We roll.

Rolling Interfaces for any motion.

• High efficiency and low wear.
• Motions not achievable from legacy mechanisms.
• Top 10 CO₂ mitigation potential.
• > $10 Billion market.

Roll's custom numerical methods optimize the raceway curvature to achieve pure rolling through the entire cycle. Optimized against un-optimized, each shown at the start and midpoint of the motion. See the animation.

Complex systems.

A kinematic first: multiple rolling interfaces, each precisely following its own motion specification, combine into systems that legacy mechanisms cannot achieve.

A novel retainer system.

The rollers carry the load; a tongue-and-groove retainer keeps them precisely spaced and aligned as they travel the raceway. Shown below with and without it — so you can see exactly what the retainer adds.

The speed reducer.

A target of ~99% efficiency against roughly 60–80% for the harmonic drives it can replace. Even ~97% validated would be a major improvement — and pure rolling means less wear and longer service life.

Where the
energy goes.

Between a motor and its load sits a speed reducer. For millennia the answer has been gears — teeth that mesh and slide against one another. That sliding is a built-in energy loss: it sheds power as friction and heat and wears the parts out. Multiplied across millions of machines, the loss is enormous — and every watt of it traces back to generated power and its emissions.

Carbon impact on the scale of LEDs and carbon capture.

If adopted globally, the technology could mitigate an estimated 1.2–3.0 gigatons of CO₂ a year — a modeled figure from a five-system cross-validation, with central estimates clustering around 1.5–2.5 Gt under moderate adoption.

A modeled estimate, not a measured result. Efficiency figures are targets; incumbent baselines are confirmed from literature; physical validation is the next phase.

Anything that spins is a candidate — pumps, robots, EVs, compressors, engines. Prove the rolling interface once, then apply it across a growing family of products.

The missing building block.

Every machine is built from the same handful of parts — gears, bearings, shafts, linkages — all more than 500 years old. Where they fall short, no amount of engineering has ever fixed it: the crankcase still can't trace an efficient thermodynamic cycle after 150 years of failed attempts to replace it. It was never the engineers — it was the toolkit. Rolling interfaces add a new part to that box: precise, specified motion carried through pure rolling contact. Drop it where the old parts fall short and the results may astound.

One mechanism, two ambitions — built by one engineer.

Roll Labs (formerly 21Geo) invented rolling interfaces — a genuinely new type of mechanism, and arguably the first new mechanical building block in centuries. It carries two ambitions at once: replacing the meshing and sliding inside legacy motion parts with pure rolling, for far higher efficiency and longer life; and unlocking precise motions the old building blocks can't produce. The torque de-rippler is our first entry into that “any motion” frontier, with biomechanics and thermodynamic devices to follow.

The work is founder Tom Dobroth's. An MIT mechanical engineer (SB '83, SM '85), he once cracked a major industrial problem by inventing a numerical method for estimating free-surface fluid-flow shapes — and today he applies that same approach to computing the pure-rolling raceway geometries at the heart of Roll Labs. The vision has always been larger than one person; the goal now is to carry it from validated models to physical proof. Two granted U.S. patents protect the work, with a third pending.