Does Caffeine Improve Exercise Performance? What the Evidence Shows

Most “performance supplements” are marketing wrapped around a placebo. Caffeine is the rare exception — the ergogenic aid with the deepest evidence base in all of sports science.

The short answer is yes, and the effect is real across a wider range of tasks than you’d expect. A 2020 umbrella review in the British Journal of Sports Medicine (Grgic and colleagues) synthesized 21 separate meta-analyses and found caffeine ergogenic for aerobic endurance, muscle strength, muscle endurance, power, jumping, and exercise speed. The catch isn’t whether it works — it’s that the effect is larger for endurance than for pure strength, the honest dose is lower than most people take, and the size of the benefit varies a lot between individuals. Here’s the version worth acting on.

Does caffeine actually improve performance — or just feel like it?

It genuinely improves it, and the evidence is unusually broad. In that Grgic 2020 umbrella review, every one of the pooled meta-analyses pointed the same direction: caffeine improved performance across aerobic endurance, strength, muscular endurance, and power. The authors graded the overall quality of evidence as moderate — which, in a field full of underpowered single studies, is about as strong as ergogenic evidence gets.

The most important nuance in that review: the magnitude of the effect is generally greater for aerobic than for anaerobic exercise. If you’re a cyclist, runner, or rower, caffeine is one of the highest-leverage legal interventions you have. If you’re chasing a one-rep max, it still helps — just less.

On the strength side, the numbers stay honest. A 2025 meta-analysis in Frontiers in Nutrition (Xiao and colleagues, 12 studies, 230 participants) found caffeine improved movement velocity during resistance exercise with a standardized effect of 0.42 and power output with an effect of 0.21 — small but real. For maximal strength (one-rep max), prior meta-analyses they cite put the effect even smaller, around 0.17 to 0.20 — a nudge, not a transformation. Caffeine helps you move a load faster more than it helps you move a heavier one.

How much, and when?

The consensus dose has held steady for years. A 2025 review in Sports Medicine (Silva, Del Coso, and Pickering) states the current recommendation plainly: ~3-6 mg per kg of body mass, about 1 hour before exercise. For an 80 kg athlete that’s roughly 240-480 mg — one to three strong coffees’ worth, ideally in a measured form so you actually know the dose.

The useful surprise is how much you get from the low end. A 2026 randomized trial in the Scandinavian Journal of Medicine & Science in Sports (Montalvo-Alonso and colleagues, 94 resistance-trained men and women) used just 3 mg/kg and still improved both strength and muscular endurance across the board, with movement velocity rising 4-12% in the most responsive participants. Their conclusion called low-dose caffeine “an effective and safe ergogenic aid… regardless of sex.” Timing matters as much as amount: nearly every study in this literature dosed 60 minutes before exercise, because that’s when blood caffeine peaks. In the Xiao meta-analysis, all but one trial used that one-hour window.

Why do some people barely respond?

This is the caveat the supplement ads skip. The response to caffeine is highly individual, and part of that variation is genetic. In the Montalvo-Alonso trial, participants were split by their CYP1A2 genotype — the gene that governs how fast you metabolize caffeine. Velocity gains were 4-12% in “AA” (fast-metabolizer) carriers, 3-9% in “AC” carriers, and 4% or less in “CC” carriers. Same dose, meaningfully different payoff.

Habitual intake matters too. In the Xiao meta-analysis, low-habitual-caffeine consumers saw far bigger velocity gains than heavy users (an effect of 0.87 versus 0.21) — evidence that day-to-day tolerance blunts the acute boost, even if it doesn’t erase it. The practical implication: if you drink caffeine all day, the pre-workout dose is competing against your baseline. Cycling down your habitual intake before an event is one way high responders and non-responders alike sharpen the effect.

None of this makes caffeine a coin flip. The average effect is positive and well-replicated. But “it works on average” and “it will work for you at this dose” are different claims — and the only way to know your number is to test it in training, never on race day.

What’s the catch?

Two real costs. First, the side effects climb with the dose while the benefit plateaus. Since 3 mg/kg already delivers most of the performance gain, pushing to 6+ mg/kg mainly buys you jitters, elevated heart rate, and — the big one — wrecked sleep. The same Sports Medicine review that endorses caffeine spends most of its length on this tension: caffeine taken for an evening event can measurably degrade that night’s sleep quality, which can quietly cost you the next day’s session.

Second, the gut. A 2025 study in Physiological Reports (Davison and colleagues) found that 3 mg/kg of caffeine before cycling exacerbated exercise-induced intestinal cell damage, especially in sensitive individuals — a plausible mechanism behind the GI distress some endurance athletes get from a pre-race hit. If caffeine reliably upsets your stomach mid-effort, that’s not in your head, and no performance gain is worth blowing up a long event.

The takeaway

Caffeine is the closest thing to a proven performance enhancer you can buy legally. Aim for 3-6 mg/kg — start at the low end — about an hour before you train, and test it in practice before it matters. Expect a bigger return on endurance than on brute strength, and don’t be alarmed if your response is modest: genetics and tolerance make it deeply personal. Then respect the ceiling. More caffeine doesn’t mean more performance past a point — it means more side effects and worse sleep.

For the harder end of your training, it pairs naturally with a well-built aerobic base — see Zone 2 training: what the science actually says. And caffeine isn’t the only compound with real evidence behind it: the same skeptical lens applies to creatine for cognitive function and to how hydration shapes cognitive performance. For the full system, see our performance optimization for high-performers work.