How muscle actually grows: tension, signaling, repair

Nothing about you gets bigger while you're in the gym. The set is a message. The building happens afterwards, over the next day or two, while you're at your desk or asleep.

And the message your muscle listens to isn't the burn, and it isn't how sore you are the next morning. It's tension: how hard the fibers had to pull against the load. Get that part right and most of the rest follows. Chase the burn instead and you can leave yourself wrecked without giving the muscle much reason to grow.

Three things have to line up. The muscle has to feel real force. That force has to become a chemical signal. And the signal has to be fed and repeated, for weeks.

Your muscle has force sensors in its walls

How does a muscle "feel" anything? Through mechanotransduction, which is a long word for turning a physical force into a chemical message.

When a fiber contracts hard against a load, the pull lands on structures called costameres. They're anchor points that stitch the machinery inside the fiber to its outer membrane, and they work like strain gauges. High force trips them, and they set off a cascade that switches on the cell's master growth regulator, mTOR (mTORC1, if you want the exact one).

Once mTOR is on, it tells the fiber's ribosomes to start building more contractile protein. That's the construction order. Pull hard on the sensors and they flip the switch that says build. No hard pull, no switch, no order.

The reviews have converged on this over the past few years. A review published in 2025 says it in one line: mechanical tension is the primary stimulus behind the growth you get from resistance training.¹ Everything else in this article sits downstream of that.

The build order stays open for about two days

After a hard session, the rate at which your muscle assembles new protein, called muscle protein synthesis, climbs and stays up for a day or two.

We've had solid numbers on this since 1997, from a study in the American Journal of Physiology that measured eight untrained people at rest and then after a hard leg session. Synthesis was about 112 percent above rest three hours later, still up around 65 percent at 24 hours, and about 34 percent elevated at 48 hours.² Nobody ate during the measurements, so net protein balance, what you build minus what you break down, stayed negative the whole time. Training moved it a long way toward the building side without ever tipping it over.

That last part is the whole game. Your body is building and breaking down muscle protein all day, every day. Growth isn't about building fast. It's about building a bit more than you tear down across the day, and keeping it tipped that way long enough for the difference to show. Lifting raises the building side. Eating enough protein is what gets the day over the line.

One session on its own does almost nothing

The window closes, and one window's worth of net gain is tiny. You can't feel it and you can't see it. Real growth is a stack of them, which is why showing up often beats any single heroic workout.

Picture a sawtooth. Each hard session bumps your muscle protein up a notch over the next day or two, then it drifts back toward baseline until the next session bumps it again. Train the same muscle before it settles all the way back, week after week, and the notches add up into size you can see.

So the week matters more than the workout. The weekly number has its own post, how many sets per week your muscles actually need, and the short version is that hard sets per week is the dial, not what you pull off in any one session.

What about the pump and the soreness?

This is the part of the story that got rewritten.

For years the working model listed three drivers of growth: mechanical tension, metabolic stress (the burn and the pump, from byproducts piling up), and muscle damage (the microscopic tearing that leaves you sore). That three-part model was published in 2010, it was useful, and it's still quoted everywhere.³

Since then the field has walked it back toward tension being the one that counts. The other two look more like passengers than drivers.

Damage is the clearest case. When researchers tracked untrained people through their first weeks of lifting, the early spike in protein synthesis, at the point where damage and soreness were highest, had no relationship to how much anyone eventually grew. That early synthesis was busy repairing damage, not adding muscle. Only later, once the damage settled down, did synthesis line up tightly with real growth, at a correlation around 0.9.⁴

Soreness tells you that you did something your body wasn't used to. It doesn't tell you that you grew. Early on, the repair crew and the building crew are the same crew.

Past a certain size, a fiber has to borrow nuclei

Up to a point, a fiber grows by cramming more protein into the space it already has. But its nuclei, the control centers that run protein production, can each only manage so much territory. To keep growing past that limit, the fiber needs more nuclei, and it gets them from satellite cells, a reserve of muscle stem cells parked just outside the fiber.

One 16-week training study makes this hard to ignore. Researchers trained 66 adults' knee extensors and then sorted them by results. The extreme responders grew their fibers by about 58 percent. Modest responders got about 28 percent. Non-responders essentially didn't grow at all, on the same program. The split tracked satellite cells almost perfectly: the extreme group expanded its satellite cell pool by roughly 117 percent and added nuclei to its fibers, and the non-responders did neither.⁵

How readily your satellite cells show up isn't something you get to choose. The tension that asks them to is.

Tension has two dials, and you need both

The first is load. Heavier weight, more force per rep, no subtlety to it.

The second is effort. As a set gets close to failure, your body recruits its biggest, most growth-prone fibers and drives them under high tension even when the weight is only moderate. That's why a hard set of 12 near failure can grow muscle about as well as a heavy set of 6. How close to failure you need to get is an argument of its own, but the working answer is 0 to 3 reps short.

Metabolic stress and damage still come along for the ride. They just don't seem to add much by themselves. They're what hard training produces, not what makes it work.

What the lab work can't tell you

Be careful with anyone who makes this sound settled.

Most of the detailed signaling work, the mTOR switches and the synthesis spikes, is measured in the hours and days after a single session, and those short-term signals predict long-term growth badly. Studies keep finding that a bigger acute synthesis response doesn't reliably mark out who grows more over months.

A lot of the mechanistic detail also comes from animals, cell cultures, or small groups of young untrained men in short studies. How cleanly that carries over to a 55-year-old woman ten years into training is anyone's guess. And the spread between people is huge: in that 66-person study, some gained close to 60 percent while others on the identical program gained nothing, and nobody can fully explain why.

Mechanisms tell you which levers to pull. They don't promise the same result when you pull them.

What that looks like in your week

• Make something harder every week or two. A bit more weight, an extra rep, or one rep closer to failure. If nothing gets harder, the signal stays flat.
• Take your hard sets close to failure, 0 to 3 reps short. A set you could have doubled barely counts.
• Come back to each muscle in a day or two. The synthesis window runs about 24 to 48 hours, so training a muscle two or three times a week beats one weekly blowout.
• Eat enough protein, spread through the day, so there's raw material when the build order goes out. Then sleep, because that's when the growth you signaled for gets built.
• Judge it by your logbook, not by how sore you were. Weights and reps trending up over months is the scoreboard.

Common questions

How long does it take to build noticeable muscle? For most people, visible change takes a couple of months of consistent, progressive training, and often longer past your 40s. Early strength gains arrive faster because your nervous system gets more efficient before the muscle itself is much bigger.

Do I have to lift heavy to grow, or do light weights work? Both work if the set is taken close to failure. Heavy loads make tension through the weight. Lighter loads make it through fatigue as you near failure. What doesn't grow much is light weight stopped well short of hard.

Is muscle soreness necessary for growth? No. Soreness reflects unfamiliar or damaging work, not growth, and it fades as you get used to a movement. You can keep growing for years with very little soreness, and a session that leaves you sore didn't necessarily build more.

Why do some people grow so much faster than others? A big part is how readily their satellite cells add nuclei to fibers, which varies a lot between people, plus genetics, age, sleep, and protein. It's why the same program produces dramatic gains in one person and modest ones in another. Muscle is also far more than a mover, it's an endocrine and metabolic organ, which I cover in what muscle does beyond moving you.

Knowing the mechanism is the easy half

A notebook, a pair of dumbbells, and the discipline to nudge the weight up every couple of weeks will grow muscle for years. The biology doesn't care who's watching.

What a coach changes is the execution, not the mechanism. Tension only shows up when the load is genuinely hard and the set really is close to failure, and most people are quietly miscalibrated on both. They stop three reps early, or they grind junk reps that pile on fatigue and nothing else. A coach watching live sees your rep speed drop, knows you've got two hard reps left, and says so while the weight is still moving. Then next week's load comes from what your last few sessions actually produced instead of a guess. The how it works page shows how that runs.

Founding membership is $49 your first month, then $299/mo locked for life, and there are 25 of those spots. If you'd rather run the experiment on yourself, you've got the levers.

Sources

1. Van Every DW, Lees MJ, Wilson B, Nippard J, Phillips SM, 2025. Load-induced human skeletal muscle hypertrophy: mechanisms, myths, and misconceptions. Journal of Sport and Health Science. ↩

2. Phillips SM, Tipton KD, Aarsland A, Wolf SE, Wolfe RR, 1997. Mixed muscle protein synthesis and breakdown after resistance exercise in humans. American Journal of Physiology. ↩

3. Schoenfeld BJ, 2010. The mechanisms of muscle hypertrophy and their application to resistance training. Journal of Strength and Conditioning Research. ↩

4. Damas F, Phillips SM, Libardi CA, Vechin FC, Lixandrão ME, Jannig PR, Costa LA, Bacurau AV, Snijders T, Parise G, Tricoli V, Roschel H, Ugrinowitsch C, 2016. Resistance training-induced changes in integrated myofibrillar protein synthesis are related to hypertrophy only after attenuation of muscle damage. Journal of Physiology. ↩

5. Petrella JK, Kim JS, Mayhew DL, Cross JM, Bamman MM, 2008. Potent myofiber hypertrophy during resistance training in humans is associated with satellite cell-mediated myonuclear addition: a cluster analysis. Journal of Applied Physiology. ↩