What is the fastest bike in the world? We tested 12 of the latest all-out aero bikes in the wind tunnel to find out

After half a decade dominated by the 'aero all-rounder', 2025 is the year in which the all-out aero bike truly made a comeback.

Evidence of this is Tadej Pogačar and Jonas Vingegaard racing their respective aero bikes, the Colnago Y1Rs and Cervélo S5, for every road stage of the recent Tour de France, even the high-mountain stages, despite both having a nominal 'lightweight bike' at their disposal too.

It comes thanks in part to the updated UCI rules allowing deeper tube profiles, but also the advancement of carbon fibre, enabling deeper tubes to be made lighter, and thus more aerodynamic machines to be built closer to the UCI's 6.8kg limit.

Most of the best aero bikes nowadays are still reasonably traditional-looking, at least in the context of modern carbon road bikes, but as brands look to eke out every last watt, a few of the more recent designs resemble spaceships on wheels. I'm looking at you, Factor.

But which of them is fastest? As the aero arms race hots up and bike design gets wilder and more creative, who gets the bragging rights? How many watts are being saved? And for those of you looking to upgrade your own bike with one of the latest aero superbikes, which one should you buy?

To answer these questions, in our latest Cyclingnews Labs test, we've been back in the wind tunnel with 12 of the latest dedicated aero bikes on the market. Our selection includes the Colnago Y1Rs, Factor's UCI-legal pre-production prototype, the latest Cervélo S5, Ridley's Noah Fast 3.0, Argon 18's new Nitrogen Pro and many more.

And just like we did in last year's superbike test, we compared everything to my old 2015 Trek Emonda ALR, complete with box-section wheels, rim brakes and exposed cables to allow comparison across both tests.

The bikes

While our previous test centred on popular all-rounders and bikes found in the WorldTour, this time round, our focus was on dedicated aero bikes.

In alphabetical order, they are as follows:

• Argon 18 Nitrogen Pro (pre-production prototype)
• Bianchi Oltre RC
• Cervélo S5
• Colnago Y1Rs
• Cube Litening Aero C:68X
• DARE Velocity Ace-AFO
• Elves Falath EXP
• Factor Prototype
• Merida Reacto 9000
• Ridley Noah Fast 3.0
• Scott Foil RC
• Van Rysel RCR-F
• And our baseline: My very own 2015 Trek Émonda ALR

We requested a top-tier spec from every brand, mainly to avoid any brands gaining an unfair advantage, and because we wanted each of our aero bikes to be 'equal' in spec level to ensure as fair a test as possible.

The difference between the groupsets' aerodynamics may be negligible, but the fairness of the result would be hampered with a 105 vs Dura-Ace test as a result of the remaining components. Top-spec bikes often come with the fastest components (one-piece cockpits, for example), whereas second-tier specs might come with two-piece bar-stem combos or cheaper wheels.

All but four of the bikes were sent with Shimano Dura-Ace. The Cervélo S5 came with SRAM Red AXS XPLR (1x), while the Merida Reacto, Ridley Noah, and Cube Litening Aero C:68X were all sent with Ultegra. Their use of a one-piece cockpit negated the detriment somewhat, but it's worth noting for the results below.

As ever, we are at the liberty of what bike brands are willing to send. We tried to get the BMC Teammachine R, for example, but our request went unanswered; Ribble and Orbea refused our requests for the Ultra Aero SLR and Orca Aero, respectively; the new Merida Reacto that leaked at the Tour de France isn't yet available; and the Cannondale SystemSix is almost entirely out of stock.

If you're wondering why bikes like the S-Works Tarmac SL8, Canyon Aeroad and Trek Madone aren't here, long-time readers will recall that we already tested them in our wind tunnel superbikes test in 2024. That test featured 11 aero all-rounders, including the aforementioned trio, the Pinarello Dogma F, Cannondale SuperSix Evo, Factor Ostro VAM, Giant Propel and others. You can head to that test to see how those bikes compare in isolation, or keep reading to see how their results compare to the aero bikes here.

The test

As with our previous test, we headed to the wind tunnel at the Silverstone Sports Engineering Hub.

This measures an item's Coefficient of Drag x Area, or CdA, which is effectively a metric of how aerodynamic something is.

As might seem obvious, we were trying to find out which of the bikes was the fastest, or more precisely, the most aerodynamically efficient.

We also wanted to know how said aero bikes compare to our baseline, with the Emonda ALR serving as our constant, and as a flagbearer for a typical 'entry level' bike. This year, we're also interested in seeing how the bikes compare to last year's cohort mentioned above. The protocol for a fair comparison of the two tests is outlined below.

We tested each bike three times, using the following setups:

1. Bike only - this offers a highly repeatable and accurate test, but without the real-world 'realism'. Bikes can't pedal themselves, but we think for comparative purposes, it's a test worth doing alongside a rider-on-bike test.
2. With rider - This is a little less repeatable, because humans are notoriously... mobile. As the rider on the bike, I could theoretically spoil the test by simply holding my body in a different position, turning my head halfway through the capture, or so on. It is, however, more realistic, since the interaction between my legs and the bike's frame will affect how aerodynamic a bike is in the real world.
3. Bike only, standardised wheels - Calculating that we'd have time to spare at the end of the day, I decided to take along a set of Enve SES 4.5 wheels and test each bike with a consistent wheelset too (not a sponsored inclusion, we just had them in stock for a review already). This allowed us not only to see the aerodynamic performance of each bike as you can buy it (i.e. with its stock wheels), but also the aerodynamic performance of each frame as a platform for upgrading, and each bike's stock wheelset too.

The protocol

The protocol for the test matched our 2024 test, which was designed by myself with some guidance from the experts at Silverstone Sports Engineering Hub.

This included seven yaw angles: -15°, -10°, -5°, 0°, 5°, 10° and 15°; resulting in a total of 21 aerodynamic data captures per bike (three setups, each at seven yaw angles).

For the uninitiated, yaw is essentially the angle at which the wind hits the rider. Zero degrees is a direct headwind, while a bigger number means the wind is more of a crosswind. A 90-degree yaw would mean the wind is hitting a rider perpendicularly from the left, while minus 45 degrees means it's hitting them from the right (driveside) halfway between head-on and side-on.

Importantly, however, when you ride at 40km/h on a perfectly still day, you're effectively creating your own 40km/h headwind, so the likelihood of ever seeing much above 20 degrees either side is incredibly low, and even then, in short bursts rather than constantly. This is explained a little further below, but hopefully explains why we only tested to 15 degrees either side.

We tested each bike at 40km/h, which we deemed representative of an amateur road race pace, a competitive amateur's road bike time trial pace, and longer days in a breakaway in the professional peloton.

For the bike-only tests, we captured data for 10 seconds per yaw angle and ensured the wheels were spinning at the same speed.

For the rider-on tests, we captured data for 30 seconds per yaw angle. This is a good middle-ground between a long enough exposure to get accurate results that smooths out any small movements, and short enough to avoid the fatigue of holding such a fixed position.

My cadence was held at an average of 90rpm, to ensure 45 full rotations per capture (and not start/finish midway through a complete revolution).

The wind tunnel underwent a tare process before each test, which is similar to a zero-offset on your power meter or the tare on your kitchen scales. The tare data was captured and analysed for any inconsistencies, and repeated if anomalous.

Standardisations

We care about a fair test as much as our readers, so everything that was within our control was standardised.

This includes the bike setup and rider position. Each bike was a 56cm (or the brand's closest equivalent), and we spent the whole previous day adjusting the setup for each bike to ensure they matched as closely as possible.

For the initial 'bike-only' test and the 'bike-with-rider' test, we ran each bike with the wheels it would come with if you bought the bike off the shelf. We standardised with a 25mm Continental GP5000 ST R front tyre, as this is the thing that hits the wind before anything else, and as such, can have an outsized influence.

When doing so, we also used inner tubes of the same valve length. The one exception to this is the Cube: its Newmen wheels come with a hidden valve system, which places the valve inside the rim and uses a special tool to inflate it. Since owners of the Cube bike would be forced to use those tubes / tubeless valves, we tested using them too.

Tyres were inflated to the same pressure (to avoid any avoidable difference in tyre shape/size).

For the 'bike-only-standardised-wheels' test, we used a pair of Enve SES 4.5 wheels shod with 28mm GP5000 S TR tyres. The Enves gave us a fair standardised wheelset that has no direct affiliation with a bike brand, whereas if we used something like a Swissside wheel, it might have unfairly benefited the Van Rysel RCR-F, since the frame was designed in conjunction with the Swiss brand.

Each bike was tested with two bottle cages and bottles, because this is likely how the bike will be set up in the real world. We used the Elite Vico Carbon cage with Elite Fly bottles (both staples in the WorldTour).

Where a bike comes with its own aero-focussed bottle cages, we used those, because we believe that's what you'd do if you owned the bike.

We removed the out-front computer mount from each bike, since some didn't come with one.

For the rider-on-bike test, I wore the same kit, helmet, shoes and overshoes, and the position of the kit was kept as constant as possible throughout the day, even going so far as to draw lines on my legs for where the shorts should reach.

Present on the day

• Tom Wieckowski: On the tools. Preparing each bike for testing, helping to ensure the bike fits were matching, and data capture (bike weights, etc) on the day.
• Will Jones: Behind the lens, capturing the incredible photography you see here. Also, the presenter for our upcoming long-form video plans. More to be announced on this soon.
• Jamie Williams: Social video and a spare hand: supporting with bike fit, adjustment, bike swaps, wheel swaps, and more. Also, video director for upcoming long-form video.
• Ed Westrop: Behind the video camera, responsible for the long-form video capture and some of the footage on our social media.
• The staff from Silverstone Sports Engineering Hub: Operating the wind tunnel, swapping bikes between runs, and ensuring smooth, accurate testing throughout the day.
• Josh Croxton (me): A bit of everything. Chief planner, van driver, protocol designer, decision maker, test rider, and zip-tie-fitter. Also ate a lot of granola.

Confidence margin

This was calculated by testing the Emonda ALR in the morning, then again later in the day. From here, there's plenty of discussion surrounding the best way to calculate the error. Some will take the average measurement across all yaws and find the difference between the two tests. Others suggest going more granular and calculating the maximum difference between two individual yaw captures, which means a bigger margin (and thus, less confidence in each result's accuracy).

Since there are other variables at play here (as outlined below), we've opted for the latter approach, measuring the biggest variance between individual yaws. For simplicity, I've averaged the two tests for the Emonda ALR when showing them in the results below.

Reassuringly, our error margin this year is smaller than our margin in our previous bike test, and this is perhaps suggestive of a few things:

Firstly, with half a dozen big tests under our belt, our ability to perform repeatable, accurate tests has potentially improved over the past year. Secondly, for the with-rider test, I feel my ability to hold position has improved, too.

There are other things that may affect the accuracy of our results, too. These are:

• The handlebar position. Some bikes were sent with wider bars than others; they ranged from 37cm to 42cm. Using the 'Edge' markings that the wind tunnel projects in front of me to help me hold a repeatable position, I was able to adjust my hands to compensate for this a little, but imperfectly. With this said, though, our test was reflective of the way bikes are sold, so if it comes with 42cm bars, testing it as such is a fairer reflection of the bike you'd receive if you bought it, much the same as if it shipped with shallow wheels or round bars. Also, further to this point, two different bikes specced with 38cm bars could still put my hands in a different position (given we tested in the drops) as a result of differing drop depths, reaches and degrees of flare.
• Bikes were shipped with varying degrees of fit-adjustment opportunity, and since they're not our bikes, we were limited to how far we could adjust them. Stem lengths, as a result of one-piece cockpits, were set largely in stone, so we adjusted reach by changing saddle setback and my hand position on the drops, to ensure my torso angle remained as consistent as possible. Some one-piece cockpits couldn't be lowered without cutting the steerer, whereas some were cut to the minimum and gave very little room for raising. We chose to measure each bike, understand the options for adjustment, and then worked out a position roughly in the middle. In the end, we ran some with steerers clamped rather precariously, and I compensated for the position in the rider-on tests using the Edge outlines.
• In the bike-only test, the results would be affected by different bar widths as a result of the added surface area, albeit without the knock-on effect it has on rider position.
• Bike-only tests could also be affected by the shape of the saddle. Short of removing the saddle and spending much longer per bike, we chose to accept the caveat.

To counter all of the above points, we spent an entire day adjusting the bikes to get positions as consistent as possible, and I'm confident we got everything close enough that it didn't affect the results beyond our error margin.

Additional disclaimers and caveats

Cyclingnews does not claim that this data is the final word on the aerodynamic performance of the bikes featured, but rather an additional stream of independent, unbiased testing and information for our readers.

The data is simply a result of our day of testing. We understand that most brands have their own wind tunnels (or at least have access to one) and may see different results using different protocols. We hope that by being clear about our testing procedure, our readers (and maybe even the brands themselves) can understand the data and appreciate the results without losing sight of the bigger picture.

Speaking of the bigger picture, a reminder that there are plenty of other factors that affect the overall performance of a road bike.

Its weight, how it rides, the stiffness, comfort, component specification, and tyre clearance are all key factors that will change how fast it goes when you press on the pedals.

The way it looks is also important, as is the price, the pre- and post-sale experience, the build quality, and the warranty.

Perhaps not the latter few if you're a WorldTour pro getting bikes for free, but it certainly is for us mere mortals.

Beyond the initial CdA readings, our averages and subsequent calculations have been weighted to reflect the yaw angles in line with how often they are experienced in the real world.

It's well documented that when cycling, you experience low yaw angles more often than high yaw angles (more headwind than crosswind), so our results have been weighted in accordance with Nathan Barry's 2018 paper, A New Method for Analysing the Effect of Environmental Wind on Real World Aerodynamic Performance in Cycling. This ensures the data and our conclusions are more real-world applicable.

All the data below is subject to the respective confidence margins outlined above. It will be clarified and repeated where necessary to avoid any confusion or overstated accuracy.

Sponsor declaration

Before we get into the results, a brief interlude to say that this story includes paid product placement from Rapha, hence the use of the EF Team Rapha Pro Team Roadsuit.

I personally approved Rapha's inclusion in this test, as I do with all paid placement inclusions, having concluded it would offer no unfair bias – nor the appearance of bias – to the test.

I also needed something to wear, and a race skinsuit is more repeatable than a standard jersey and bibs, since the jersey portion can't twist, ride up and so on. I did initially try the Breakaway suit, which would almost certainly have been faster, but since I was on-and-off the bike all day, the less position-specific Roadsuit made sense as a more comfortable alternative.

Beyond that, the test wasn't in any way sponsored. We run these tests entirely independently. Each brand loaned us the bike specifically for the test (sometimes doubling up for a subsequent review), and the bikes will all be returned in due course. By the time you read this, most of them will already have been sent back.

As ever, we paid the standard commercial rate for the hire of the wind tunnel, and as ever, these tests cannot happen without our subscribers, so a fair and impartial protocol is essential to their success and our ability to continue doing them.

The results

And now for the nerdy fun stuff!

Here we'll start with the raw CdA for each bike at each yaw angle, shown as they came out of the wind tunnel software, to four significant figures.

We've then calculated to solve for power (watts), and worked out the difference between each bike and the baseline Trek Emonda ALR.

Below the results, we'll also solve for speed, as well as the time to cover a set distance at a set power output for each bike, and how those results compare to our baseline, too.

Not everyone can easily comprehend the significance of a 10-watt saving; some find speed easier to work with, while others prefer to think in time savings. We've got all bases covered.

These calculations will only account for aerodynamic drag, not drivetrain friction, rolling resistance, gravity, and so on. We're interested in the differences here, rather than the absolutes.

For clarity, the calculations used are as follows:

When calculating for watts, we use an air density of 1.2. This doesn't really matter since it's constant across all calculations, but it's there if you want to check my sums.

Bike only with stock wheelset

In the graph above, the green line at the top reflects our baseline Trek Emonda ALR. Reassuringly, there's clear daylight between that and the aero bikes, all of which are lower - and thus faster.

There's a clear cluster of bikes around the 0.8m² mark, a trio of slower bikes that sit a little higher, at around 0.9m², and then our baseline.

It may be hard to comprehend exactly what this means, but that's all explained below.

After weighting the yaw angles, here's that same data with the error margin overlaid.

Below, using that weighted average and solving for power, we have calculated the watts required to hit 40km/h for each of the bikes.

The above graph shows that the Factor Prototype requires 61.51 watts to hold 40km/h; two watts less than the Van Rysel in second place. Even with our 0.33-watt error margin accounted for, that's a clear victory for the Factor, by 1.34 watts. Error margins go both ways, though, so the victory could be as big as 2.66 watts.

The Van Rysel landed in second place, but thanks to that 0.33w swing, it could have been as low as 4th on another day.

Third place on the day went to the DARE VA-AFO, with a required input of 63.48 watts, but again, taking our error margin into account, that could drop to 5th on a different day.

The Colnago Y1Rs and Cervélo S5 landed surprising 6th and 7th places respectively, but could have been as high as 4th and 5th.

The Elves Falath EXP was the slowest of the aero bikes on the day, with a required power of 77.58 watts. There's a 3.2-watt deficit to the Bianchi Oltre RC above it, making the loss a clear-cut one, but even that is significantly faster than the baseline, which required 101.79 watts.

We fully appreciate that bikes can't pedal themselves and that the interaction of airflow between the legs and the bike isn't being accounted for here, but given the improved accuracy testing this way, we still think this test is useful for comparison purposes alongside the with-rider test.

With rider

Here, as before, there's a clear difference between the baseline Trek Emonda ALR and the aero bikes. Once again, those bikes all sit relatively clustered together, but with the Factor Prototype and Cervélo S5 sitting slightly lower. Remember that there's an error margin of 0.0021 to account for here, so many of the bikes clustered around 0.330-0.335 are statistically equal.

Here's the weighted CdA averages, with error margin overlaid:

As before, we will now use that weighted average and solve for power to calculate the watts required to hit 40km/h for each of the bikes.

Here, the Cervélo S5 landed a power of 273.12 watts, edging ahead of the Factor by 0.05 watts. This is well within our 1.72-watt error margin, so both bikes can lay claim to victory.

Both of those sit at an impressive four watts higher than the Colnago Y1Rs in third, meaning the top two spots on the podium are uncontestable outside of those two bikes.

With our error margin accounted for, the deficit between the Factor and the Y1Rs could be as little as 0.16 watts, but the Y1Rs could just as easily have landed as low as 6th place, with the DARE, Ridley Noah and Van Rysel RCR-F all being close enough behind to climb above it on another day.

The middle of the table is extremely close, with the DARE down to the Merida Reacto all sitting within the error margin of each other, and so any of those bikes could swap positions on a different day. Within this, the Elves Falath EXP landed a better result, sitting 8th on the day (and able to claim as high as 4th).

Below this mid-table cluster, the Bianchi Oltre RC landed a poor result at 285.24 watts, and was the slowest of the aero bikes on test. It's close enough to the Merida that it could leapfrog it on a different day, but only by 0.14 watts.

As expected, our baseline bike was still significantly slower, sitting at 300.7 watts. This is at the very least 12.00 watts behind the Oltre RC, and at least 24.12 (as many as 31.06) watts behind the Cervélo S5.

Bike only with standardised wheels

Here, we tested each bike with the same pair of Enve SES 4.5 wheels, Continental GP5000 S TR 28mm tyres, and a Shimano cassette (no gear shifting needed, so no detriment to the SRAM-equipped S5).

You can't fit thru-axle wheels to my old Trek Emonda with its quick releases, so that's not included in this test.

Many of the same bikes occupy the fastest spots, with the Cervélo S5 fastest at zero degrees, the Factor Prototype being fast across the board, and the Colnago Y1Rs performing well too.

Next up, the weighted averages with error margin, which are shown below:

And below, using those above averages, we've solved for power.

The Factor was once again the fastest, by 1.38 watts, making it another clear-cut win.

Where things get interesting, though, is when we see how each of these bikes fares with a switch to the Enve wheels from the ones they're sold with. Eight of the 12 bikes netted a higher (worse) CdA with the Enve wheels, whereas the other four saw an improvement.

The Van Rysel, which landed in 2nd place with its own wheels, was down in 5th place with the Enves, but close enough behind the Ridley Noah Fast 3.0 that it can claim a tie for 4th place (actually edging ahead slightly if we go to the third decimal place).

The 1.41-watt difference confirms the Swissside wheels on that bike are either outright faster than the Enves, or at least they work well with the Van Rysel bike. Given the bike was designed with input from Swissside, my money's on the latter.

The DARE continued to perform well with the Enves, losing just 0.21 watts. It actually closed the gap to the Factor by around a watt. The Colnago Y1Rs did a similar thing, but even more significantly, more than halving the 3.15-watt deficit; a combination of the Vision SC45 wheels slowing it down, and the Black Inc 62 wheels helping the Factor.

The result for the Argon 18 swung the other way, though. It will be sold with an updated version of the Scope Artech 6.A wheels, and landed 4th at 64.06 watts. With the Enves, however, it dropped to 7th, 1.65 watts worse off at 65.71 watts. We know those Scope wheels are faster than the Enves – they beat them in our 2024 wind tunnel wheelset test – so it was clearly being helped by those.

This is none more prevalent than for the Cube, though, which benefited from its Newmen Streem Aero wheels – complete with its pain-in-the-backside hidden valve system – to the tune of 2.54 watts.

That loss contributed somewhat to the 8th place the Cube landed in the table, a sizeable 4.59 watts off the Argon 18 above it.

Elsewhere at the foot of the table, we find the same suspects as above, but with very clear-cut positions for the bottom four. The Elves Falath EXP came in last, 14.43 watts down on the Factor. The Merida Reacto was 2.77 watts better, the Bianchi Oltre another 1.75 watts up, and the Cube Litening Aero C:68X a further 2.26 watts better off.

Watts saved vs baseline

Using the results above, this section will focus on the difference vs our trusty Trek Emonda ALR baseline. We've already seen that each of our aero bikes is faster than my 10-year-old winter commuter, but here we'll quantify exactly how much.

Again, we know bikes can't pedal themselves, but the Factor is an impressive 40.28 watts more efficient than our baseline. Switching to the Elves Falath EXP will net you the smallest saving, but a still sizeable 24.21 watts (+/- 0.33w)

With me aboard the bikes, the difference is slightly smaller, but still a handsome saving that will definitely be noticeable on the road or in a race result.

The Cervélo S5, fastest of the lot, will net you a 27.57-watt (+/- 1.73) saving, while bringing up the rear, the Bianchi Oltre RC will save you a still significant 15.46 watts.

2024 and 2025 tests combined

While it's not a fair test to compare the raw CdA of two bikes over two separate testing days, our use of a baseline allows accurate comparisons to be drawn.

By testing against the baseline bike on each day, and keeping that baseline the same each time, we can be confident in concluding that if a bike is 10 watts faster than the Emonda in 2024, it should also be 10 watts faster in 2025, irrespective of changes in conditions. So long as we keep the Emonda consistent, our results should be fair.

With that said, we did make one change to the Emonda for 2025. We lowered the stack by 10mm. This was to make it more closely match the rest of the bikes on the day. The fairest achievable position across the 12 bikes this year was slightly lower than last year's cohort, so to make for a fairer baseline comparison, we matched the position.

If we matched the 2024 cohort to the Emonda position, then slammed all of 2025's bikes without adjusting the Emonda too, then there would be an unfair advantage given to the 2025 bikes, as their relative position would be lower.

Everything else was kept the same, including the tyres, the bottle cages, and so on.

Evidence that this works is that we tested the same Scott Foil in 2024 and 2025, and it compared similarly in both tests (bike-only and with-rider) over the two years. In the bike-only test, it was 34.98 watts faster in 2024, and 35.66 watts faster in 2025; a difference of just 0.68 watts. With the rider on, it was 19.79 watts faster in 2024 and 19.15 watts faster in 2025; a difference of 0.64 watts. This is despite the slightly adjusted position on both bikes and a different rear tyre for the Foil this time round, but it falls well within our error margins of the two testing days.

Somewhat unsurprisingly, and reassuringly, the best of the dedicated aero bikes outperformed the best of the all-rounders. The Factor prototype sits clear atop the tree, with its 40.28-watt saving over our trusty Trek.

The fastest of our 2024 cohort, another Factor in the Ostro VAM, sits in eighth place at a saving of 35.06 watts, 5.22 behind its younger, mega-aero sibling.

The new Cervélo S5 beat its predecessor by exactly 2 watts, and even with our error margin, we can be confident that it's faster. How much of that came from the switch to SRAM's 1x Red XPLR groupset is untested, so a small pinch of salt to go with that result.

Meanwhile, the all-out-aero Van Rysel RCR F beat its all-rounder stablemate by 6.68 watts.

The Look Blade 795 RS, which was a notable outlier in last year's test, remains bottom of the table, but within the respective error margins when compared to the Elves Falath EXP above it.

The with-rider tests have a slightly more mixed bag of aero bikes and aero all-rounders, adding weight to the fact that a rider on the bike is a key part of the aero equation.

But, reassuringly once again, (albeit dishearteningly for anyone who wants to preserve the traditional bike aesthetic), the podium of our combined table goes to the trio of spaceships: the Cervélo S5, Factor Prototype, and Colnago Y1Rs.

The bigger error margin in our 2024 rider-on test means we're less confident in those results. As such, on a best-case scenario day, the S-Works Tarmac SL8 and Factor Ostro VAM can still lay claim to top spot in that field, as could the Trek Madone.

Down at the other end of the table, the Bianchi Oltre RC landed below the Look Blade 795 RS, but seven other bikes could have landed at the bottom on a different day.

Conclusions and takeaways

Which bike is fastest?

There are a few results here, and a couple of bikes atop the tree, but when analysing all three of the tests, I think it's fair to conclude that the bike with the bragging rights goes to the prototype Factor with its wide-set fork, super-deep bayonet-style head tube and progressive geometry.

It was the fastest in both of the bike-only tests by a significant margin each time, and it fell short of the Cervélo S5 in the with-rider test by just 0.05 watts; well within our error margin. The Cervélo also, theoretically, benefited from its 1x groupset.

How many watts can you save vs your old bike?

If we accept that our baseline Trek Emonda ALR from 2015 is a good representation of an entry-level aluminium bike (or any 'generic' old rim brake bike), then there are 27.57 watts on the table.

This is the difference between the Emonda and the fastest bike on test in the rider-on-bike test, ignoring our error margin. With that added in, the saving ranges from a minimum of 24.12 watts to a maximum of 31.03.

If you're brand new to cycling, then such an upgrade won't replace some consistent training. And if you're running entry-level tyres or wearing budget kit, then similar improvements can be found by upgrading those items first, and for a lot less money, but 31 watts is absolutely noticeable.

And if you already own one of these bikes, the switch from the winter commuter to your summer race bike is more than just a fancy-bike morale boost.

Funky bikes are fast bikes

I mentioned up top that nowadays, as bike design gets more creative and brands look to eke out as many watts as possible from their machines, a few of the latest cohort resemble spaceships more than bikes.

It'd be a bit of a PR disaster for those brands if those bikes weren't fast, but fortunately for them, it seems that they are.

The Ridley Noah Fast, with its super deep head tube, held its own, especially at lower yaw angles. The Cervélo S5 did too. The Colnago Y1Rs, which set the internet ablaze with its radical looks, was also one of the better performers, while the Factor prototype (which was almost banned by the UCI days after it leaked at the Critérium du Dauphiné) is a standout performer.

I've heard well-sourced tales of other brands doing equally wacky things too - design tweaks which fall foul of the UCI's proposed new rules - so watch this space, because aero bikes are only getting wilder.

More money doesn't always equal more fast

The retail prices of each bike will be listed below, but it's fair to conclude that spending more money doesn't by proxy mean you get a more aerodynamic bike.

The Elves Falath EXP might not have been the fastest on show here, but at a little over £5,000 for a Dura-Ace spec, it held its own impressively for a bike that costs less than half of some others in this test.

Likewise, Decathlon's Van Rysel is a brand that's building its reputation on value for money. The RCR-F is a hard sell to describe as 'cheap' at £10,000, but it far outperformed the Bianchi Oltre RC, despite the Italian machine costing over £1000 more.

The UCI is stuck in a contradiction cyclone

Much talk this year has surrounded the UCI's proposal to change the rules surrounding narrow handlebars, the maximum depth of rims, categorisation of helmets, and trials on gear restrictions.

The UCI's motivation is to slow down the peloton with the aim of improving rider safety, but just a few years ago, it announced rules that massively upgraded the aerodynamic opportunity for bike designers.

It dropped the 3:1 ratio rule (in which a tube could measure a maximum of 300% of its opposite plane) in favour of a 'minimum dimension' of 10mm in any plane and a maximum of 80mm. Head tubes, though, can now be up to 160mm deep, thanks to updated rules surrounding compensation triangles in that area.

The DARE and Ridley have clearly taken advantage of this, while Colnago, Cervélo and Factor have adopted a bayonet-style steerer, which essentially adds even more depth.

The result, evidenced not only here but more so by the sheer fact that bikes are being designed this way, is that bikes are now much faster than they were 10 years ago. Just ask Bianchi, whose Oltre RC was designed before those rules came to be.

To now announce proposals to slow down the peloton just two years later, while inadvertently negatively affecting the bike fit of smaller riders, costing brands hundreds of thousands of pounds in research and development, and forcing groupset manufacturers into a product overhaul, despite the proposal being debunked by anyone with a basic grasp of physics (and more eloquently put by Red Bull-Bora-Hansgrohe's Head of Engineering, Dan Bigham), is both deeply damaging to the industry and utterly ineffective in aiding rider safety.

It smacks of a whimsical, capricious approach to equipment rulemaking that is already proving bad for the bike industry. That itself is suggestive of a shortsighted vision for the future of our sport as a whole, which likely bleeds into other areas of rulemaking in the sport, such as the financials, television rights, equality, and especially safety.

Individual bike results

The individual results listed below are done so in ascending order based on our rider-on results, as I had to pick one of the three datasets and it's the most real-world of the bunch.

Baseline: Trek Emonda ALR 2015

Image 1 of 3

This is our baseline, the comparison, the zero point against which we'll compare all of the modern 'aero' bikes.

It's a Trek Emonda ALR, launched around 2014 and this is the 2015 colourway. I bought it around 2017 as a winter bike, and it served me well for around three winters, but has long since become shed fodder, worth too little to bother selling on Facebook.

Our reason for using this is twofold. Firstly, it's ours. We can keep it in this very state and bring it back next year, the year after, and have a constant across various tests.

Secondly, we think it loosely represents Your Generic Road Bike: round tubes, aluminium frame, round handlebars, external cables, rim brakes and shallow wheels. If we accept that its performance would be similar to other entry-level bikes, you can get a sense of roughly what performance will be gained if you choose to upgrade your own bike.

Bianchi Oltre RC

Image 1 of 3

There's no mistaking the Bianchi Oltre RC's intentions. Its wild design, unique cockpit with a low-pressure-generating hole through the middle, and F1-inspired 'air deflectors' on either side of the head tube confirm the Italian brand went all-in in its pursuit of aerodynamic performance.

Sadly, those deflectors were unavailable for our test, as I'd have been intrigued to see how much performance they added (or Arkea-B&B Hotels were losing as a result of the UCI's decision to forbid them), but as ever, we're at the whim of what brands are able to send us. In this case, it was a size 55cm model equipped with Dura-Ace, and with a front-end that was slammed as low as it would go. This should have advantaged the bike, but evidently it didn't.

The brand might have called it a 'hyperbike' at its 2022 launch, but our testing has it in the relegation zone of both bike-only tests, and in last place in the rider-on test. In the latter, it is still over 15 watts better than our baseline Emonda, but it netted the worst performance we've seen to date, including last year's test.

Regardless, I must give credit to the Italian brand. Having spoken to Bianchi's Global Marketing Manager, Claudio Masnata, it's clear they knew the Oltre RC wouldn't compare well against newer bikes like the S5. They contemplated refusing to be involved, but instead decided to participate.

I'm also told that although it was launched in 2022, it was delayed by the pandemic; its design was finalised two to three years prior. This means it was working according to the old UCI rules, which in Masnata's words, "compares apples to pears."

Going back to the mention of Arkea-B&B Hotels, if one of the team's riders was in a flat breakaway with a rider from Visma-Lease A Bike (on their new Cervélo S5), holding 45km/h and sharing the work evenly, we've calculated they'd be disadvantaged to the tune of 13.8 watts.

Regardless, Bianchi is strongly rumoured to be switching its allegiances to Bahrain-Victorious next year. Masnata wouldn't give anything away there, but if I read between the lines, a new team and an aero bike designed five to six years ago suggests another Oltre might be on the way.

Merida Reacto 9000

Image 1 of 3

This version of the Merida Reacto has been around since 2020, and if the Merida aero bike that leaked at the Tour de France is anything to go by, it's about to be replaced.

We tried to get that newer bike, but to no avail as the brand was unable to supply it in time. Still, we thought it pertinent to include this model, which has been a staple in the WorldTour for half a decade beneath the legs of Bahrain Victorious riders.

It is the only model in this year's test with a second-tier spec, with its Shimano Ultegra groupset and Reynolds 'Expert' Black Label 60 wheels (rather than the top-tier 'Pro' wheels. Regardless, the cockpit was a one-piece Vision affair, and our switch to Enve wheels only netted a 0.06-watt improvement, so the Expert options weren't holding them back much.

Still, like Bianchi above, I don't think Merida will be too upset that its five-year-old bike with second-tier spec was within 10 watts of the latest mega aero machines.

Scott Foil RC

Image 1 of 3

Scott's Foil RC aero bike is a year older and a year closer to being upgraded since we first tested it in a wind tunnel, but we've not heard any rumours of that happening any time soon.

Scott, as a brand, deserves a lot of praise for its work in recent years. The brand launched the Addict RC back in November as a 5.9kg lightweight all-rounder, which was described as one of the very best road bikes on the market right now in our Addict RC Ultimate review. Its Cadence aero helmet did well in our first helmets wind tunnel test and Tom's subsequent review, and sister brand Syncros' Capital SL wheels performed very favourably in our wind tunnel wheels test last year, as did the Cadence helmet.

The Foil RC Pro also did well in our test last year, landing 2nd in the bike-only test, albeit close enough to those around it that it could climb to 1st or drop to 7th once the error margin was considered. In the with-rider test, the result was a little more average, but still close enough that it could claim victory.

This year, it performed very similarly in comparison to the Emonda (as it should, given the minimal change to either bike), but it landed a more mid-table performance when compared to the cohort of modern bikes. In the bike-only tests, it landed 8th with its own wheels and 6th with the standardised wheels. And in the with-rider test, it can claim no better than 4th.

Argon 18 Nitrogen Pro

Image 1 of 5

A case of right-place, right-time led to us including the Nitrogen Pro from Argon 18.

The brand's engineers came to visit Silverstone Sports Engineering Hub while we were testing aero wheels last Autumn. There, with Scope's permission, we left behind our loaned pair of Artech 6.A wheels for them to test the following day.

Fast forward to July, a month ahead of our bike testing day, I received an email about a mystery new aero bike from the brand, and by chance, the brand was returning to the wind tunnel a week prior to our August testing day. It was an opportunity to repay the favour, leaving behind its latest creation for us to include in this test.

The bike, I've since learned to be called the Nitrogen Pro, is being launched at the time of publishing this article (Weds 10th September), concurrently with the results of our testing.

Among its standout features is an updated version of the Scope wheels, designed in conjunction with Argon 18's R&D lab, ATTEN, specifically to work (in aerodynamics and comfort) with the Nitrogen Pro frame and a 30mm tyre.

Somewhat ironically, given this is a wind tunnel test, the brand's marketing details a bike that has been designed to be fast "in real-world conditions, not just a wind tunnel."

Take our weight figure with a pinch of salt above, though, since it's a pre-production sample designed for aerodynamic profiling alone, rather than riding. The brand has kindly sent photos of its official weigh-in, which shows an unpainted 56cm frame at 6.77kg with SRAM Red 2x and bottle cages fitted.

Elves Falath EXP

Image 1 of 5

For this year's test, we wanted to include a 'challenger' brand born out of the OEM space, in response to our research into Chinese OEMs' designs on the US market.

Thanks to its 2023 launch of a UK distribution arm (meaning simplified access), and the recent launch of the Falath EXP aero bike (complete with its Madone-like hole in the seat tube), Chinese brand Elves, which also has offices in Taiwan and Hong Kong, was the perfect fit.

Incredibly, the Falath EXP is priced at £5,099 for a frame made using a mix of T800 and T1000 carbon, equipped with Dura-Ace Di2, deep carbon wheels and a one-piece cockpit. That, admittedly, is an introductory price with a £1000 discount, but even at the eventual retail price of £6,099 it undercuts the next cheapest bike in this test by £1400, and is less than half the price of the Colnago and Argon 18.

Beyond the price, the bike's biggest USP is undoubtedly the Inverted Triangle Airflow Channel (ITAC), which it says reduces drag. It's a divisive aesthetic, but one that is becoming more common thanks to Trek's Madone and, to a lesser degree, Hope's latest HB.T track bike with its split seatpost. While we're on the subject of aesthetics, I'd be remiss not to call out the curved seat tube's inability to uniformly follow the shape of the rear wheel, though.

Like all other brands here, Elves has seen these results before they were published, and described it as a "phenomenal result" to hold its own against household brand names.

Despite bringing up the rear in both bike-only tests, it still outperformed the Look Blade 795 RS from last year's test, and it climbed to a mid-table result with me aboard. Elves attributes this to the tube profiles, claiming they were specifically designed to channel airflow efficiently around the rider.

The brand also champions bike fit and its importance on aerodynamics, adding that customers can "ensure correct positioning" at the point of purchase.

Cube Litening Aero C:68X

Image 1 of 3

Taking the Elves out of the equation briefly, the Cube Litening Aero C:68X is a very good value bike, at least in terms of the spec list. At just £7,499, you get a dedicated aero frame, one-piece cockpit, integrated cabling and Shimano Ultegra groupset.

Although there's nothing dramatic or radical about its design, there's no mistaking its aero intentions, with its deep tubes and deep Newmen Streem wheels, complete with their bladed aero spokes and hidden valve system.

Those wheels certainly offered a performance gain, since the bike's performance dropped by over two watts when we swapped to the Enve wheels. I can't mention them without discussing how difficult they are to work with, though. Instead of a standard valve and small hole, they were shipped with a TPU inner tube with a short Schrader valve, which is locked into place using a nut. Access to the nut, and indeed to the valve to reinflate the tube, is available only via a special tool, which you'll need to carry everywhere you go.

The Litening Aero C:68X performed reasonably well, especially given the price difference between it and some of the fastest bikes here, but I feel was let down slightly by its wider-than-average handlebars.

Van Rysel RCR-F

Image 1 of 3

While Van Rysel has built its reputation on price, it's hard to describe the RCR-F as a value bike, given its £10,000 price. In this group, it's among the cheaper of the top-tier bikes, with the Colnago and Cervélo tipping the scales at over £2000 more, but it's still massively undercut by the Elves and the Scott, both with the same Dura-Ace groupset.

It landed a reasonably mid-table position in the with-rider test, but with our error margin taken into account, it can lay claim to the third spot on the podium (it could also drop to 11th).

In the bike-only test, it snuck a 2nd place behind Factor's outright victory, but when we swapped to the standardised wheels, it dropped down to 5th place.

Ridley Noah Fast 3.0

Image 1 of 4

At face value, the Ridley Noah Fast 3.0 is just a bike with a mega deep head tube; everything else looks surprisingly normal... at least in the context of deep-tubed aero bikes. But it is designed to put the rider into a 'modern', forward, aggressive race position, and in Will's Ridley Noah Fast 3 review, he raved about it. Of all the bikes in this test, he was adamant on keeping hold of this one a little while longer for 'further testing'; he enjoyed it so much.

Ridley's designers tell me there was also a focus on low yaw performance, which shines through in the data here.

On our day in the tunnel, it performed relatively consistently across all three tests. It was 5th in both of the tests with its own (DT Swiss) wheels, and climbed to 4th in the Enve wheels test. It can lay claim to a podium finish in the rider-on-bike test, and could finish as high as 2nd in the standardised wheels test.

One downside I noticed when setting up was that the cockpit couldn't be lowered without cutting the steerer, even just for a test ride to check it's comfortable. As a result of this and its shallow drops, it was one of the taller-stack bikes in our test, approximately 10-15mm higher than the Factor, Colnago and Cervelo.

Ridley would no doubt argue that chopping the difference off the steerer tube might have pushed it higher up the list, and given it offers a configurator at the point of purchase on its website, there are opportunities for further gains with bike fit there too. Whether that would have been enough to climb above the Factor or the Cervélo is impossible to say, but to take a step back for a second, the margins are small enough that if I were choosing which to buy, I'd be looking beyond the aero figures anyway.

DARE Velocity Ace-AFO

Image 1 of 4

The DARE Velocity Ace-AFO, (from here on in to be known as the VA-AFO), was first spotted in the pro peloton during the Spring Classics in 2024, and made headlines with its uniquely deep head tube (this was before Ridley's Noah Fast 3 had broken cover), and similarly deep cockpit.

It was arguably one of the first to really capitalise on the UCI's updated rules that governed frame/handlebar tube dimensions and compensation triangles, at least in the road bike category.

Alongside the Bianchi, it's one of two bikes that were received slammed to the hilt. This made our lives slightly tougher when it came to matching positions, but a bit of precarious clamping on the very top of the steerer tube allowed us to run it higher. I wouldn't want to run it on the road that way, but for a few minutes in a wind tunnel on a bike fixed in place, I took my chances. This meant a small amount of steerer tube (which the proprietary spacer would otherwise fill if we had it) was exposed. This is a tiny detail in the grand scheme of things, but one that might see the VA-AFO land a marginally better result with a perfect setup.

Speaking of results, it didn't do too badly all told: 4th with a rider and the ability to lay claim to third, with a minimum saving of 18.48 watts over our baseline bike; 3rd in the own-wheels test, a worst-case of 5th with our error margin; and 2nd in the Enve-wheels protocol, just 1.38 watts behind the Factor.

Colnago Y1Rs

Image 1 of 3

The Y1Rs nigh-on broke our small corner of the internet when it first leaked, with its unusual floating seat tube and crazy Y-shaped cockpit causing uproar among cycling's purists; it's certainly one of the more creative takes on aero bike design, and I'm not shy to admit I was quick to put it on blast in the Cyclingnews office chat.

At its launch in December 2024, the Italian brand made the bold claim that it was the fastest in the WorldTour, and it's since been ridden to Tour de France domination by Tadej Pogačar (though I'm aware that he'd likely do the same on any of the aero bikes listed here).

Given it came third in our with-rider test and the two bikes that beat it were released more recently, we can't argue with the fastest-in-the-WorldTour claim, although our error margin - which is likely bigger than Colnago's, since they use a pedalling mannequin - does mean it could drop to as low as 6th place.

It is the most expensive bike on this test, though, at £12,600. It nets you a 23.91-watt advantage against our Emonda, which works out at £526.98 per watt. There are certainly cheaper ways to go faster, I'll tell you that for free.

Factor Prototype

Image 1 of 6

Undoubtedly the most radical design here, Factor's UCI-legal pre-production prototype features a wide-set fork and seatstays, über-deep head tube with bayonet-style steerer, and hints at a very progressive geometry if you look beyond the unorthodox shape.

It first leaked at the Critérium du Dauphiné, and within 24 hours, the UCI had issued statements banning forks outside a certain width. At the time, it seemed reactionary from the governing body, and it appeared the prototype was going to be banned before it could even be launched.

But Factor tells me it had an inside line on the dimensions for that upcoming rule, and evidently its designer, Graham Shrive, squeezed it right up against the limit.

Since the bike hasn't been launched yet, there's still a whole heap we don't know about it, including its name, the price, and exactly how it rides (sadly, this one was a pre-production prototype, so I was forbidden from taking it home).

We do know that it was insanely fast across all three tests, though. It outright won both bike-only tests, and only lost to the Cervélo S5 in the with-rider test by 0.04 watts (well within our error margin).

Look at the yaw graphs above, and it's clear that the Factor retains a low CdA across the full yaw sweep. In the bike-only tests, while most bikes create more drag at higher yaw, the Factor's CdA actually reduces. It's not quite as significant as when a wheel 'sails', but it's impressive.

It also performed fastest at zero yaw in the with-rider test, and despite weighting low yaw more heavily, it's the slightly better result of the Cervélo at +/- 5 and 10 degrees that netted it the marginally better result. Given the Cervélo S5 had a 1x groupset, I'm minded to think this would climb above it under equal protocol, though.

And you'd be forgiven for thinking all these deep tubes and super-wide forks make for a bike that weighs as much as a small horse, but you'd be wrong. At 7.35kg (without pedals and cages), it was the second-lightest on test (behind the Bianchi).

Cervélo S5

Image 1 of 4

Cervélo's old S5 came out as one of the fastest in last year's test, and the same can be said for this year's test, netting a marginal with-rider victory with a worst-case 2nd place considering our margin of error. That puts it at the top of our combined results table, with the Factor close behind.

In the bike-only tests, it landed a slightly more mid-table result; 7th and 6th, but Cervélo will likely be quick to point out that bikes don't ride themselves.

The small caveat to this data is that it's the only bike on test this year with a SRAM groupset, and the only with a 1X chainset, too. Regardless, a win is a win and we're sure the Canadian brand will take it.

During setup, I particularly appreciated the bike's ability to adjust the height of the cockpit without having to awkwardly place the proprietary aero spacers back atop the stem or cut the steerer tube. This allowed us to test the Cervélo with 20mm of spacers removed too, to see the difference in a controlled setting.

The result was expectedly faster; 0.45w more efficient in the bike-only test, and 1.79w faster with rider. The latter is within our error margin, but gives us a clue to the advantage given to the Bianchi and DARE, which had lower front ends.

Thank yous

Many words later, I'd like to round off this feature with a few obrigados to the people and brands that helped to make it happen.

Rapha: For the paid placement support and loan of the Rapha Pro Team Road Suit. It doesn't affect the impartiality of our test, but helps us do more of them.

Each of the brands featured: Thanks to the 12 manufacturers of the bikes featured in this test. Such an empirical test as this must have a loser, and we appreciate that accepting our request involves a risk, as well as a logistical (and oftentimes financial) cost to get the bike to us.

Argon 18 and Factor: It's not uncommon for a bike brand to send a production-ready sample to media for an embargoed pre-launch review of the bike, but to send a pre-production sample for an empirical test like this carries a risk. If it does badly, it's not just one reviewer's opinion, but data laid bare for all to see, and the bike's launch could well become a flop as a result.

Trusting that the bike is good enough to take that risk shows a confidence that should reassure potential buyers, and trusting us to perform a fair test is truly an honour. We take the responsibility very seriously.

A special thanks to Factor, too, from which a staff member personally flew from Girona to the UK with the bike and hand-delivered it to Silverstone Sports Engineering Hub to ensure we could test it.

Silverstone Sports Engineering Hub: For their guidance on protocol, operating the wind tunnel, swapping the bikes over between tests, and for opening up early so we could get through our long list of tests in the day.

Madison UK: For the loan of Elite Vico Carbon bottle cages and Elite Fly bottles.

Continental Tyres: For the loan of GP5000 S TR tyres.

And most importantly, Cyclingnews Members: Without your support, these tests wouldn't happen. I enjoy making them, and the data tells us you enjoy reading them, so thank you.