'19 minutes and 31 seconds later, I had a new 5K PB one almost a full minute better than Id ever achieved before': our writer tried a robotic exoskeleton and ended up setting himself a new running record

Date: Tue, 20 Jan 2026 14:42:14 +0000

Description:
Does your running PB count if you run like Robocop? Our writer got to grips with the Hypershell X Ultra exoskeleton

FULL STORY

Is it important to have a running goal? Some folks need to work towards a marathon, or a personal-best 5K time, in order to maintain the motivation to train.

I havent really felt that way in the five years since I started running with intention. But one benchmark goal has quietly wormed its way into the back of my mind since I noticed my 5K Parkrun event times gradually come down over
the years. I'm working on achieving a sub-20-minute 5K.

The closest I got to reaching it was back in October 2025, at a Parkrun event by the sea on Englands south coast, in Worthing, with a time of 20:28.

A sub-20 5K? Its kinda within reach, if I were to do a training block for it. But, to be frank, that doesnt seem all that appealing. With short days and frequently awful weather in the UK at the moment, running has become a way to get outside whenever possible, for a mental health boost. And cobble together what meagre amount of natural Vitamin D production actually happens under cloudy, grey skies. Short and sharp sessions arent where Im at.

So I started wondering if I could cheat my way to a notable 5K personal best, using some new technology to power my legs through those last 30 seconds.

Im not talking about one of the best smartwatches , fitness trackers , or
even efficiency savings via a pair of the best running shoes . Earlier in
2025, I got to try out a Hypershell exoskeleton, not that long after first bumping into the brand online. Ive been for a short-ish hike with a
Hypershell on, months ago, and while I did race a friend a few hundred
metres, for fun, this was just two minutes after putting the rig on, when its at its most unnatural-feeling. It felt weird. Wrong, even.

But if youre anything like me, you have to be curious about whether this Hypershell stuff actually works, even if you hate the idea of it being used
in any form of competitive running scenario. So I asked the company if I
could borrow one for a few weeks, to see if robotic leg assistance would help me shave 30 seconds off my time. And they said yes.

Its probably about time we deal with how a Hypershell works. I have the Hypershell X Ultra, the latest and most powerful version of the exoskeleton.
It uses a carbon fiber reinforced frame that sits over your legs, using
anchor points just above your ass, and just above each of your knees.

Motors reside in the enclosures that sit roughly above your hip joints, used
to generate the force that pulls your legs upward as you stride as you walk
or run. Youre the puppet pulled, not pushed.

The Hypershell X Ultra uses algorithms to judge when to apply force and how much of it is needed at any given millisecond. And what it needs to do on
that front is going to change with your pace, your stride arc all sorts of stuff all the time. And as it can deliver up to 1000W of peak (not
sustained) force according to Hypershell, getting this stuff right is important.

Dont focus on that 1000W figure too much, though. If the Hypershell X Ultra were to output at 1000W, the 72Wh battery would last less than five minutes. But its obviously only going to hit anything like its peak power in certain modes, for a fraction of a second, during a certain point of your stride. The more important bit is this is the most powerful of Hypershells models, which
it says can assist at a pace of up to 25km/h.

We should have plenty of headroom for our goal. At 25km/h, youre finishing a
5K race in 12 minutes, and beating Joshua Cheptegeis 12:35 world record for that distance. Im looking for a modest 20.

Test 1: A local running route

But you know the real first challenge of a Hypershell X Ultra? Getting out of the door with one on. You dont half feel self-conscious with this kit
strapped in, and youll actually look worse if you put clothes over the top. Dont even try to hide it, youll look ridiculous either way.

To steel myself for the proper attempt at a 5K PB, I ran to my local park for
a taster session. And on the way there I got a kinda shocking lesson in quite how much the Hypershell actually does .

You can switch between Transparent (ie no assistance at all) and your
currently selected mode by pressing the single button on the Hypershell X Ultras right motor housing. I had the active setting set to the exoskeletons most punchy Hyper mode max power and an accidental shift to Transparent
while jogging was so jarring I almost face planted into the pavement. The
shot of adrenaline this fired off made it feel like I was tasting gravel for the following 20 minutes.

Perhaps that helped the run, though, because the results I got when upping
the pace I found genuinely surprising. Mildly stunning, even. This was just meant to be a practice run for the big effort, but my Garmin Forerunner 970
was telling me I was hitting kilometer paces well under four minutes. I
cannot ordinarily run that fast.

And so I decided to carry on at a decent effort level, just to see what would happen. 19 minutes and 31 seconds later, I had a new 5K PB one almost a full minute better than Id ever achieved before at that October 2025 Parkrun (20:28).

The odd part was it didnt even feel like I was truly going all-out. And this course was roughly based on my local Parkrun which is rated as the 505th fastest course in the UK by Power of 10 , while my actual non-assisted PB
from October was recorded on the 25th fastest course in the UK, at Worthing.

I was kind of blown away by the difference the Hypershell X Ultra made. Even
if running with it isnt exactly a dream in all respects.

How using an exoskeleton actually feels

You never get away from the sense theres a contraption attached to you, a
1.8kg one in the Hypershell X Ultras case, as it naturally moves about a bit
as you run, and not in lock-step with your own body. Because its not part of it, were a ways off Robocop here, and it weighs heavily around your waist and hips.

But the actual assistance the Hypershell gives you quickly becomes near-invisible, only showing its face when the exoskeletons algorithms take a fraction of a second to adjust to a change of pace. A sudden downhill section can make it feel as though the Hypershell is launching you into the ground,
and there were a lot of these in my initial trail-adjacent park-based route.

The next day or so also came with DOMS (delayed onset muscle soreness) the likes of which I havent felt in a while. It wasnt all that intense, but the assistance the Hypershell gives is naturally going to alter which muscle
groups are actually working the hardest. Your quads and glutes may feel it. Mine did.

My main takeaway, though, was that if I can score a 19:31 5K on a non-flat, awkward course, what could I do on a running track? Thankfully I live near Crystal Palace Athletics Stadium, which any old nobody can book access to,
for as little as 2.45 ($3.26) an hour. So I did just that, with a slot on one Monday afternoon in December 2025.

Test 2: Track run

Was it a triumph? Absolutely not. My second, super-flat attempt ended up with
a 5K time of 20:33, not even beating my previous unassisted best result.

But you know what? I dont blame Hypershell for this one. Its a reminder these exoskeletons are an assist. You still have to turn up with the base stuff for the job. And I didnt.

Its my complete lack of experience on the run track. Its the weekend I spent partially inebriated (and not sleeping enough) before that Monday. Its the
16km run I did the day before on Sunday. Its perhaps that self-consciousness
at wondering what the two hurdlers training at the same time thought of the idiot wearing the exoskeleton on the track.

Heck, it might even be the cortisol built-up having spent 30 minutes hunting through a labyrinthine sports centre trying to work my way to the run track.
I also knew I was cooked the moment my Garmin Forerunner 970 gave me a +1 Performance Condition score part-way through the attempt. As any owner of one of the best Garmin watches knows, Garmins Performance Condition score represents a representation of your run compared to your average fitness
level. A score of +0 is a baseline, +5 or higher is significantly better than average, and -5 or lower is significantly worse. A score of +1 represents minimal improvement.

Is using a Hypershell still cheating if you still need to be able to run well to make the most of it? Of course it is. My 5K PB remains 20:28. But theres still a part of me that wants to have the stones to turn up to my local
Parkrun wearing a Hypershell X Ultra.

======================================================================
Link to news story: https://www.techradar.com/health-fitness/19-minutes-and-31-seconds-later-i-had -a-new-5k-pb-one-almost-a-full-minute-better-than-id-ever-achieved-before-our- writer-tried-a-robotic-exoskeleton-and-ended-up-setting-himself-a-new-running- record
$$
--- SBBSecho 3.28-Linux
* Origin: Capitol City Online (42:17/1)