Two federal numbers describe how much protein you should eat. They do not agree.
- The RDA — 0.8 g per kg of body weight. Unchanged since 2005.
- The 2026 Dietary Guidelines — 1.2 to 1.6 g/kg.
For a 70 kg adult, that is 56 g a day versus 84 to 112 g — one to two meals' worth of protein between the two answers.
Most coverage treated this as an update: the advice went up. That is not quite what happened. The old number is still there, still official. The new one simply appeared beside it.
The Short Answer
Use 1.2 to 1.6 g/kg as your target. Aim for the top of that range if you lift, are losing weight, or are over 60.
The RDA of 0.8 is not wrong. It answers a different question — what amount keeps nearly everyone out of deficiency. That is a floor, and it was never meant to be a goal.
How the old number was measured
There is a second reason these numbers differ, and it is the more interesting one: the floor itself was probably measured too low.
Start with how you would measure "enough" at all. If you eat less protein than you lose, your body breaks down its own muscle to cover the gap. If intake matches losses, you are covered. So researchers fed people different amounts and looked for the intake where the two sides matched — a method called nitrogen balance, since nitrogen is how protein gets counted. That balance point came out at 0.8.
The problem is the losses side. The studies collected urine and stool — but protein also leaves in shed skin, lost hair, and sweat, which nobody can collect. It is the household-budget mistake: count the card payments, miss the auto-drafting subscriptions, and a smaller income looks like enough. Losses looked smaller than they were, so "enough" landed lower than it really is. (A second bias in the analysis leaned the same way — PMID 17921376.)
What happened when they re-measured it
A newer technique starts from how the body actually uses protein. Building anything out of it — muscle, enzymes, skin — requires every amino acid the blueprint calls for, and the body keeps no storage bin for spares. A building block that cannot be used is burned for energy, and what is burned leaves in the breath as carbon dioxide.
That burning can be watched. Researchers put a trace tag on one amino acid and measure the breath. Feed too little protein and construction stalls — the tagged block goes unused, gets burned, and the tag shows up in the breath. Feed more and the burning drops, until adding protein stops changing it. The intake where the burn stops falling is "enough." The method is called indicator amino acid oxidation (IAAO).
Re-measured this way, "enough" came out at 1.2 g/kg — 50% higher than the old floor (The American Journal of Clinical Nutrition, 2007, PMID 17921376). Which means the lower bound of the 2026 target is not a new ambition; it is the old deficiency question, answered with a better ruler.
Why two numbers exist at all
Because two different bodies published them, for two different purposes, and only one of them moved.
| Number | Set by | Asks | |
|---|---|---|---|
| RDA | 0.8 g/kg | National Academies, 2005 | What prevents deficiency in nearly everyone? |
| Dietary Guidelines | 1.2–1.6 g/kg | HHS and USDA, 2026 | What protects lean mass during weight loss? |
The government's own 2026 review calls the RDA a deficiency floor and puts actual US intake at about 1 g/kg a day — above the floor, below the new target. That is why "Americans already eat enough protein" and "Americans need more" are both true: each side measures against its own number.
The 1.2–1.6 range itself came from trials of mostly overweight adults during weight loss, where extra protein preserved muscle and sped up fat loss. A specific situation — and the target drawn from it now addresses everyone.
One more thing: that review never cites the re-measurement work (PMID 17921376). The strongest reason to think the floor was too low plays no part in the case for raising it.
Does more protein mean more muscle?
Only up to a point, and only alongside training.
A large meta-analysis of resistance-training trials found that extra protein does add strength and muscle — and that the benefit flattens at about 1.62 g/kg (PMID 28698222). The government's upper bound sits at 1.6. The ceiling is not arbitrary.
The condition matters as much as the number. Protein is permissive, not causal — it supplies the material; training supplies the signal. Raising intake without lifting does not produce these results, a point that also holds for muscle during rapid weight loss on GLP-1 medication.
What This Means for You
How much. Your weight in kilograms × 1.2 is the lower target; × 1.6 is the upper. Aim high in three cases: training, a calorie deficit, or age over 60.
How to add it — this is the trap. Protein carries calories, about 4 per gram. Stack a shake on top of meals that already covered your day, and the protein goal is met while the day quietly gains 300 calories. Months of that is weight gain — the opposite of the point. So swap instead of stack: half the rice comes off the plate, two eggs go on. Same calories, more protein. The guideline's own bullet ends with exactly this adjustment — the half of the sentence nobody quotes.
When. The muscle-building response has an on-switch, and one meal needs roughly 20 to 25 g to flip it — more as you age (PMID 29497353). Ninety grams at dinner flips the switch once; thirty at each meal flips it three times. For most people the empty meal is breakfast.
Watching only protein hides the trap. "98 g today" reads the same whether it came from rearranging the plate or from stacking 300 calories on it. Only the calorie line next to the protein line tells those apart — which is why Vivledia draws both against their targets on one screen. The wider habits still outrank the macro: for most people, sleep and daily movement move the needle further than any gram target.
Where these numbers come from
| Study | What it looked at | What it found |
|---|---|---|
| Am J Clin Nutr 2007 (PMID 17921376) | Protein requirement re-measured by IAAO in young men | Average 0.93 g/kg, population-safe 1.2 g/kg — 41% and 50% above the nitrogen-balance figures |
| Br J Sports Med 2018 (PMID 28698222) | 49 trials, 1,863 adults, protein plus resistance training | Strength and lean mass improved; the benefit flattened at about 1.62 g/kg |
| J Nutr 2018 (PMID 30383278) | 28 trials, 1,358 participants, kidney filtration | No difference between higher- and normal-protein diets in healthy adults |
| Br J Nutr 2003 · 2023 (PMID 12908885, PMC10167660) | The case against IAAO: readings taken within hours may run high | The bias is real but small — several times too small to close the gap back to 0.8 |
The first row is the engine of this article: it is why the floor and the new lower bound landed on the same number. The last row is the counterweight — measured, found small, not settled.
This article is general health information, not medical advice. Protein needs differ by person — if you have kidney disease, are pregnant, or are managing a diagnosed condition, your own clinician's target replaces the arithmetic here.
