Feeding the Fade: What 120g of Carbs an Hour Buys You in Watts

Most fuelling advice stops at the gut: how many grams an hour you can tolerate, and how to train your stomach to take more of them. That's a real problem and TriForward has covered it at length elsewhere. It is not this article's problem.

This article is about what happens after the carbohydrate clears your stomach and gets to work: how many watts of usable power it actually buys you back, late in a long ride, when you're fatigued and still have to hold a number. A 2026 randomised crossover trial has now put an unusually precise figure on that question, and the answer is more interesting — and more honest about its limits — than "more carbs, more power".

Why this shows up in your pacing, not just your energy levels

If you've ever felt like your legs "had it" on a long ride despite still having food in your special needs bag, you were probably right that something changed — just not necessarily what you assumed. The instinct is to treat carbohydrate purely as fuel: run low, feel flat, eat, feel better. The research reframes it as something closer to a ceiling: how much sustainable power your body can still produce when it's several hours in and tired.

For an age-grouper riding a 5-hour 70.3 leg or a 6-hour-plus Ironman bike, that ceiling is the number that decides whether you run off the bike or survive it. Small, real differences in watts at hour three or four compound over that much time — which is exactly why a study that measures the effect in watts, not just minutes to exhaustion, is worth reading closely.

What the evidence says

The headline number: carbohydrate defends your critical power

The most directly relevant study put 16 trained cyclists and triathletes (15 men, 1 woman; age 35 ± 9; V̇O₂max 51.8 ± 6.8 mL·kg⁻¹·min⁻¹) through 180 minutes of riding at 95% of gas exchange threshold on three separate occasions — once with water only, once with 60 g·h⁻¹ of a maltodextrin:fructose mix, and once with 120 g·h⁻¹ — then tested critical power immediately afterwards with a 3-minute all-out effort.

Fresh, non-fatigued critical power was 277 ± 27 W. After 3 hours: 236 ± 30 W on water only (a 14.7 ± 7.1% loss), 257 ± 28 W at 60 g·h⁻¹ (a 7.4 ± 4.7% loss), and 266 ± 29 W at 120 g·h⁻¹ (a 4.2 ± 3.7% loss) — a clean, statistically significant dose-response relationship (120 > 60 > water, p < 0.05 throughout).

That is a randomised, counterbalanced crossover design with a clear dose-response curve, which is about as strong as sports nutrition evidence gets. The honest caveat is the sample: 16 athletes, 15 of them men. Treat the exact percentages as representative of trained, mostly male cyclists and triathletes rather than a universal law, and expect some individual variation.

Diminishing returns — the part the 90–120g discourse tends to skip

Here is the detail that gets lost when "just take more carbs" becomes the whole message: the watts don't scale evenly with the grams. Going from zero to 60 g·h⁻¹ bought back roughly 20 W of critical power (p < 0.001, a large effect). Going from 60 to 120 g·h⁻¹ — doubling the intake rate — bought back a further 9 W (p = 0.006, a smaller but still real effect).

That's not a reason to skip the second 60 grams if your gut can handle it — 9 W held for the last hour of a long bike leg is still worth having. But it is a reason to stop expecting the second dose to work as hard as the first. If you've read TriForward's earlier pieces on carb intake rates and gut tolerance, you already know the guidance on how to get to 90–120 g·h⁻¹ in the first place; this is the part those pieces didn't cover — what that last increment is actually buying you, in watts, once you're fatigued enough for it to matter.

The mechanism caveat: carbohydrate isn't an unlimited match book

It would be tidy if more carbohydrate simply meant more matches in the book. The data don't support that story. In the same trial, W′ — the finite, high-intensity energy reserve above critical power — declined regardless of feeding condition; carbohydrate protected the sustainable ceiling, not the size of the anaerobic reserve above it.

More strikingly, a separate biopsy study measuring the same power-duration parameters found that the change in whole-muscle glycogen did not correlate with the change in critical power (r = 0.19) or W′ (r = 0.07) after 2 hours of heavy-intensity cycling — the authors concluded explicitly that changes in these parameters "are not primarily determined by changes in muscle glycogen." That's a limited, single-study finding, but it is a direct challenge to the "empty tank" mental model: your legs can fade even when there's still glycogen sitting in the muscle, and topping up glycogen doesn't fully explain why fuelling helps.

The more plausible mechanism, per the wider evidence base, is that glycogen depletion isn't uniform across muscle fibres. The fibres that empty out first tend to be the ones with more mitochondria, so the muscle is forced to lean on less efficient fibres to hold the same power — which looks, from the outside, exactly like fading legs and a rising oxygen cost for no change in speed.

Carbohydrate feeding blunts that process rather than eliminating it, which is why 100 g·h⁻¹ in a group of elite and professional cyclists across a 4-hour intermittent protocol still left them 10% down on 6-minute time-trial power and 6% down on peak power. Fuelling is the single biggest lever an age-grouper has over their own late-ride watts — but it is a partial fix, not a cure, even at high intakes and even in elite athletes.

It isn't only about your top-end power

The dose-response finding above used a short, hard 3-minute effort to detect critical power. A separate study of 12 trained cyclists and triathletes riding 150 minutes with and without carbohydrate found the same protective pattern at gentler intensities: power at the first ventilatory threshold fell 3 ± 2% with carbohydrate versus 6 ± 4% without it (p = 0.019), and 5-minute time-trial power fell 4 ± 3% with carbohydrate versus 10 ± 10% without — though the authors note that second, between-condition comparison did not reach statistical significance (p = 0.186), so treat the threshold result as the firmer of the two. That matters because most of an Ironman or 70.3 bike leg is spent near that lower threshold, not at 3-minute-effort intensity — so the protection carbohydrate offers isn't confined to a short, unrepresentative burst.

The near-all-male evidence base, and one study that isn't

Nearly every number above comes from samples that are 80–100% male, which is a real limitation of this literature, not a footnote. One study offers a useful counterpoint: 9 well-trained female cyclists, glycogen-depleted then fed either a high (≥9 g·kg⁻¹) or low (≤1 g·kg⁻¹) carbohydrate diet, showed power at the first ventilatory threshold of 152 ± 28 W on the high-carbohydrate diet versus 133 ± 24 W on the low one (a 19 W gap, p = 0.011), alongside reduced gross efficiency and a shift in muscle recruitment toward higher-threshold motor units on the low-carbohydrate diet. The sample is small (grade: limited), but the direction agrees with the larger male-dominated studies, and it's the only piece of this evidence base built specifically on female physiology — worth knowing given how thin that coverage is everywhere else in this area.

The framework for why glycogen availability changes efficiency and recruitment this way, rather than simply running the tank dry, comes from the broader "fuel for the work required" literature on carbohydrate periodisation — a mechanistic account, not a training prescription; we'll come back to what happens when people try to turn that mechanism into a training method in a separate article on train-low protocols.

What to actually do with this

None of this changes the intake-rate guidance TriForward has already published — it changes why you should bother following it. The practical read for a long-course bike leg:

• Treat carbohydrate as a power-preservation tool for the back half of the ride, not just as a way to avoid bonking. If your goal includes holding a specific bike power target into hour three or four, your fuelling plan is part of your pacing plan.
• Fuel at the top of what you can comfortably tolerate, because the dose-response effect is real up to at least 120 g·h⁻¹ — but don't expect the jump from 60 to 120 g·h⁻¹ to feel as dramatic as the jump from 0 to 60. Roughly two-thirds of the benefit shown in this trial came from the first 60 grams.
• Don't abandon pacing discipline because you're well fuelled. Carbohydrate slowed the decline in every condition in this trial — it didn't stop it. Even 100 g·h⁻¹ left elite riders down 10% on short, hard efforts after 4 hours in a separate study; a well-fuelled hour four is still a fatigued hour four.
• If your race includes any hard, short efforts late on — a surge past a competitor, a steep pinch, a final push into T2 — remember that W′ depletes regardless of fuelling. Carbohydrate defends your sustainable ceiling, not your matches.

Common mistakes

• Chasing the highest tolerable carbohydrate number as an end in itself, without ever rehearsing it. 120 g·h⁻¹ is a meaningful, evidenced target, but it also requires deliberate gut training over weeks; showing up to race day with a new fuelling rate you've never used in training is a GI-distress risk, not a watts strategy.
• Assuming a flat, empty-tank model of fatigue — "I ate enough, so I shouldn't be fading." The evidence says otherwise: critical power fell in every fuelling condition in the dose-response trial, including at 120 g·h⁻¹. Fading despite good fuelling is normal, not a sign you did something wrong.
• Ignoring pacing because "I'm fuelled." Good fuelling buys you a higher ceiling; it doesn't remove the ceiling. Riders who go out too hard early still fade hard late, just from a slightly higher starting point.
• Extrapolating this data to a female-specific race plan without caveats. The core dose-response numbers above come from an essentially male sample; the only female-specific study in this evidence base is small (n = 9) and used a different protocol. The direction agrees, but treat the exact wattage figures as provisional for women until better-powered female-specific trials exist.

How to apply this this week

You don't need a lab to act on any of this. On your next long ride of 2.5 hours or more (the kind you're already doing most weekends if you're building toward a 70.3 or full-distance race):

• Set a specific carbohydrate target for the ride — whatever your currently tolerated rate is — and hit it on a schedule (a timer or lap alert works better than "when I remember").
• In the final 45–60 minutes, hold a genuine effort check: pick a short, honest effort (a 3-minute push, or simply your normal race-pace number) and see how it compares with fresh numbers from earlier in the week. That gap is your personal, real-world version of the critical-power decline this study measured.
• If you're not yet comfortable above 60–90 g·h⁻¹, don't jump to 120 for a race. Add 10–15 g·h⁻¹ every couple of long rides and note GI symptoms; this is gut training, and it takes weeks, not one session.
• Log how your race-relevant power or pace at the end of the ride compares with the start, across a few long rides. If the gap is shrinking as your fuelling rate climbs, that's your own dose-response curve forming.

References

• Norte et al., "Carbohydrate ingestion attenuates the decline in critical power," Scandinavian Journal of Medicine & Science in Sports, 2026;36(6):e70326
• Clark et al., American Journal of Physiology-Regulatory, Integrative and Comparative Physiology, 2019;317(1):R59–R67
• Clark et al., Journal of Applied Physiology, 2019
• Dudley-Rode, Zinn, Plews, Charoensap & Maunder, European Journal of Applied Physiology, 2025;125(5):1349–1359
• Main et al., European Journal of Applied Physiology, 2026
• Ørtenblad, Zachariassen, Nielsen & Gejl, European Journal of Applied Physiology, 2024;124(7):2193–2205
• Impey et al., "Fuel for the Work Required," Sports Medicine, 2018;48(5):1031–1048