BLOGS

In brief: On September 2, 2026, Schaeffler announced formed strain wave gearboxes built specifically for humanoid robots — with a manufacturing claim that matters more than any torque spec: key components formed in seconds instead of minutes, cutting manufacturing cost by more than 25% and material consumption by more than 75%. Mass production starts in 2027. Here is what actually changed, why it matters for anyone buying robot joint modules, and where the engineering caveats sit.

Why the Reducer Became the Bottleneck

Start with the cost structure. Actuators — motor, reducer, encoder and drive packaged as a joint — are widely estimated to account for around half of a humanoid robot's total manufacturing cost. Within the actuator, the strain wave (harmonic) gearbox is the precision-critical core: it converts a fast, small motor into a slow, strong joint with near-zero backlash.

The problem Schaeffler is attacking is structural. Conventional strain wave gears are produced by precision machining — material removal to create the gear geometry. It works, but it is time-consuming and capital-intensive, and as several tier-one suppliers have warned, it could become a genuine bottleneck if humanoid volumes scale from thousands toward hundreds of thousands of units per year. You cannot buy your way out of a process that takes minutes per part when your customers want millions of joints.

Why Forming Changes the Economics

Forming is not new to Schaeffler — it is transplanted expertise. The company has produced formed strain wave gearboxes for the automotive sector for over a decade and says it has shipped more than 2 million units to all major markets in the last ten years. The humanoid program transfers that automotive-grade forming know-how to robot-specific gear geometries.

Three things make the process work:

• Speed. High pressing forces bring the component to its final geometry in seconds, rather than removing material for minutes. That is the difference between a machine shop and a mass-production line.
• Material efficiency. Machining cuts away a large fraction of the input metal. Forming near-net-shape cuts material consumption by more than 75% — a direct cost line and a supply-chain resilience win.
• Process stability. According to Schaeffler, formed components achieve comparable torque and efficiency to conventionally manufactured ones, with high dimensional accuracy and improved process consistency.

David Kehr, President of Humanoid Robotics at Schaeffler, framed the strategy plainly: the company deploys humanoids across its own global value chain, giving it first-hand requirements knowledge, and combines that with decades of automotive manufacturing experience to make key components "geared toward quality, economy, and scalability."

The Engineering Caveats Buyers Should Keep in Mind

A 25% cost reduction is a headline; a joint module spec sheet is where careers are made and lost. When evaluating formed strain wave gears — from Schaeffler or anyone else — the questions that matter do not change:

• Flexspline fatigue. The flexspline is the wear-critical element in any strain wave gear; it transmits torque through elastic deformation. A new forming route must demonstrate equal or better fatigue life over millions of cycles, not just comparable static torque.
• Batch consistency. Formed gears live or die on tooling wear and process control. Ask for backlash and lost-motion distributions across production batches, not just the single prototype's numbers.
• Stiffness and hysteresis. Torque transmission stiffness and torsional hysteresis determine control bandwidth. "Comparable torque and efficiency" is necessary but not sufficient.
• Thermal behavior. Joint modules are sealed, thermally constrained spaces. A reducer's efficiency is only as good as the heat path around it.

Engineering note: These are solvable engineering problems, and Schaeffler's completed validation testing is a serious signal. But validation under the manufacturer's own load spectrum is not the same as validation under yours.

What It Means for Joint Module Buyers

For teams specifying humanoid joint modules today, the announcement reshapes the purchasing math in three ways:

1. Second-source leverage is arriving. A tier-one automotive supplier entering volume production in 2027 gives humanoid makers a credible alternative to incumbent precision-machining capacity. Procurement teams should qualify both now, before demand forces the issue.
2. The cost curve bends. If actuators are roughly half of BOM cost and the reducer is a major actuator line item, a >25% reducer cost reduction compounds into real per-robot savings at scale — the kind that moves total cost of ownership across the payback threshold that factory operators care about (commonly 1.5–2 years).
3. Volume risk shifts from process to allocation. Once forming capacity exists, the question becomes who gets allocated output when demand spikes. Early design-in and long-term agreements will matter more than spot purchases.

Our own view at Ti5ROBOT is straightforward: joint module buyers should evaluate the transmission route — harmonic, planetary, cycloidal — per joint and per load case first, and only then optimize supplier cost. A lightweight joint module with the right ratio and thermal design beats a cheap reducer in the wrong joint every time. That is why our Lightweight Joint Module series pairs ultra-thin harmonic reducers with high-torque-density frameless motors and dual absolute encoders, and why our humanoid robot actuator lineup is built for integration rather than spec-sheet heroics.

The Bottom Line

Schaeffler's formed strain wave gearbox is not a new gear — it is a new way to afford the same gear. If the 2027 ramp delivers on the stated cost, material and cycle-time numbers, the constraint on humanoid scaling shifts further away from precision manufacturing and toward integration quality: encoders, force sensing, thermal design and control software. That is a good trade for the industry, and it is exactly where joint module suppliers differentiate. We compared the harmonic and planetary routes for joint selection in a companion article — the forming revolution changes the price tags, not the physics.