There is no single answer, and the reason is not caution. It is that “500 mg” means a different quantity of the thing you are actually buying depending on which molecule is in the bottle — and then a different quantity again once route and conversion are accounted for. Before anyone can answer how much, the number has to be made comparable, and almost nobody does that.
This page is the arithmetic and the reasoning. If what you want is the list of doses researchers administered, that is the dosage page for NAD⁺ broadly and the NMN page for that compound specifically; if you want what sellers print, the injection figures are collected separately.
A milligram of what?
NAD⁺ and its precursors are different sizes, and a milligram is a weight rather than a count. NAD⁺ itself has a formula weight of about 663 daltons. NMN is about 334. Nicotinamide riboside chloride, the form used in the trials, is about 291. So the same weight of each contains very different numbers of molecules — and since one molecule of NMN is converted to one molecule of NAD⁺ in a single enzymatic step, molecules are the unit that matters.
Do the division and the market inverts. A 300 mg NMN capsule carries about 0.90 millimoles of NAD⁺-equivalent. A 500 mg NAD⁺ vial carries about 0.75. The capsule is the larger dose, by roughly a fifth, at a fraction of the price.
Extend it over a month and the gap stops being a curiosity. A stated 1,000 mg monthly injection supply — a figure some sellers on the injection board do publish — is about 1.5 millimoles of NAD⁺. Thirty days of a 300 mg NMN capsule is 9 grams, which is about 27 millimoles. Roughly eighteen times the NAD⁺-equivalent, for a small fraction of the cost.
Two honest caveats, and they cut in opposite directions. The injected route bypasses the gut, so a milligram given parenterally is not expected to behave like a milligram swallowed. And oral conversion is not one-for-one in practice: studies tracking the metabolites find the picture messier than the textbook pathway, with intermediates rather than the parent compound rising in blood. Both of those are reasons the milligram figures are less comparable, not more. Nobody has published the conversion factor that would let you line them up, which is precisely why no seller can tell you what their dose means relative to anyone else’s.
The dose you take is not the blood level you get
There is a genuine dose-response in the biomarker, and it is the most reliable finding in the field. An eight-week randomized, placebo-controlled trial in overweight adults found that 100, 300 and 1,000 mg of NR daily raised whole blood NAD⁺ by 22%, 51% and 142% respectively within two weeks, and held those increases for the rest of the study [1]. A pharmacokinetic study in eight healthy volunteers that started at 250 mg and uptitrated to 1,000 mg twice daily found whole blood NAD⁺ doubled, with the change in NR and the change in NAD⁺ highly correlated (R² = 0.72, p = 0.008) [2].
Then read the trial that did not get that result. A placebo-controlled, double-blinded phase 2 study of oral NR in a cutaneous nerve model reports that NR supplementation did not result in elevation of plasma NAD⁺ levels at all, producing only a small increase measured in skin samples [3]. Same molecule, same route, no blood response.
So even the one thing these products reliably do is not reliable in every study or every tissue. A milligram figure does not determine a blood level, a blood level does not determine a tissue level, and nobody is measuring yours.
The studied territory, by format
Kept deliberately short, because the detail belongs on the pages above. Oral precursors have been given at 250 to 2,000 mg a day for 14 days to 24 weeks across the randomized trials pooled in a 2026 meta-analysis, without an increase in adverse events or liver enzymes [4]. That is the territory where a dose question has any published answer at all.
Intravenous NAD⁺ has one randomized, placebo-controlled trial, and its dose is the number to hold onto: 10 mg a day for seven days, given to 180 hospitalized adults with heart failure alongside guideline-directed therapy [5]. A single wellness infusion commonly sells 250 mg or more. The only other human data is a pharmacokinetic pilot which infused NAD⁺ at 3 µmol per minute over six hours and found plasma NAD⁺ and its metabolites unchanged until after two hours, concluding the infused molecule was being rapidly and completely removed from plasma for at least that long [6].
There is no published dose for NAD⁺ injected into muscle or under the skin, because there is no published trial of that route to take one from — a 2026 PRISMA-guided systematic review covering January 2010 to October 2025, 113 eligible studies and 33 human intervention studies, states that no eligible outcomes trial evaluated intravenous or intramuscular NAD⁺ itself [7].
Why there is nothing to titrate against
A dose for a medicine is chosen by finding the point where benefit stops growing faster than risk. That requires a benefit that moves with dose. The same systematic review describes the human evidence as consistent biochemical target engagement with effects on functional, metabolic and vascular outcomes that were heterogeneous and often null [7].
Which leaves blood NAD⁺ as the only thing that responds to dose. You can aim at it — more milligrams do produce a higher number — but you are then optimizing a biomarker rather than an outcome, and the outcomes have mostly been measured and mostly have not appeared. That is the substance of the answer to “how much”: as much as a trial used, if you want to stay inside tested territory, and no more, because there is nothing more to buy.
The four questions that replace a number
Which molecule, and how many milligrams of it? Not of a blend, not of a proprietary complex — of the named compound. If a provider will not give you the figure, the product cannot be compared with anything, including a cheaper version of itself.
Per what? A milligram figure attached to an injection, a vial and a month are three different quantities that look identical on a checkout page. Our dose-value tool keeps the basis attached for exactly this reason.
Is it inside the studied range? For oral precursors that means roughly 250 to 2,000 mg a day. Below it, nothing has been tested; above it, nothing has either.
What does it cost per gram, across formats? Run the molar comparison above and the answer usually settles the format question before the dose question arises — the cost calculator turns competing offers into one unit.
Then take those answers to a clinician who knows your history. That is not a disclaimer bolted on the end; it is the only route to a number that applies to you rather than to a trial population, and what the safety evidence can and cannot establish is worth reading before the conversation.