5.5 kg, TT bars, and a fairing: The best Everesting bike

A deep dive and video on the challenges and thinking behind optimising an Everesting bike.

Don’t miss out on the latest CyclingTips updates.

Assuming you visited CyclingTips at least once over the past few days, you’ll have seen that I broke the Everesting world record last week and documented that ride in a video. I briefly held the Everesting record last summer with a “backyard” style attempt. For that ride, I had focused on maximising the bike and components I had, with virtually zero budget for buying speed. 

Having set a time with a short-term training focus and shoestring budget approach, I was keen to see what my absolute best could be with months of focus and the ability to pick and choose which components to use. Here’s what I ended up with.

What goes up must come down

With the support of people and brands worldwide, I set about formulating the perfect Everesting bike. On the face of it, the solution seems simple. Just aim for the lightest setup possible. But that would completely ignore a large chunk of the equation. I would need to complete a descent for each of the ascents, and time spent descending is time I am not climbing towards the height of Everest. I needed to optimise for both weight and aerodynamics.

The spreadsheet that just kept growing.

I had a huge spreadsheet of component weights, calculations, lap times from the previous Everestings, gear ratios, and theoretical build options. This data, combined with support and suggestions from a group I have now named the Secret Sheep Club, was the starting point for what would eventually evolve into the Frankenbike we see today.

The Secret Sheep Club is my version of British Cycling’s Secret Squirrel marginal gains group. My fellow secret sheep included David Bailey, Robert Chung, Tom Anhalt, Josh Poertner, Dan Bigham, Chris Bloomfield-Brown and Federico Fontana. Look out for videos on the science of Everesting from this group soon. 


A lightweight and aerodynamic frame was an essential starting point. The wrong choice here would make the weight and aero targets very difficult to achieve. 

I had never ridden a TCR previously, but when Giant offered a 2021 rim brake TCR Advanced SL and suggested they could remove all the paint to save every gram possible, I was all ears. I had spent the winter moving my cleats as far back as possible in a bid to shrink myself onto a smaller frame. This is a tactic often used to achieve CdA (drag) reductions for time trials, but the smaller frame, less seat post, lower handlebars, and a smaller frontal area would also help for Everesting. 

I had spent an unholy amount of time on Zwift over the winter partly to avoid the miserable weather but primarily to achieve specific adaptions in both my training and setup. I spent most of the rides with my front wheel propped up at an 8° angle to replicate the gradient on Mamore Gap and train the muscles and pedalling technique required to Everest at speed. 

With the new cleat and saddle positions as low as possible, I then used motion sensors to validate and replicate this new position on the TCR. 

Ultimately the TCR I used is a size medium and weighed in at 772 g before I cut the seat post. 


The drivetrain was undoubtedly the area of focus that caused me the most brain ache. Again, both weight and aero were the main considerations, but efficiency and gear ratios also had to be considered. 

As detailed as I got with all my other calculations, I opted for pure ignorance with the gearing. My theory was, if I Everested in 7 hours and 4 minutes with a 39:32, by opting for a bigger gear, I just had to ride quicker. I justified this by telling myself I was also lighter and faster, so I wouldn’t need the same gearing anyway. 

Ultimately I erred on the side of caution and used a combination of XTR and Dura-Ace sprockets to create a custom seven-speed cassette (eight if you count the 11 sprocket holding the cluster together, but that I couldn’t engage). The plan was to use the 44:35 most often, with a 44:40 option for when fatigue really set in. 

While the gear size was slightly bigger, the combination used to achieve this gearing was greatly increased. I switched from a 39 upfront to a 44 and from a 32 on the rear to a maximum 40T sprocket. While this would result in a small weight penalty, the increased sprocket sizes should have resulted in a considerable efficiency gain. The decreased chain articulation with the larger sprockets offers an estimated third of a watt saving for each extra tooth, a marginal gain that can quickly add up to a substantial saving. 

A lot of time and thought went into the drivetrain choices.

While gearing for the uphills was crucial, I wanted to have some options to help with accelerating back up to speed at the start of the descent each lap. With that in mind, I included sprockets right down to a 17. This made a significant difference for acceleration after the turnarounds at the top and bottom.

While removing three sprockets from the cassette had a weight-saving benefit, the primary reasoning was to shift the remaining sprockets more centrally on the freehub. This ensured a straighter chain line, less cross-chaining and, overall, a more efficient drivetrain. I used a single-speed spacer to replace the sprockets removed. 

While the more centrally located and larger sprockets offered efficiency gains, they did present a problem. Derailleurs are not designed to run this combination so early in their range. If I ran a standard setup, the derailleur would undoubtedly catch on the cassette. The Wolf Tooth road link offered a solution and with an adjustment to the b-screw, the shifting was surprisingly good. 

On the topic of shifting, a Di2 groupset was selected rather than a lighter mechanical setup. Di2 offered two main advantages: firstly, using TT base bars (more on that later), I had to use TT brake levers. Secondly, by integrating an SM-EWWU111 wireless unit, I could analyse shifting patterns during the test rides to further validate my gearing selections. 

I used an R9150 rear derailleur paired with R9160 time trial brake levers with integrated shifters. The battery was placed as low in the frame as possible to help lower the centre of gravity and speed up those turns a little. The front derailleur was ditched to save both weight and aero drag.

I had initially included Di2 climbing shifters on the bar tops but removed them after my ‘Basecamp’ (Half Everesting) ride due to lack of use.  

Last but by no means least on the hunt for drivetrain efficiency, I turned to CeramicSpeed for some help. The Danish brand duly obliged with its coated OverSize Pulley Wheel system, UFO-treated chain, coated ceramic bottom bracket, and coated hub bearings. In the ultimate race-day-only upgrade, these bearings were oiled rather than greased to reduce drag even further. 

Front end

Not much to see here.

I guess now is a good time to tackle that front end. While I optimised and hacked the frame and drivetrain as much as possible, they still resembled a normal road setup. The same cannot be said for the front end. 

I decided to opt for a flat time trial base bar very early in the process. I had cut the drops off my Tarmac SL6 last year to save a tiny bit of weight and reduce the aero drag of the handlebars. Josh Poertner of Silca and Marginal Gains fame first suggested running a base bar to further improve aerodynamics. 

I had initially decided to use a TriRig Alpha integrated bar and stem, but the shift to the TCR and OverDrive steerer tube meant this was no longer an option. That was when I found the Deda Tribar. At 200 g it was far from the lightest option but should have offered some significant aero gains. With the adjustability of a standard stem and bar interface, I could also adjust the bar angle to provide a comfortable grip on the bull horns. 

Deda Tribar and TriRig Delta.

The TT base bar provided a nice aero gain for minimal weight gain, but I still wasn’t happy. I still had cables, and a brake calliper stuck out in the wind. After considering many options, including a drill (enough said), I opted for a TriRig Delta. With its claims of “instant superbike” aerodynamics by hiding the cables and callipers, plus its cut-to-fit adjustability, it seemed like the perfect solution. 

However, the Delta is designed to integrate with those Alpha bars that didn’t fit the new forks. Thankfully the hacksaw, some tape, and a file were on hand to make it work. Warning – I am confident this hack voids the warranty.

The final sized and cut Delta did add approximately 150 g to the total weight, but cleaned up the bike’s front end.  



Somewhat controversially and against some of the Secret Sheep Club’s recommendations, I decided to opt for rim brakes. A rim brake setup is significantly lighter than disc brakes, and I was confident I was already at the limit of my skills for late braking on the descent. 

Once I’d decided on rim brakes, the logical option was Cane Creek’s eeBrakes. At just 172 g for the pair, they weighed similar to a single Campagnolo calliper I had used last year. More importantly, the El TD limited edition in pink, turquoise, and silver offered the chance to show off my own design tastes and add some much-needed colour to the very stealthy bike. The only problem was that I then went and hid one of them behind the Delta fairing. 


I had initially planned to run the race-day-only (and discontinued) Speedplay Nanogram pedals. At a mindbogglingly light 65 g per pedal, they were the natural option. Unfortunately, part of the weight-saving with these pedals comes from the shorter spindles and ultimately, these proved too short for my large feet. 

Thankfully I had a set of the new Wahoo Speedplay Nano pedals in for review, and these duly stepped up and onto the Everesting bike. I persisted with the Nanogram version one style Speedplay cleats and single-use only mounting hardware for a few precious grams of savings. 

I mounted the cleats to Giro Imperial shoes. I had tried several lighter shoe options but ultimately settled on the Giros and their double Boa dial, which offered more adjustability on the fly during the seven-hour ride. I forgot to take the heel pad off, costing me tens of milliseconds I’m sure.


I opted for a pro edition 131 mm-wide Cadex Boost saddle. While I did have the option of a carbon-only saddle from AX lightness at 93 g, I decided to opt for the additional padding on the Boost at the expense of 27 g. While I had used the full carbon option for a Basecamp ride last year without issue, the Boost was more akin to the shape of the saddle I usually ride. 


Sub-1 kg carbon wheelset from AX Lightness.

This proved another difficult decision to make. I had the choice of AX Lightness Ultra 38T and Cadex Tubular 42 mm rims. Ultimately the decision was taken out of my hands as the front Cadex got damaged in transit, and I had to at least start with matching wheels for the ‘gram snaps. 

I was somewhat relieved the decision was made for me as AX Lightness offered a crazy-light sub-1 kg for the pair, whereas the Cadex wheels had claimed aerodynamic benefits and looks to die for.

Mounted to these wheels were Vittoria Corsa Speed 2.0 tubulars. The Corsa Speeds are known for incredibly low rolling resistance and low weight. This does come at the expense of robustness, which I was all too aware of when selecting these tyres. While the same model of tubular had survived Everestings on the same road last year, I was not quite so lucky this year. This was mostly down to me pushing the braking point as late as possible on the descent and at least once locking up the rear wheel for a split second. 

The Corsa Speed tubulars measure out at 24 mm when mounted, and I opted for 74 psi up front and 76 on the rear. 

Data collection

Let’s get to the juicy stuff. I am a data nerd, at least when it comes to cycling performance metrics. On the day of the Everesting, I ran a Dura-Ace R9100p power meter and Wahoo Tickr linked to an Elemnt Bolt. While this is a fairly straight forward setup and certainly on the lighter side of the options list, things got much more detailed during preparation for the attempt. 

Also linked to the Bolt was a CORE body temp sensor. I had used this in preparation for the Everesting to track core body temp during training. I performed a heat ramp test to get my heat training zone, which I then used to achieve some heat adaption physiological gains. I also used the CORE on my Basecamp test ride to assess cooling strategies and clothing selection. On the day, I opted to wear it for post-event analysis purposes and opted not to have core temp on screen for the ride.  

In the weeks ahead of the ride, during the Basecamp test, and for the Everesting itself, I wore a Supersapiens continuous glucose monitor. This paired to an app on my phone to give insight into my glucose levels, response rates to different foods, and help validate my fuelling strategies. 

Stay tuned for reviews of these items in the near future. 

The pacing plan was simple: stay under 5:22 average lap times.

These validated fuelling requirements were met with Science in Sport Beta Fuel drink mix and Go energy and caffeine gels. 

I mentioned earlier using motion sensors to validate and replicate my new position. I used the Leomo Type S for this. Having bought the Leomo more than a year ago, I have had plenty of experience with it. Although frustratingly unreliable at times, it is useful for this purpose. I used the motion performance indicators to track pelvic angle, rock, and rotation, plus leg and foot angular range.  

I also used a Notio Konect for the Basecamp ride and some training rides to track CdA and environmental parameters for the Secret Sheep Club. These metrics were used to help with skinsuit and overshoe selection, but ultimately I opted for the most ventilated aero options I had after noticing some overheating during the Basecamp ride.


Lastly and while on the topic of clothing, I used a Pactimo Summit long-sleeve skinsuit. While the design and colour choice are reminiscent of a traffic cone and while this suit was originally designed to comply with the night-time rules of the Donegal Wild Atlantic Way ultra-endurance ride, I was happy to be a bit more visible on the descents during the Everesting ride.  

I opted for the new MET Manta MIPS helmet. I had this helmet in for review recently, and it tipped the scales at just 247 g, the lightest of the aero helmets I had on hand. The inclusion of the MIPS technology was another large part in the decision to use the Manta. 

Ultimately I opted to skip the overshoes, which slipped down during last year’s Everesting, and instead opted for an old Specialized Sub-6 cover I had in my “random stuff” drawer.

So there you have it — all the inner workings of my Everesting bike brain. Be sure to check out the video above for more on this record-breaking bike and drop us a comment below with how you might have built this Frankenbike.

Editors' Picks