Tofu Calories and Protein, by Firmness

Two USDA records both say "firm," and they are not the same food#

USDA's entry for tofu, raw, firm, prepared with calcium sulfate lists 144 calories, 17.3 grams of protein and 683 mg of calcium per 100 grams¹. Its entry for tofu, firm, prepared with calcium sulfate and magnesium chloride (nigari) lists 78 calories, 9.04 grams of protein and 201 mg of calcium². Same database, same word on the label, and an 85% gap in energy, a 91% gap in protein and a 240% gap in calcium.

USDA does not leave that unexplained. The second record carries a note stating that "descriptive terms (soft, firm, extra firm) vary in usage between manufacturers as there are no standards for the various types of tofu." So the answer to how many calories and how much protein are in tofu is: between about 55 and 145 calories and between about 5 and 17 grams of protein per 100 grams, and the word on the package is not what tells you which. Two things do — how much water was pressed out, and which salt was used to set the curd — and neither is printed in a form most shoppers read. This is one of the rare rows in the common-foods reference where the variation is manufactured rather than biological, and therefore knowable if you look in the right place.

Press out water and nothing else really changes#

Here is the whole tofu ladder, from USDA records. The final column is my division of their protein figure by their energy figure — arithmetic on their numbers, not a fifth measurement.

USDA entry	Energy	Protein	Fat	Water	Calcium	Protein per 100 kcal
Firm, calcium sulfate¹	144 kcal	17.3 g	8.72 g	69.8 g	683 mg	12.0 g
Firm, calcium sulfate + nigari²	78 kcal	9.04 g	4.17 g	82.9 g	201 mg	11.6 g
Regular, calcium sulfate³	76 kcal	8.08 g	4.78 g	84.6 g	350 mg	10.6 g
Silken, soft⁴	55 kcal	4.80 g	2.70 g	89.0 g	31 mg	8.7 g

Read the energy column and the water column together and they move in lockstep in opposite directions. From silken to the firmest entry, water falls from 89.0 grams per 100 to 69.8 — a difference of 19 grams — and energy rises from 55 to 144. Water carries no calories, so nothing was added; something was simply removed, and what remained got concentrated. Tofu firmness is a dewatering step, not a recipe change.

Which is why the last column barely moves. Protein per 100 calories runs from 8.7 grams in silken to 12.0 in the firmest entry — a real spread, but a narrow one against the 162% range in energy density. Per calorie, every tofu on this page is roughly the same food. Per block, they are not remotely the same food. If you are counting calories, firmness matters enormously. If you are asking whether tofu is an efficient way to buy protein, it is about equally efficient at every firmness, and comfortably efficient in absolute terms — firm tofu sits near the top of the practical plant list precisely on this ratio.

The practical size of it: a 350-gram retail block works out to about 504 calories and 61 grams of protein at the firm calcium-sulfate figures, or about 273 calories and 32 grams at the firm nigari figures. That is a 231-calorie and 29-gram difference between two packages that could carry identical wording on the front. Weighing the block does not resolve it. Reading the panel does.

The coagulant, not the soybean, sets the calcium#

The calcium column deserves its own look, because it swings by a factor of 22 across four records of the same food and the reason is a manufacturing choice with a name.

Tofu is made by curdling hot soy milk with a coagulant. Two are standard. Calcium sulfate — gypsum, the traditional Chinese choice — sets a softer, higher-yield curd and deposits calcium in it. Magnesium chloride, sold as nigari and traditional in Japan, sets a firmer curd with a cleaner flavor and deposits magnesium instead. USDA's records show exactly what that does: the calcium-sulfate-only entries carry 683 and 350 mg of calcium per 100 grams, the nigari-blended one 201 mg, and the silken entry 31 mg.

At 683 mg per 100 grams, calcium-set firm tofu is one of the densest non-dairy calcium sources in an ordinary supermarket — a 100-gram serving covers roughly two thirds of an adult's 1,000 mg daily requirement. At 31 mg, silken tofu contributes essentially nothing on that axis. The two are separated by no visible property and no regulated term. The only reliable signal is the ingredients line: if it says calcium sulfate, the calcium is in there; if it says magnesium chloride or nigari, it mostly is not.

That is the mirror image of the way straining removes minerals from strained dairy, an effect worked through in Greek yogurt's calcium arithmetic. Here the mineral arrives with the processing rather than leaving with it — but in both cases the nutrition panel records the outcome and the name on the tub records nothing.

The question people actually ask, and who is funding the answer#

Almost nobody worries about coagulants. They worry about isoflavones — the plant compounds in soy that bind weakly to estrogen receptors — and specifically about whether eating soy feminizes men or feeds hormone-sensitive cancers. The clinical literature is reassuring on both counts, and it is also, almost uniformly, funded by people with an interest in that answer. Both halves of that sentence matter.

On male hormones, a meta-analysis pooled 15 placebo-controlled treatment groups with baseline and ending measures, drawn from a wider set of 32 reports covering 36 treatment groups, and tested soy protein and isoflavone intake against total testosterone, sex hormone-binding globulin, free testosterone and the free androgen index. The finding is a flat null: "no significant effects of soy protein or isoflavone intake on T, SHBG, free T, or FAI were detected regardless of statistical model"⁵. The caveat belongs beside it rather than in a footnote: one of the authors, Mark Messina, is a career soy-industry scientist who has directed the Soy Nutrition Institute, an industry body, and who consults for soyfood and isoflavone companies. That does not make a null result wrong — nulls are the hardest findings to manufacture — but it is the same standard this blog applies to dairy-funded fat research and supplement-funded protein research.

On cancer, the strongest data are observational and come from a population that eats far more soy than Western ones do. Among 5,042 Chinese women treated for breast cancer and followed for a median 3.9 years, the highest soy protein intake after diagnosis was associated with a hazard ratio of 0.71 (95% CI 0.54–0.92) for total mortality and 0.68 (95% CI 0.54–0.87) for recurrence, against the lowest intake — four-year mortality of 7.4% versus 10.3%⁶. This is a cohort study, so it cannot establish cause, and the intake contrast is wide in a way no Western diet reproduces. Its primary author disclosed a 2005 research development fund from the United Soybean Board, with the board stated to have had no role in design or interpretation.

So the fair summary is narrower than "soy is proven safe" and much narrower than "soy is dangerous." The direction of every well-designed measurement points away from harm; nearly every one of those measurements sits somewhere downstream of the soy industry; and the fear that generated the question was never supported by a controlled human trial in the first place.

Reading a package that nobody regulates#

Three things follow from all of this, and the first two take one glance each.

Read the panel, not the front. Since "firm" is not a defined term, the calorie and protein figures on the label are the only ones that describe the block in your hand. The numbers in the table above are population values from a database, and a specific brand can sit a long way from them.

Read the coagulant line if calcium matters to you. Calcium sulfate or magnesium chloride is the whole story, and it appears in the ingredients rather than in any claim.

Then stop optimizing. Tofu at any firmness delivers roughly 9 to 12 grams of protein per 100 calories, which puts it in the same efficiency bracket as lean animal foods and near the top of any plant list. Soy also scores well on the amino-acid quality measures that separate plant proteins from each other, and where it lands on that ranking is a question of its own — but once the daily total is met, the source stops deciding the outcome, which is the settled result in plant vs animal protein. The variable that is actually worth your attention on a tofu package is the one printed in grams, and the reason it is worth your attention is that the word next to it means nothing.

FAQ#

How much protein is in a block of tofu?#

Between about 32 and 61 grams for a 350-gram block, depending on the firmness and the coagulant. USDA lists firm tofu set with calcium sulfate at 17.3 grams of protein per 100 grams and firm tofu set with a calcium sulfate and nigari blend at 9.04 grams (FDC 172475; FDC 172448). Since "firm" has no standard definition, the nutrition panel on your specific block is the only figure that describes it.

Does eating soy lower testosterone in men?#

The pooled clinical evidence says no. A meta-analysis of 15 placebo-controlled treatment groups, from a wider set of 32 reports, found "no significant effects of soy protein or isoflavone intake" on total testosterone, SHBG, free testosterone or the free androgen index, under any statistical model tested⁵. Worth knowing that one of its authors has led an industry body, the Soy Nutrition Institute — a disclosure that argues for weighing the pattern of evidence rather than any single paper.

Is silken tofu a good protein source?#

Per calorie, nearly as good as firm; per spoonful, much weaker. USDA puts silken soft tofu at 55 calories and 4.80 grams of protein per 100 grams, which is 8.7 grams of protein per 100 calories against firm tofu's 12.0 (FDC 174292; FDC 172475). The difference is water. Silken works as a protein food if you eat enough of it, which for a meal-sized dose means a lot of volume — one reason running the calories-per-gram check on soft plant foods usually confirms they are lighter than they look.

Sources#

This article was researched and drafted with AI assistance and reviewed for accuracy by the BurnWeek team. It is general information, not medical advice. How we research and correct our articles →