What if you wanted to completely rethink a bicycle drivetrain to have less drag, less weight, and take up less space? That’s what CeramicSpeed has done with their new Driven drivetrain concept, relying on a pair of roller pinions on a carbon driveshaft mated to a completely flat cassette.

CeramicSpeed’s tests show that a typical Dura-Ace drivetrain’s peak efficiency is around 98%, and a CeramicSpeed optimized version of that is 98.5%. We truly are talking about incremental gains. The new Driven system achieved 99% efficiency in the same test. When you start adding in things like chain lube degradation and cross-chain friction, things get worse for the traditional drivetrain. And those peak numbers? That’s at an average power output around 500 watts. If you’re not putting out that kind of power, the numbers get a little worse. At 200 watts, DA is around 96.75% and the optimized version is 97.75%. The Driven system is up at 98.75%. Putting the numbers into another format, they say at 250 watts, Driven has 32% less friction than the CeramicSpeed optimized DA drivetrain, and 49% less friction than a stock Dura-Ace group.

ceramicspeed concept drivetrain with shaft drive and flat cassette

If the drag numbers don’t impress you, there are other benefits, too. It’s potentially far more aerodynamic, and it’s definitely lighter. The power meter can be integrated into the drive shaft, and the gear combinations are nearly endless. The cassette can add as many speeds as they want (within reason, of course) because it doesn’t need additional width to add more speeds.

ceramicspeed concept drivetrain with shaft drive and flat cassette

The roller pinion slides fore and aft to shift gears.

CeramicSpeed Driven concept shaft drivetrain with roller bearings and flat cassette driven by a carbon shaft

The fixed front pinion can be designed to move, opening up the opportunity for a double gear set up front, too. This concept is limited to a single front cog, though.

CeramicSpeed Driven concept shaft drivetrain with roller bearings and flat cassette driven by a carbon shaft

CeramicSpeed Driven concept shaft drivetrain with roller bearings and flat cassette requires no derailleur

This one’s left open to show how it works, but it can easily be placed inside a cover, or possibly even inside a chainstay, to protect it from the elements. As is, it requires an elevated chainstay, and therefore a custom frame. Should it make it into production, frame builders would be required to build the bike specifically for this. Don’t hold your breath…the concept has only been in development for nine months and still has a few bugs to workout.

ceramicspeed concept drivetrain with shaft drive and flat cassette

CeramicSpeed Driven concept shaft drivetrain with roller bearings and flat cassette requires no derailleur

There’s also the opportunity for using different materials to save more weight, and angling the teeth and/or bearings to further optimize its motion. Even double-row bearings could be implemented.

ceramicspeed concept drivetrain with shaft drive and flat cassette

For now, it’s purely a concept, but this is how great ideas start. Well done, CeramicSpeed!



  1. So the entire power of the rider is delivered onto just 2 or 3 engaged teeth of the chainring and cassette (or whatever they’re calling that thing)? Surely that’s just going to strip the teeth or wear extremely quickly?

    • An easy solution to deal with the front chainring is to add a 2nd set of teeth closer to the crank spindle and then simply add an additional “bevel gear” or whatever you want to call it timed appropriately.

      This design has issues from a mechanical standpoint but I don’t think they’re insurmountable. I can already envision how to fix their shifting problem without using very fancy electronics. That said, the biggest issue they will face is the need for a proprietary frame.

    • Notice that the engagement point is a set of bearings. The drive doesn’t slide in and off from the engaged teeth like a traditional bike chain.

      • Chains don’t slide on & off from an engaged tooth either….
        They have rollers for the express purpose of rolling on & off the tooth…

    • Not it if never gets produced. This is destined to be a wall hanger. Shaft drive has been done to death, and it’s just not effective for bikes. I already thought this company was BS, didn’t think they could go farther, but somehow they’ve outdone themselves. A sucker is born every minute.

      • So even though it’s a prototype youre convinced it’s not going to work… interesting. It’s interesting to see critics who have negative comments about profitable, successful companies who work their asses off and take all the risks. This is the type of fuel companies use to prove their critics wrong all the time… it’s a concept which drives technological innovation at their company. It may never come to light but they may take what they learn and apply to other areas.

        • maus haus – why do you think you have to invalidate the comment from Shafty. a good skeptic is a good engineer. a lot of corporations rely on the ‘gee whiz’ factor of a new design, and fail to illuminate the fact that most designs are derivative and rarely is anything truly ‘new’.

          cycling is absolutely plagued by ill-informed and deceitful companies that rely on the general ignorance of their customers. that’s why you find so many meaningless acronyms and questionable advertising.

          the fact that the main point of this driveshaft design (that is anything but new) is promising AT MOST 1-2% increases in efficiency. obviously, this small change is well within the margin of error of any possible measuring instrument used to determine that efficiency. in that regard, most people will not notice, or benefit from such a small gain in efficiency.

          so, in the nicest way possible, i have come to the same conclusion. this driveshaft design is unlikely to reach the market, simply based on cost alone. the increase in performance is completely negligible.

          the average cycling enthusiast should be able to realize the purpose of price vs. performance. personally, i won’t pay beyond $1 per gram to save weight.

          in this case, any true enthusiast would benefit from simple improvements on efficiency by optimizing their tires, and wheels(reducing weight, friction, rotational inertia, tire pressures, etc) and working on their aero position ( which is free). Once the bike and rider is optimized completely, if the rider is still desperate for more performance, then an esoteric frame and drivetrain should be considered, but people must realize, that all the low hanging fruit have priority first.

      • Typical “crank” response… Fortunately, “progressive thinking” fellows like yourself are the exception and not the rule… did you finally sell your Buggy Whips shares since that car thing has finally been validated…;-)?

    • The bevel gears a car’s differential also have only 2-3 teeth engaging at a time, and they handle slightly more torque than a rider can put out.

    • The small ball bearings and teeth are transferring far less torque than the rider is putting out, so they do not need to be extremely bulky. The crank ring greatly increases the rotational speed of the driveshaft from what the rider is pedaling at (probably about 1-to-5) which means it’s operating at 1/5th the torque output.

      If a rider is outputting 500W of power @ 60 rpm (sort of slow cadence), the input torque to the crank is about 60 ft-lbs. The driveshaft would then be transferring about 12 ft-lbs of torque, which is well below what those bearings and drive teeth should be rated for.

      To calculate power: P (W) = T(N*m) * W (rads/sec)
      Power remains constant through the drivetrain, so an increase in rotational speed would require an equal decrease in torque.

      In case you still don’t believe me: https://www.wolframalpha.com/input/?i=500W+@+300rpm

  2. I’d love more info on the testing used to determine efficiency. I know – as long as the same test was used, relative comparisons are valid, regardless the numbers, but I’d still like to know more about the testing.

  3. it looks like if you put 500w at 0kph the thing would instantly rip itself off. also doubt the numbers a whole lot.

    would love to know more

  4. Ha…and the UCI just legalized disc rotors…let’s put a huge midieval, torturing spiky disc on bikes now. Yet it looks super cool though!

  5. Solution in search of a problem. I applaud them for thinking outside the box but there is probably no replacing the derailleur / sprocket drivetrain for sheer engineering simplicity.

    • and if man were supposed to fly he’d have been born with wings..right?

      Simplicity??? over 100 parts in the chain (weakest link & all), a derailleur hanging in the air vulnerable to impact, a design almost 100 years old & you believe it is the final/best solution??

      How is that horse working out for you?

    • It solves a couple of problems: it eliminates the constraint of having to make the freehub wider to add additional gears, eliminates the issue of cog spacing on high range cassettes, and eliminates the problems of bad chainline on high range cassettes.

      Wondering how it could be adapted to FS though, you’d need some mechanism to grow/shrink the length or a jockey pulley type mechanism.

  6. Do not believe the numbers. They go against everything that is known about power transfer involving plane changes. A minimum 5% effciency loss at each end.

    • Could you provide some lecture ? There is mostly no sliding in this design because there are rollers on ball on every sliding point. I believe it could work. As said before i think it needs to be seriously built to whitstand some use because all the power goes on those tiny ball bearing.

  7. Whoa! Very cool. It’s so hard to come up with truly new ideas for bikes though that I half expect someone to dig out a patent drawing from 1893 that shows something similar.

  8. That’s definitively crazy cool, and better than their overpriced rollers… this allows better hub geometry, could integrate it to the frame for aero, No dégradation with dust on the chain, quicker wheel change, probably better in mud
    Now definitively few challenges power transmission with angles use to have lower efficiency (I got there are bearings) and involve lateral forces. With actual design no way the “cassette” won’t flex under low gear / high intensity acceleration… but still Cool!

  9. Nope, stick to selling snake oil boys. Let’s see…

    So there’s the inconsistent torque as the teeth engage on the roller bearings insofar as the distance from the centre of the drive shaft to the tooth engagement point won’t be constant.

    Next I bet there’s a constant clatter as one tooth hands off to another with either a dead point as one let’s go before the next picks up OR as the effective drive from one tooth sellers u (due to the afore mentioned torque inconsistency) whilst the next is decellerating into its radius or closest approach.

    Next up is the block full of neutrals between gears because unlike a conventional cassette where the chain can be engaged on two sprockets during the gear change, that would sieze this solid.

    Please note, for anyone who wants to call me an ‘armchair engineer’ I’ve got two degrees in engineering and it’s been my career for a quarter of a century.

  10. Hey Bike Rumor, when posting some of these unusual designs can you please include some video or draw some arrows on the pictures to make clear how it works? I *think* I understand how it works but there are far too many times I am not sure from reading BR articles. I am not asking for high production quality; a quick video of the product moving without any narration would still be a big help.

    • I found a you video on how shaft drives work: https://youtu.be/u-P3xPdz35s

      This uses those ideas, adds the cartridge bearings at each shaft end, and a shifting mechanism.

      The shift mechanism basically telescopes the end of the shaft to line it up with concentric gears.

      That last one is what I don’t love about this – the minimum gear change is roughly the width of those bearings, and the distance between them changes how much the shaft end must move. Unless indexing is managed by the cassette, it’d be some complex mech to handle more than a single given gear range.

    • Thanks for the feedback. We try to make things as clear as we can given the information we have (and can get). If you have a specific question, leave it here and we’ll see what we can do. Thanks!

      • Hi Tyler. My comment was general to many BR articles. Many could use some better explanation.

        For this specific article: how does this thing actually shift? It seems from the comments that it doesn’t, which is a pretty big deal and should have been mentioned in the article.

  11. i think its suspicious that theres no videos of this thing shifting. also, the “cassette” needs to be EXTREMELY STIFF especially for the larger cogs. any deflection and the thing will skip.

  12. Looks interesting for sure. In terms of removing derailleurs though, I think there is more advancement coming in the future from belt drives and internal geared rear hubs.

  13. Does this company sell a single product which lives up to their claims? Even the name of the company is bogus marketing hype.

  14. UCI has pre-emptively banned it because a roadie has already complained about the possibility of falling off his bike, bouncing off a curb, hitting a tree, landing on his head…then having his pinky pricked by those spikes on the cassette.

  15. As with all concepts there will be issues . The biggest issue is the single contact point of each tooth on the bearing . Teeth will need to be much stronger (perhaps ceramic coated) . Other issues like indexing can easily be solved by simply measuring the distance of each gear and implementing it into the shifter gear . The strength of the bearings shouldn’t be a problem as bearings are incredibly strong . I would expect them to be full ceramics for extra hardness and ;longevity . I hope the carbon shaft is also well coated as the constant rubbing from the bearing sleeve will wear down the resin very easily .

    • Single point of contact = definitely a big issue in regards to durability. That said, they could go with a thicker tooth to spread that contact over a larger line.
      Problem with using bearings for drive is that you can’t use a proper gear tooth profile and therefore I don’t think you could have multiple teeth engaged at a time.

  16. “Wild CeramicSpeed Driven 13-speed drivetrain concept shifts with no derailleur”

    Except that it doesn’t shift at all.

  17. I’m at Eurobike. It doesn’t shift. It self destructs if the sprocket wears at all. The frame has to be overbuilt to handle lateral forces. The bearings are unsealed to look cool and keep efficiency up. The bearings make point contact with the sprocket, forcing wear and putting large loads on tiny portions of tiny ball bearings.
    I wish them luck and it’s good to see new (old) ideas. But this was a publicity stunt. Bike roumor should ask more questions before gushing.

  18. Interesting concept, but there’s no way this goes anywhere. The concept alone has so many obvious flaws. Based on videos showing CeramicSpeed discussing how shifting would work, there are only 5 “shift channels.” The Roller bearings would be moving perpendicularly to the direction of wheel rotation during shifting, with the result being that any misalignment would cause teeth on the drive wheel to bend or shear off and bearings to be damaged or destroyed. I.e., the design is inherently less forgiving than current drivetrain designs. Also, this system would require significant changes to frame construction that would be incompatible with existing drivetrains. And for what? As long as current technology is available to all (i.e., everyone is on a level playing field), why does it matter if current drivetrains have a bit of excess drag? We all know that the current, double diamond standard for bicycle frames is not the most efficient, but we aren’t abandoning that design for the most part.

  19. Nice marketing bullsh*t. All the websites are focusing on that instead of real innovations with less carbon and shiny CNC parts….
    it’s not even 20% finished, all the shifting is not solved and that will be the tricky point of this concept.
    Smoke and mirrors!

  20. Why do people get so upset when they see companies play around with prototypes? This is what happens everyday in the real world – trial and error – prototyping – etc. It’s how “things” get developed.

  21. It is in an interesting collection of old and very old designs using new materials. There are a significant number of patents from the 70’s, 80’s, and 90’s on this type of shaft transmission but I suspect the complexities of a functioning design over a chain drive for 0.5% efficiency increases (+2 watts at 400 watts) might be an issue. For a beautiful single speed ‘gravel’ bike with the same drive shaft look here: http://www.oldbike.eu/chainless/?page_id=131

  22. Thats a crazy looking drivechain. The concept is awesome and I hope it does come into production. It really sucks that it needs a custom bike frame though. I’m sure that will run the price up a lot if and when these do come out. I’m new to biking but even small gains are still gains. I’m assuming based off of what I just read because its a 13 speed it has less friction which allows you to ride faster?

  23. As long as they’re thinking out of the box, this makes it one step closer to a single-sided drive-side stay…kind of like a Lefty fork, but a Righty-“chainstay” with the “cassette” doubling as a disc brake rotor.

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.