onshoring, plus the MRO distributors who service the installed base as an aftermarket annuity.

Why bearings, specifically

Content scales with complexity. Bearing count multiplies with a robot's degrees of freedom. A small quadcopter drone uses roughly 8–12 bearings. A humanoid robot uses 70 or more — spread across shoulders, elbows, wrists, fingers, hips, knees, ankles, and neck. As the industry pushes toward more dexterous, more articulated machines, bearing content per unit rises even before unit volumes do. Pricing spans under $1 for miniature deep-groove ball bearings to $100+ for the precision crossed-roller and harmonic-drive bearings used in load-bearing joints, so a mix shift toward high-precision robotics applications also lifts average selling prices.

The demand signal is no longer theoretical. In January 2026, @OpenAI issued a formal RFP for U.S.-based hardware manufacturing capacity and named precision bearings (ball, roller, harmonic) as one of six critical inputs in its robotics category, alongside actuators, harmonic drives, gearboxes and motors, permanent magnets, and power electronics. When the best-capitalized AI company in the world lists your product as a strategic bottleneck it wants reshored, that is a demand signal, not a narrative.

Low obsolescence risk. Bearings face minimal substitution, in-sourcing, or technological displacement risk. There is no credible roadmap for moving machines that eliminates rotating interfaces. Robot OEMs are far more likely to in-source software, actuator assembly, or even motors than they are to vertically integrate into precision bearing metallurgy, a discipline measured in microns and built on a century of accumulated process knowledge.

Market structure favors incumbents

The global bearings industry is consolidated at the high end. The top six global manufacturers — $SKF, $Schaeffler, $NSK, $NTN, $JTEKT, and $Timken — control the majority of the global roller bearing market, with Chinese manufacturers holding roughly 25% of overall supply, concentrated in the low and mid tiers. High-precision segments (the ones robotics requires) remain effectively an oligopoly of Swedish, German, Japanese, and American producers, because precision bearing quality depends on clean steel, heat treatment, and grinding tolerances that are difficult and slow to replicate.

Today's demand mix of roughly 40% industrial equipment OEMs, 30% automotive, and 30% distribution channels means robotics is currently a rounding error in these companies' revenue. That rounding-error status is the appeal. The theme has no weight in their earnings yet, and the base businesses in industrial machinery, automotive, aerospace and defense, rail, and wind energy provide cash flow and dividends while the robotics option matures. The global bearings market generated an estimated $58–100+ billion in 2024 revenue (estimates vary by scope) and is growing mid-to-high single digits before any robotics contribution.

Among the pure bearing makers, $Timken and $RBC Bearings trade on the NYSE and carry a direct reshoring tailwind from the U.S. manufacturing push. RBC Bearings in particular is a high-precision aerospace and defense component specialist and appears in Morgan Stanley's Humanoid 100 basket alongside Timken.

The joint is a bill of materials: industrial automation and motion control

The same demand logic that applies to bearings applies to every precision component packed into a robot joint. A humanoid actuator module pairs crossed-roller bearings with strain-wave gears, ball screws or linear guides, frameless motors, and encoders. The companies that sell across that bill of materials are the industrial automation and motion control suppliers.

Linear motion and precision guides. $THK makes linear guides, ball screws, and crossed-roller bearings, giving it some of the most direct per-joint content exposure of any listed company; it is a Humanoid 100 constituent. $Nippon Thompson (IKO) supplies needle and crossed-roller bearings engineered for compact robot joints. $HIWIN, the Taiwanese linear motion leader, sells guides and ball screws into robot makers globally.

Miniature and specialty bearings. $MinebeaMitsumi is the world leader in miniature ball bearings, the exact category humanoid hands, fingers, and small high-speed motors consume in volume. $Nachi-Fujikoshi makes precision bearings and also builds its own industrial robots, giving it exposure on both sides of the trade.

Strain-wave gears and precision drives. $Harmonic Drive Systems makes the strain-wave gears named explicitly in the OpenAI RFP; its reducers sit inside the joint actuators of competing robot OEMs, making it one of the most closely watched pure-play component names in Japan. $Regal Rexnord owns motion-control and power-transmission portfolios spanning bearings, gearing, and motors, and is also a Humanoid 100 constituent.

Electric motors and motion systems. $Nidec is the dominant global supplier of precision electric motors, each of which needs bearings and many of which will end up inside robots. $AMETEK supplies precision motion, electromechanical devices, and instruments into factory automation and aerospace. $Allient, a U.S. small cap formerly known as Allied Motion, sells integrated precision motion systems including motors and gearing.

Upstream: specialty steel, advanced materials, and rare earth magnets

Precision bearings begin as clean, high-carbon chromium steel, a specialty product with few qualified suppliers. $Metallus, the former TimkenSteel, is the leading U.S. producer of the clean alloy steel that bearing races and rollers are machined from. $Carpenter Technology supplies the specialty alloys and powder metals used in high-performance bearings, gears, and actuator components for aerospace, defense, and medical markets.

The same OpenAI RFP that named precision bearings also named permanent magnets as a robotics critical input. Every motor in a robot needs rare earth magnets, and $MP Materials is the primary U.S. producer of the neodymium-praseodymium feedstock and, increasingly, the finished magnets themselves. The rare earth magnet angle is the same architecture-agnostic logic one step upstream: more robots means more motors means more magnets, regardless of whose robot wins.

Reshoring and distribution: the installed-base annuity

The U.S. manufacturing reshoring push, formalized in commitments like the OpenAI RFP and broader industrial policy, favors domestic precision component producers across every theme above.

Separately, roughly 30% of global bearing demand flows through industrial distribution channels rather than direct OEM sales, and distribution is where the aftermarket annuity lives. Bearings are wear components. Every robot deployed becomes a recurring MRO replacement customer for its service life. $Applied Industrial Technologies is the largest U.S. distributor of bearings and power-transmission components. $Genuine Parts owns Motion Industries, the other dominant North American industrial distributor. Both monetize robot fleets in operation rather than robot units produced, which makes them the longest-duration expression of the theme.

Bottom line

This is a picks-and-shovels position with an unusually clean logical core: bearing demand is a function of total robot joints built globally, regardless of who builds them. The full expression runs the length of the value chain, from the specialty steel and rare earth magnets the components are made of, through the bearing, motion control, and industrial automation suppliers themselves, to the distributors who service the installed base. The incumbents are profitable, dividend-paying industrial companies where the robotics option is cheap or free at current multiples. The cost of that optionality is patience. The payoff window is measured in years, and the stocks will trade like industrials until it arrives.