How Kilian Jornet Trains (and Why He Counts Metres, Not Kilometres)

Ask most runners how their week went and they'll tell you a number of kilometres. Ask Kilian Jornet and the unit is metres climbed. That sounds like a quirk of mountain running, but it's actually the clearest window into how he trains: for a mountain athlete, the load that matters is the one measured upwards, and almost everything else in his method follows from taking that seriously. His 2026 season is also the best argument yet that a huge aerobic engine and a durable body are two different projects.

Vertical is the unit

Flat distance tells you almost nothing about a mountain session. An hour on a road and an hour going up a 20 percent slope are the same duration and completely different training, and pace as a metric collapses entirely the moment the gradient gets steep enough that everyone is power-hiking. What survives is vertical gain per hour, which is why Jornet's own logs are dominated by metres of ascent rather than distance covered.

That gives you a usable intensity anchor too. Your rate of climb, in vertical metres per hour, behaves like pace does on the flat: repeatable, measurable, and directly comparable week to week. It's exactly what our ski mountaineering and uphill calculator is built around, so if you train on real gradients, this is the honest way to set your bands:

Do a hard, sustained climb, feed in the vertical you gained and how long it took, and you get zones expressed in the unit the mountain actually charges you in.

Three or four months of easy, every year

The structural feature of Jornet's year is a long, unglamorous aerobic block. In his own words:

I always do a long aerobic base period for 3-4 months with moderate to long distance journeys. That's a must for me to build the base fitness to then do the specific training.

Specific work, the race-pace and threshold sessions, comes late, roughly six or seven weeks out from a target. Everything before that is volume at an intensity he could sustain all day. His reported intensity distribution sits somewhere around 80 to 90 percent easy, which puts him in the same territory as polarized training even though he arrived there by feel rather than by reading Seiler.

The transferable part is the ordering, not the hours. Most amateurs invert it: they start sharpening in January, stay sharp all year, and never build the base that sharpening is supposed to sit on top of. Base first, specificity last, and be patient about it. Our guide to aerobic base training is the same argument at amateur scale.

Winter is for skis

From roughly November to May, Jornet's training is dominated by ski mountaineering rather than running. He races it seriously, and the weekly vertical in that block is enormous.

This is the smartest structural idea in his whole method, and the one almost nobody copies. Skinning uphill produces the same cardiovascular load as running uphill with a fraction of the eccentric damage, because there's no downhill pounding. He gets a winter of huge aerobic volume and gives his tendons and connective tissue a seasonal break from impact. If you live anywhere with snow, ski touring is not cross-training as a consolation prize, it's the closest thing to free aerobic volume that exists. Our ski touring zones page covers the nordic side of the same idea.

Self-coached, by feel

Jornet has no personal coach. He consults sports scientists, but the plan is his own, and he's explicit that the day-to-day steering is done by sensation rather than by a prescribed file:

I train based on feel, and use my heart rate to see if I have a cardiac response that might suggest if I'm tired or fresh.

Note what heart rate is doing there. It isn't setting the session. It's a check on whether the athlete is fresh, which is the more useful job for HR in a sport where terrain makes pace meaningless. A morning resting heart rate that has drifted up, or an effort that produces an oddly low heart rate on a climb that should be easy, tells you more about your readiness than any single workout number.

He also says the quiet part out loud, which for a professional is rarer than it should be:

If you don't have pleasure when you train, you will never improve!

The number nobody has actually measured

You will read that Jornet's VO2max is somewhere around 90. Treat that with suspicion. It gets repeated everywhere, but it traces back to informal estimates and back-calculations from race performances rather than to a published lab test, and Jornet himself has said he has not done the standard protocol. He is obviously an extraordinary aerobic engine. The specific number is folklore, and it's worth knowing that, because the same is true of a lot of numbers attached to famous athletes.

If you want your own figure, an estimate from a field test is at least honest about being an estimate.

2026: the engine was never the problem

Jornet built 2026 as a compressed, ambitious season: Western States in June, Sierre-Zinal in August, UTMB at the end of August, chasing a record fifth UTMB title.

It has not gone to plan. At Zegama in May, a race he has won a record eleven times, he picked up a leg problem and faded to 43rd. He took time off running, started Western States anyway, ran with the lead group early, and pulled out at the Dusty Corners aid station around mile 38 as the knee got worse. His team's statement was plainly put: when he reached Dusty Corners, continuing no longer felt like the right decision. Sierre-Zinal and UTMB are still ahead of him as I write this, so the season is unfinished.

There's a pattern here worth naming, because we hit it in the Ingebrigtsen post too. Two of the most physiologically gifted endurance athletes of this era, training in completely different systems, both lost seasons to a joint or a tendon rather than to a lack of fitness. Aerobic capacity and structural durability are separate accounts. Almost all popular training advice, including most of what we write about zones, is about the first one. The second one is quieter, slower to build, and does not show up in any calculator.

What you can actually steal from it

• Measure the load your sport actually applies. For mountain athletes that's vertical gain, not distance. Set your bands from rate of climb rather than pretending flat pace means something on a 20 percent slope.
• Build base before sharpness, every year. Three to four months of genuinely easy work, then six or seven weeks of specific preparation. Not the reverse, and not both at once all year.
• Use winter for low-impact volume. Ski touring, or anything else that loads the engine without pounding the tissue, is the cheapest aerobic gain available.
• Let heart rate report on freshness, not dictate pace. In hilly terrain it's a much better fatigue sensor than a target.
• Treat famous physiological numbers as gossip until you see the protocol.
• Enjoy it, or you won't keep doing it. He's not being sentimental. Consistency over years is the actual mechanism, and nobody is consistent at something they hate.

More in this series: Jakob Ingebrigtsen, Eliud Kipchoge, Tadej Pogačar and Jonas Vingegaard.

Sources

Jornet's base-period and train-by-feel quotes are from Coros; the "pleasure" quote is from Suunto. His reported intensity distribution and training volumes are compiled from secondary reporting (Sweat Elite, Runner's Tribe) rather than a published training log, so treat the exact figures as approximate. The status of the widely repeated VO2max figure is examined by Positive Nation. His 2026 schedule, the Zegama result and the Western States DNF are reported by Run247, Run247 and Run247.

Cover photo by Christian David, CC BY-SA 4.0, via Wikimedia Commons.