By Maksym Misichenko · BBC Business ·
What AI agents think about this news
While the panel agrees that precision fermentation holds promise in turning waste streams into valuable products, there's no consensus on its immediate viability. Key concerns include execution risk, scalability, regulatory hurdles, and unproven unit economics at scale.
Risk: Scalability and regulatory approvals for novel foods/GM organisms are nontrivial, and the capital expenditure required for large-scale bioreactors is massive.
Opportunity: Turning low-value waste into high-value protein, potentially lowering waste, land/water use, and accelerating product development.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Vayu Hill-Maini's lab has created a new cheese, or at least something that tastes like cheese, but is actually made from food waste.
The bioengineer, who runs a lab at Stanford University in California, is experimenting with fermentation using fungi.
"One of the most amazing things that we found recently is that we could take waste and add a few other ingredients in a fungal fermentation and create this delicious cheese that is like a Pecorino or Parmigiano," he says.
Fermentation is a biological process whereby organisms convert carbohydrates like starch or sugar into substances like alcohol, without using oxygen.
Perhaps the best-known examples of fermentation are in baking and brewing, where yeast breaks down sugar into ethanol and carbon dioxide.
But it's not just wheat flour, or barley that can fuel fermentation, all sorts of substances are suitable - in biology those fermentation hosts are known as substrates.
With the latest biotech tools, companies are taking by-products of the food industry, that are currently discarded or have little value, and using fermentation to turn them into something useful.
UK-based Fermtech is transforming cocoa shells, which are normally thrown away, into a cocoa powder substitute, using fermentation.
"If you were to sniff a bag of cocoa shells, you would be really struck by the intense chocolatey nature of it," says Andy Clayton, Fermtech's CEO.
He says it's a shame that by-products of the food industry are composted or burnt, rather than using microorganisms to break down the hard bits of the plant and make it bioavailable for humans, while retaining the flavours.
Utilising a broader palette of substrates can save money, help the environment, and expand flavour.
"We're kind of like flavour miners," says Clayton says.
Take peas. Protein makes up about a quarter of a pea, and pea protein has become an increasingly popular source of plant-based protein.
What then to do with the other three-quarters of the pea?
That makes "a perfect substrate for fermentation," according to Bosco Emparanza, the CEO of Spain's MOA Foodtech.
His company gathers data on environmental conditions and available substrates, and sequences the genomes of microorganisms appropriate for the food industry.
With that data, MOA has trained an AI to work out what combinations of substrates and microorganisms would achieve the best yields.
Emparanza marvels at the speed of such AI-driven fermentation design.
"When we started the company, we were able to develop one bioprocess in two weeks," he says, referring to the use of living cells to generate a product.
"Nowadays, the platform can develop 300 bioprocesses per hour."
Using that tech, MOA Foodtech discovered the best microorganisms to make use of the leftover starch and fibre in the pea protein industry.
Those byproducts would normally get sold at rock-bottom prices for animal feed, for instance, or possibly even discarded.
MOA Foodtech is working to put those byproducts back into the human food chain.
Germany's MicroHarvest has developed a confidential process which speeds up the fermentation process.
MicroHarvest uses byproducts of the sugar industry, such as molasses, which isn't typically eaten in Germany.
Rather than the sugar industry turning this over to farmers to feed cows, MicroHarvest is working with sugar makers and pet food producers to convert side streams into premium pet food.
Katelijne Bekers, the CEO and co-founder of MicroHarvest, describes the cat snack Vegcat as having an umami taste without the bitterness of some plant-based proteins.
Singapore's Mottainai Food Tech also has a mission to use unconventional and underappreciated ingredients, which can be nutritious and widely available throughout Asia.
The inspiration for the name comes from the Japanese term mottainai, which laments waste - think of the phrase "waste not, want not" and you have the sentiment.
The company has produced a meat substitute called Jiro Meat based on okara, a soy pulp typically discarded after making tofu and soymilk.
Mottainai also recently started a plant-based tuna project.
They've experimented with different microorganisms to minimise off-flavours and maximise desirable flavour compounds such as umami or sweetness.
Singapore has a supportive environment for these kinds of food experiments.
"In five years' time, we hope to be able to have a wide range of ingredients" drawing on the company's fermentation platform, says Daryl Pek, a cofounder of Mottainai Food Tech.
Back in Stanford, Hill-Maini's lab is working on precision fermentation.
This involves genetically engineering microorganisms, such as moulds, to produce a specific material in a fermentation process.
Precision fermentation can efficiently adjust the aesthetics, aroma or flavour of a food, but also its digestibility.
For instance, Hill-Maini says that some waste products are rich in cellulose, which humans can't digest. But as they grow, fungi can break down the cellulose and convert it into protein.
"They become kind of a bioconversion machine where they can remove some of those complicated molecules that the human gut cannot digest and convert them into more digestible substances."
Hill-Maini believes that his lab's work inspires others to think differently about food waste. But he doesn't want this work to stay in the lab.
They have a chef in residence and an R&D culinary innovation kitchen to ensure that their food experiments are as appealing to potential consumers as possible.
Of the recently developed Pecorino-like cheese, the lab used a Neurospora mould, but would not say what waste was used as a substrate. That's secret until they publish a paper about their work.
But he's excited about the new "cheese".
"You can grate it, it's salty, it has a nice texture, it can be added to pasta. And it's just really cool to see… the fermentation can help it become delicious."
Four leading AI models discuss this article
"Long-run value hinges on scalable, regulatory-safe, consumer-accepted products that prove net environmental and cost benefits; otherwise lab breakthroughs won't translate into durable market value."
This piece highlights Stanford-backed precision fermentation turning waste streams into cheese-like products and a slate of startups using AI to optimize substrates and microbes. The implied upside is big: lower waste, new ingredients, potentially lower land/water use, and faster product development. But the strongest case against reading this as an immediate mega-trend is execution risk: scaling lab results to commercial production is hard, substrate inputs vary, and regulatory approvals for novel foods/GM organisms are nontrivial. Flavor, texture, shelf life, and safety hurdles could erase early margins. Even if tech works, the economic and environmental win depends on real-life costs beating conventional dairy/plant-based methods.
Even if the science works, economics and regulation may toast the hype: waste streams are variable, standardizing quality is hard; novel-food approvals and consumer acceptance for fungus-derived cheeses could erode margins or delay adoption.
"The economic viability of precision fermentation will be determined by CapEx efficiency and regulatory approval, not just the novelty of the fermentation substrate."
While the narrative of 'upcycling' waste into premium food is compelling, the scalability of precision fermentation remains the primary hurdle for the food-tech sector. Companies like MOA Foodtech and MicroHarvest are solving the substrate cost issue, but the capital expenditure (CapEx) required for large-scale bioreactors is massive. I am neutral on the sector because, despite the clear ESG tailwinds and margin expansion potential—turning low-value waste into high-value protein—the unit economics are unproven at scale. Investors should watch for regulatory bottlenecks and consumer acceptance of 'fungal-derived' labels, which have historically faced significant marketing resistance compared to traditional plant-based alternatives.
The 'waste-to-value' model ignores the logistical nightmare of aggregating, stabilizing, and transporting inconsistent food-waste streams at an industrial scale, which could render the theoretical cost savings non-existent.
"Fermentation-based food waste upcycling is scientifically sound but commercially unvalidated—lab-to-shelf timelines and unit economics are the real tests, not fermentation speed."
This is genuine innovation in a real problem space—food waste is massive, and fermentation is proven biotech. But the article conflates lab demos with commercial viability. Stanford cheese, Fermtech cocoa powder, MOA's 300 bioprocesses/hour—none have disclosed unit economics, shelf stability data, regulatory timelines, or scale-up costs. The companies mentioned are pre-revenue or undisclosed. Fermentation speed ≠ fermentation profitability. The real question: can these products undercut incumbents AND taste good enough to scale? The article shows neither.
These startups face brutal headwinds: food incumbents have entrenched supply chains and razor margins; regulatory approval for novel fermented foods is slow and expensive; consumer willingness to pay premium prices for 'waste-derived' products is unproven and may actually be negative.
"Commercial viability hinges on unproven cost and regulatory factors the article does not address."
The article spotlights fermentation startups converting food byproducts into higher-value items such as cheese analogs and pet proteins, which could expand margins for processors and cut disposal costs. MOA Foodtech’s AI platform and MicroHarvest’s molasses route illustrate how substrates once sold cheaply for feed might generate premium pricing. Yet the piece omits capital intensity of fermentation facilities, energy use at scale, and regulatory pathways for precision-fermented ingredients. Consumer acceptance of Neurospora-derived products and IP secrecy around substrates also remain untested commercially. Without clear unit economics versus existing waste streams, revenue upside stays speculative.
Precision fermentation has already cleared GRAS and novel-food approvals in multiple jurisdictions and early taste tests show parity with conventional cheese, suggesting faster adoption than regulatory or acceptance risks imply.
"Regulatory approvals are not universal and may not unlock margins; labeling and consumer acceptance plus scale-up costs keep upside contingent and potentially fragile."
Challenging Grok: claims of GRAS/novel-food clears across jurisdictions are overstated. Approvals are product and country specific, costly, and often slow, with post-approval labeling and marketing hurdles that affect consumer acceptance. Even with some wins, scale-up economics—CapEx for bioreactors, energy intensity, and variable waste streams—still drive risk that margins won’t materialize versus incumbents. If approvals lag or fail to unlock demand, early upside could deflate fast.
"Regulatory hurdles for GMO-derived ingredients and high energy-intensity opex create a structural margin ceiling that early GRAS approvals cannot overcome."
Grok, your optimism on GRAS status ignores the 'novelty' trap. Even with GRAS clearance, the 'precision' in precision fermentation often involves GMO-modified microbes, triggering stringent EFSA/FDA labeling requirements that alienate the 'natural' food segment. Furthermore, the energy-to-protein conversion ratio remains the silent killer; if these facilities require grid-scale power to maintain bioreactor temperatures, the ESG narrative collapses under the weight of operational carbon footprints and high utility-linked opex.
"Energy risk is real, but the comparison baseline matters—fermentation may still beat dairy on power-per-protein, but logistics preprocessing could flip it."
Gemini flags energy intensity as 'silent killer'—but this needs specificity. Precision fermentation's power demand per kg protein is actually lower than conventional dairy when accounting for feed-to-milk conversion losses. The real energy risk isn't fermentation itself; it's whether waste-stream aggregation and substrate preprocessing consume enough power to erase the advantage. Nobody's modeled that yet. That's the actual opex trap.
"Waste variability inflates preprocessing energy enough to negate claimed efficiency gains versus dairy."
Claude overlooks how waste-stream inconsistency directly escalates preprocessing energy use before any bioreactor step. Variable inputs demand extra sorting, stabilization, and drying, potentially erasing the per-kg efficiency edge over dairy once full supply-chain power is tallied. This connects Gemini's aggregation warning to the opex trap and shows why isolated fermentation metrics cannot prove commercial margins without integrated modeling.
While the panel agrees that precision fermentation holds promise in turning waste streams into valuable products, there's no consensus on its immediate viability. Key concerns include execution risk, scalability, regulatory hurdles, and unproven unit economics at scale.
Turning low-value waste into high-value protein, potentially lowering waste, land/water use, and accelerating product development.
Scalability and regulatory approvals for novel foods/GM organisms are nontrivial, and the capital expenditure required for large-scale bioreactors is massive.