Fueling at 90 to 120 Grams of Carbs per Hour

Keep a tolerated carbohydrate intake when it suits the session. For demanding efforts beyond 2.5 to 3 hours, working toward 90 grams per hour can be useful. Treat 120 grams per hour as a rehearsed experiment for a selected long ride.

A June 2026 randomized crossover trial found that 120 g per hour preserved more late-ride power than 60 g after three hours of moderate cycling.¹ The 9-watt difference in 16 trained cyclists and triathletes gives long-distance riders a reason to test higher intake. A separate October 2026 study of eight trained men randomized 90, 120, and 150 g per hour and found no significant difference in exercise capacity among them.⁶

01Choose a target for the session

The Australian Institute of Sport recommends 30 to 60 g per hour for sessions of 1 to 2.5 hours and up to 90 g beyond 2.5 to 3 hours. It calls for mixed carbohydrate sources at higher intakes and practice that accounts for tolerance and hydration.² Its standard guidance does not extend to 120 g.

Duration and effort help set fuel demand. An easy four-hour ride and four hours with sustained threshold climbs can justify different targets.

Tolerance sets the ceiling you can use today. If 60 g per hour still brings bloating or nausea, return to the last intake you tolerated and review the plan before increasing it. Feeding opportunities set the ceiling the course allows. Bottle cages, aid station spacing, and technical descents can cap intake before the gut does.

Stay at your tolerated intake for easy long days or sessions that do not justify more fuel. Work toward 90 g per hour for demanding efforts beyond 2.5 to 3 hours once 60 g feels routine. Test 120 g on a selected long ride after repeated comfortable rehearsals at 90 g. The daily carbohydrate around these sessions is covered in endurance athlete fueling.

02Higher intake can preserve late ride power

Norte and colleagues tested 16 trained cyclists and triathletes, 15 men and one woman, in a randomized crossover trial.¹ After carbohydrate loading and a standardized breakfast, participants rode for three hours at 95% of gas exchange threshold, a moderate intensity, while taking water, 60 g per hour, or 120 g per hour.

A 3-minute all-out test at the end of each ride estimated critical power, the boundary between heavy and severe exercise intensity. Riders also did the test fresh.

Group means from 16 trained cyclists and triathletes, 15 men and one woman, after carbohydrate loading and breakfast. Critical power was estimated with a three-minute all-out test. The fresh reference was tested on a separate visit. There was no 90 g per hour condition or race-time outcome.¹

Sixty grams per hour preserved about 20 W compared with water. The 120 g condition preserved another 9 W, but remained 11 W below the fresh value. More carbohydrate reduced the decline without erasing it.

There was no 90 g arm, so the trial cannot say how much of the 9 W a 90 g plan would capture. It studied cycling and measured no race time. The difference cannot be converted into minutes saved. GI symptoms were mostly mild. Cumulative symptoms were higher with 120 g than with water, and the 60 g and 120 g conditions did not differ significantly. The carbohydrate drinks came in a fixed volume while riders drank water as they wished, so the trial says nothing about raising fluid or sodium to carry a bigger dose.

A longer ride produced a similar gain on a different test. Mazza and colleagues ran a randomized, double-blind crossover trial in 15 trained male cyclists and triathletes.⁷ After a carbohydrate-rich day, each rider completed two 240-minute intermittent rides at 60% of their fresh 15-minute time-trial power, with high-intensity intervals, taking 60 or 100 g per hour from drinks and gels at a 1:0.8 glucose to fructose ratio. A 15-minute time trial followed each ride, and mean power in that test was 9 W higher after 100 g than after 60 g (P = .028).

The two 9 W results measure different things. Norte estimated critical power from a three-minute all-out test after three hours. Mazza measured average power across a fixed 15-minute time trial after four. Neither number can be added to the other or turned into minutes saved. The higher intake improved late-ride power without a detected difference in post-ride muscle glycogen. Mazza had no 90 g or 120 g arm.

03Does 150 grams per hour improve performance

Riders burned more of the drink at 150 g per hour than at 120 g. They did not last significantly longer.

Martyn and colleagues studied eight trained men who cycled for three hours and then completed an exercise-capacity test at 150% of lactate threshold.⁶ A 10 g per hour maltodextrin visit came first as familiarization. Only the later 90, 120, and 150 g per hour trials were randomized. Each used a 1:0.8 maltodextrin to fructose drink enriched with labeled glucose and fructose, so the researchers could trace how much of the ingested carbohydrate was oxidized.

Oxidation of the ingested carbohydrate rose with each step, from 1.23 g per minute at 90 g to 1.48 at 120 g and 1.62 at 150 g. Whole-body carbohydrate oxidation was 2.6 g per minute at both 120 and 150 g. Total exercise duration did not differ significantly among the three doses. The extra 30 g at the top raised the share of fuel coming from the drink without raising total carbohydrate use or producing a detected capacity advantage.

Left, Mazza studied 15 trained men after four hours of intermittent cycling. Right, Martyn studied eight trained men with three hours of cycling followed by a capacity test. Blue values on the right measure oxidation of ingested carbohydrate, not power or exercise duration. These are separate studies, and the nonsignificant capacity result does not establish equivalence.⁷⁶

The eight-person trial was too small to establish that the doses perform equally. The authors modeled an estimate of 107 g per hour for exercise capacity in this protocol. The 107 g rate was not directly tested, and the figure describes one laboratory test rather than a target for every athlete. The studies below differ in duration, dose, and endpoint, so none of them ranks 90, 100, 120, and 150 g on one scale.

Use them to choose what to rehearse. If you already finish long rides cleanly at 90 or 100 g per hour, a rehearsal at 120 g is a reasonable experiment. Going past 120 g raised oxidation of the drink without a detected capacity gain in the one study that tested it. Count the grams you actually consumed, and let symptoms and the final hour decide whether a higher rate stays in the plan.

04Combine glucose and fructose

Ingestion is what you swallow. Absorption is what crosses the intestinal wall into the blood. Oxidation is carbohydrate used as fuel. Isotope-labeled carbohydrate lets researchers estimate how much came from the ingested source.³⁴

Glucose crosses the intestine through the SGLT1 transporter and fructose through GLUT5. Maltodextrin is a chain of glucose units, so it supplies glucose and uses the same route. A product built from maltodextrin and fructose uses both routes, which is why high-intake guidance calls for mixed sources.² The transport detail lives in the multiple transportable carbohydrates glossary entry.

A 2:1 product carries twice as much glucose as fructose. A 1:0.8 product moves closer to equal parts. At 90 g per hour, a 2:1 mix supplies about 60 g glucose and 30 g fructose. At 120 g per hour, a 1:0.8 mix supplies about 66.7 g glucose and 53.3 g fructose. The total rises by a third, and the fructose rises by more than three quarters.

Podlogar and colleagues compared exactly those two conditions in 11 highly trained men who rode for three hours at 95% of gas exchange threshold after an overnight fast.³ Oxidation of the ingested carbohydrate between minutes 120 and 180 was higher with 120 g at 1:0.8, at 1.51 versus 1.29 g per minute. The extra carbohydrate did not spare more of the body's own carbohydrate stores. Dose and ratio changed together, so the study cannot assign the gain to either one, and it measured no performance outcome.

Hearris and colleagues tested format at 120 g per hour.⁴ Nine trained men loaded at 8 g/kg the day before, ate a 2 g/kg breakfast, and rode for 180 minutes while taking the carbohydrate as a drink, a gel, a jelly chew, or a combination. Oxidation of the ingested carbohydrate and exercise capacity did not differ significantly between formats. Rehearse any product switch in training.

05Count every gel and bottle

Use 30 g of carbohydrate as a convenient counting unit. Check your labels, because gel and chew serving sizes vary.

One-hour carbohydrate totals, assuming the drink and gels are consumed. Bottle volume and product choice are separate decisions. Check the carbohydrate amount on each label.

A bottle containing 60 g of carbohydrate and finished within the hour changes the gel count.

These tables show carbohydrate consumed per hour. They do not set a bottle volume, require particular products, or prescribe a schedule that suits every course. Bottle volume belongs to the fluid plan.

Count what you consumed rather than what you carried. If the 60 g bottle comes home a third full, you drank about 40 g from it, and a 120 g plan delivered about 100 g. Estimate the leftover and subtract it in your log. A gel dropped on a descent counts as zero. The race-week nutrition plan turns these counts into a packing list.

06Rehearse before increasing intake

Gut tolerance responds to practice, and the clearest evidence describes that response at 90 g. Costa and colleagues gave 25 runners two weeks of repeated gut challenges during exercise, with a test before and after the intervention that involved a 30 g gel-disc every 20 minutes during a two-hour run, or 90 g per hour.⁵ The intervention used carbohydrate supplements, carbohydrate foods, or placebo during repeated exercise challenges. GI symptoms improved in the carbohydrate groups, while oxidation rates did not differ between groups and trials. After the two weeks, distance covered in a one-hour test run improved 5.2% in the supplement group and 4.3% in the food group, while it fell 2.1% in the placebo group.⁵ That supports practice for symptom tolerance and running performance at 90 g in this protocol. It sets no timeline for reaching 120 g and shows nothing about intestinal transporter expression.

Record actual carbohydrate consumed after subtracting leftovers, fluid consumed, temperature and humidity, intensity by power, pace, or heart rate, any GI symptoms and when they started, the timing of each serving, and how the final hour felt and performed.

Hold the current intake until you have completed it in two or three long sessions with no more than mild symptoms and a final hour that matched the effort plan. Then step up by 10 to 15 g per hour. Step back to the last clean intake if symptoms reach moderate, if you skip servings to cope, or if the final hour gets worse. This progression is a practical coaching approach rather than a trial-tested protocol.

Ketone drinks are sometimes pitched as a way to stretch carbohydrate further. In a 2026 trial of eight cyclists taking 120 g per hour, adding a ketone monoester lowered blood glucose and oxidation of the ingested carbohydrate, with no significant change in exercise capacity. Ketone supplements for endurance covers that trial and the separate evidence for ketones taken after training.

Running needs its own rehearsal. Norte, Podlogar, and Hearris all studied cycling. Their results cannot establish the intake a runner will tolerate. A runner should not assume that 90 g on the bike means 90 g at marathon pace. Gut training for race nutrition covers practical rehearsal and symptom monitoring.

07Match fluids to conditions

Fluid and carbohydrate targets come from different inputs. Carbohydrate follows duration, effort, and tolerance. Fluid planning accounts for sweat losses, conditions, drinking opportunities, and tolerance. Raising carbohydrate from 90 g to 120 g per hour does not by itself create a need for more water or more salt, and none of the trials above tested that.

Concentration is where the two plans meet. On a cool day with low sweat losses, moving part of the dose into gels and drinking plain water to a measured plan lets you adjust fuel without requiring an extra bottle of sports drink.² Fluid needs may rise with heat and sweat loss. Build the drinking plan from a sweat rate test.

If high intake brings GI symptoms, look first at the carbohydrate dose, the mix, and how often you have rehearsed it. Extra sodium is not a tested fix for those symptoms.

Before your next long ride, write the hourly target, servings, and fluid plan. Count what you consumed and use that record, your symptoms, and late-ride performance to choose the next target.