Precision Fermentation, Explained: How Microbes Learned to Make Milk Without Cows

Every few months, a startup announces it has brewed milk without a cow, eggs without a chicken, or cheese without, well, cheese. The technology behind most of these announcements is precision fermentation, and it is worth understanding precisely, because the term gets stretched to cover everything from yogurt cultures to lab-grown meat. It is neither. Precision fermentation is closer to brewing beer, except the yeast has been reprogrammed to manufacture one specific molecule, and then filtered out before anything reaches your plate.

How it actually works

The process has five stages, and each one is industrial biotechnology rather than cooking.

First, scientists select a target molecule, usually a single protein. The classic example is beta-lactoglobulin, the main whey protein in cow's milk. Second, the DNA blueprint for that protein is inserted into a host microbe, typically a yeast, a filamentous fungus such as Trichoderma reesei, or a bacterium. The engineered microbe now carries the instructions to build the target protein alongside its normal biology.

Third comes fermentation itself. The microbes are grown in stainless steel bioreactors, fed with sugar and nutrients under tightly controlled temperature, pH, and oxygen. As they multiply, they secrete or accumulate the target protein. Fourth is downstream processing: the broth is separated, the microbial cells are filtered out, and the protein is purified. This step matters more than most coverage suggests. Because the engineered microbe is removed, the final ingredient typically contains no detectable modified DNA, which is why such proteins can avoid bioengineered labeling in the United States, and why purification can represent a large share of total cost.

Fifth, the purified protein is dried or formulated and sold as an ingredient to food manufacturers, who use it exactly as they would the conventional version.

It helps to distinguish precision fermentation from its two cousins. Traditional fermentation, the kind behind beer, yogurt, and miso, uses microbes to transform a food matrix. Biomass fermentation grows the microbial cells themselves as the product, as with mycoprotein. Precision fermentation does neither: the microbe is a factory, the molecule is the product, and the factory is discarded.

Precision fermentation is not brewing beer. It is programming a microbe to manufacture one molecule, then filtering the microbe out before anything reaches your plate.

What you can actually buy in 2026

The honest answer is: ingredients, mostly, rather than finished products with the technology on the label.

Perfect Day remains the pioneer and by far the best funded company in the space, having raised more than $825 million. But the Perfect Day of 2026 looks different from the one that launched Brave Robot ice cream in 2020. That consumer brand is defunct, its parent The Urgent Company was sold, and Perfect Day has pivoted to selling its ProFerm whey protein as a business-to-business ingredient. ProFerm is approved for commercial use in the United States, India, Singapore, Colombia, the UAE, and Israel, with import approval for Hong Kong, and UK and EU applications still working through review. A dedicated production facility in Bharuch, Gujarat, built with Zydus, is expected to begin initial operations in the second half of 2026. The company has also faced headwinds: its co-founders departed in 2024, and in 2025 it was sued over its "animal-free dairy" marketing, a case it moved to dismiss.

The regulatory picture in the United States has widened well beyond one company. The FDA issued no-questions letters in 2025 covering beta-lactoglobulin, lactoferrin, collagen, and ovalbumin. TurtleTree received the first letter for precision-fermented lactoferrin in May 2025. Onego Bio secured GRAS notice 1249 for its Bioalbumen ovalbumin in September 2025. Vivici self-affirmed GRAS status for lactoferrin in February 2026 and launched in the United States, and All G received its FDA letter in April 2026. China, meanwhile, gave All G the first approval anywhere for recombinant bovine lactoferrin.

On shelves and in supply chains, Remilk launched milk made with its animal-free beta-lactoglobulin in Israel in November 2025 through a partnership with Gad Dairies, with US plans on the horizon. Impossible Foods has used precision-fermented heme for years to give its plant-based burgers their meaty character. Formo is developing fermented casein for cheese, New Culture is working on animal-free mozzarella, The EVERY Company is commercializing egg proteins, Geltor makes fermentation-derived collagen, and Phytolon is producing betalain food colors at tens of thousands of liters under a deal with DSM Firmenich. Myprotein sells a precision-fermented protein powder, and Breyers launched a lactose-free frozen dessert using fermented whey. These are real products, but notice the pattern: the technology succeeds as an invisible ingredient, not as a headline.

The cost curve

The most striking number in precision fermentation is the cost trajectory. Industry estimates put the cost of producing a kilogram of precision-fermented protein at roughly $1 million in 2000, falling to approaching $10 per kilogram today, a decline of more than 99.999 percent. The research group RethinkX projects these proteins could become five times cheaper than conventional animal proteins by 2030 and ten times cheaper by 2035, eventually approaching the cost of sugar. One market analysis puts Perfect Day's dairy protein production at $8 to $12 per kilogram, approaching parity with conventional dairy whey, though such figures should be read as company-adjacent estimates rather than audited costs.

What sets the floor? Three things: feedstock cost (mostly sugar), facility depreciation on expensive stainless steel, and downstream purification. Investors now price funding rounds against progress on those levers rather than the novelty of the molecule. One promising lever is continuous fermentation: today's standard fed-batch process is stop-and-start, with tanks emptied and cleaned between runs, while continuous systems keep microbes in the bioreactor and harvest product without interruption. A Dutch consortium called EcoFerm launched in October 2026 to demonstrate exactly that approach for proteins like beta-lactoglobulin.

The honest limits

Precision fermentation works, but its limits are structural. First, it has so far only cracked high-value proteins: whey, egg white, lactoferrin, collagen, specialty enzymes. Nobody is fermenting bulk calories; producing commodity protein this way remains far more expensive than conventional agriculture or plant-based bulk ingredients, partly because aeration and agitation make the process energy hungry.

Second, scale-up is capital intensive. Proving a strain in a lab is a different universe from running tens of thousands of liters reliably, and contract manufacturing capacity is a bottleneck the whole industry shares.

Third, regulation is a patchwork. The United States has moved fastest, while Europe trails by years: Impossible Foods' heme received positive EFSA opinions in 2024 with the Commission decision still pending, a typical timeline.

Fourth, the product still triggers dairy allergen labeling, because the protein is molecularly identical to the cow version, and consumer acceptance of the underlying genetic engineering remains contested, as the lawsuit against Perfect Day's marketing shows.

The fair summary: precision fermentation is a real manufacturing platform for high-value food proteins, with genuine products and a genuine cost curve. It is not, in 2026, a replacement for animal agriculture. It is a new ingredient industry growing up inside the old one.

References

Green Queen, AgFunderNews, Dataintelo precision fermentation market reports, FoodNavigator, Food Processing, Vegconomist, Wikipedia (Brave Robot), Non-GMO Report.