The Climate Case
Speed reducers live inside essentially everything that spins — industrial motors, pumps, compressors, HVAC, cars, trucks, wind turbines, robots. Each one loses energy the same way: parts sliding against parts. Replace the sliding contact with rolling contact and the saving per machine is a few percent. Broad use is what turns a few percent into a climate number — a single LED bulb saves almost nothing, and broad use made the LED one of the top climate solutions. Same logic.
Adopted globally, the technology could avoid an estimated 1.2–3.0 gigatons of CO₂ a year — among the top climate technologies by potential impact. Five independent analyses found similar results: 1.2 gigatons under conservative adoption, 3.0 under rapid, broad adoption. The figures are modeled, pending hardware validation. We state them anyway, because the size of the number is the point — this is more than a novelty.
The efficiency story has two halves. The first is the speed reducer — see the product page — where removing sliding aims directly at the losses every drivetrain pays today. The second half is stranger and larger: machines whose efficiency is limited not by friction but by the motion their mechanisms can trace.
Elusive efficiency
The mechanical toolkit is astonishingly static. Gears, cams, linkages, and bearings have been the mainstays of machine design for hundreds of years. Using old parts is fine when the outcome is good — but for whole categories of machinery, it isn't. A large share of the world's electricity drives compressors for HVAC and refrigeration; engines discard much of their fuel's energy when the exhaust valve opens; drivetrains tax every vehicle on the road. These machines have barely changed in decades, and all of them are built from the same parts box.
The clearest case is the crank-and-piston. Since the beginning of the internal combustion engine, it's been known that better efficiency is possible. There have been hundreds of attempts using additional linkages, cams, and sliders. Add one thing and subtract another: efficiency at the cost of durability, size, weight — like whack-a-mole. Still today, the crank-and-piston prevails. The engineers were brilliant. The tool was the limit.
Continuous processes need to be optimized with continuous variables.
A linkage adds a discrete bar to a continuous problem — a better approximation, never the optimum. A thermodynamic cycle is a continuous process; what it needs is a mechanism whose motion can be specified continuously.
A continuous tool for a continuous problem
A rolling interface takes the motion specification directly: piston height as a continuous function of shaft angle, delivered through rolling contact. The cycle defines the geometry instead of the crank constraining the cycle. One example of what that unlocks: constant torque during the power stroke, achieved today by adding cylinders — four, six, a V8 — could instead be designed into the motion itself, with the piston's velocity shaped against the force curve. If the mechanism is already a rolling interface, that shaping comes with it.
The compressor and engine case studies work these ideas through concrete design studies. Their results are modeled and their payoffs are stated as possibilities, because that is what they are until hardware runs — the honest phrasing is that the outcomes may astound, and finding out is the point of the work.
Case Studies
Robotic Walking
A leg joint that keeps the body level through the stride — walking about as efficient as rolling.
Learn More →