The plant-based boom failed to deliver. But there may yet be hope.
In
an office block in south London a chef was trying hard to prove that
vegan cheese wasn’t disgusting. Violife, a brand that makes dairy-free
products, had asked him to make a multicourse lunch to show off its cheese alternatives.
The chef, Duncan Parsonage, served up a fake-meat Cuban sandwich topped
with gooey mock mozzarella, tortilla chips with whipped Greek White (a
version of feta) and, to top it off, a doughnut filled with imitation
cream cheese. It was all tasty enough, but it was clear that the cheese
needed dressing up. Even Parsonage confessed the products would struggle
to hold their own on a cheese board. As vegans know, ersatz cheese can
be waxy and bland—and nowhere near as good as the real thing.
I swore off animal products several years ago, when the vegan-food craze was
at its peak. Propelled by wellness influencers and Netflix exposés of
the meat industry, veganism became aspirational—and lucrative. Venture
capital and celebrity finance poured into firms such as Impossible Foods
and Beyond Meat, the latter of which boasted Bill Gates and Leonardo
DiCaprio among early investors and went public in 2019 at nearly $4bn.
In 2021 Oatly, a Swedish oat-milk firm that was backed by Oprah Winfrey,
Jay Z and Natalie Portman, was valued at around $13bn. Its chief
executive hailed the stars’ investment as “a clear indication of where
the world is heading”. Food, like popular culture, was going woke.
During
the pandemic, however, it became clear that investors had overestimated
consumers’ appetite for vegan alternatives. Sales of plant-based meat
and milk have since stalled, weighed down by higher prices and growing
suspicion of ultra-processed foods. Beyond Meat is now worth less than
$400m and Oatly has lost 97% of its peak value. Yet although the
companies at the forefront of the vegan-food revolution are worth less,
most of their products are still widely available. Fortunately for
people like me—and for anyone with a dairy allergy or
intolerance—plant-based milk is pretty good. Last year it accounted for
13% of American milk sales, while Square, a payments firm, has said that
oat milk is chosen for one in three coffee orders made through its
platform. Even cafés in the back of beyond serve oat-milk lattes, and
most supermarkets stock several varieties.
Propelled by wellness influencers and Netflix exposés of the meat industry, veganism became aspirational—and lucrative
But
there is one dairy alternative you’re unlikely to spot in shopping
trolleys, even if that trolley contains dairy-free ice cream and
chocolate. Vegan cheese makes up less than 1% of the American cheese
market, and sales were 10% lower at the end of 2025 than a year earlier.
Clearly I’m not the only consumer who finds it offputting.
Why
is vegan cheese still so disappointing? The answer can be found under a
microscope. Milk consists of two proteins: whey and casein. Casein is
what makes cheese cheesy. Unlike most proteins, it doesn’t form a fixed
shape. Instead, it gathers into tiny clusters that trap fat and calcium
(these are what makes cheese stretchy when melted). As cheese ages, the
clusters slowly break down, releasing enzymes that produce complex
flavours. Plant proteins can’t replicate this behaviour: they are more
rigid and don’t bind fat and calcium in the same way.
Companies
like Violife have attempted to sidestep the casein problem by focusing
not on what cheese is, but what it does. They generally use coconut oil
and tapioca starch to replicate some of cheese’s distinctive properties.
Like cheese, coconut oil is solid at room temperature and melts under a
grill. Tapioca gives the melted oil its tackiness and stretch (think
tapioca balls in bubble tea). Synthetic flavouring and salt are added
for the final touch.
Parsonage
told me vegan cheese is best enjoyed as an ingredient on a burger or
pizza. Even then, texture is a problem. Shreds of Violife blocks “lock
up” under direct heat in a pizza oven, softening around the edges but
refusing to melt. Their fake Camembert comes with instructions: don’t
bake it for as long as a real one, or it will turn into liquid rather
than a gooey treat.
There
are other snags. Despite being plant-based, most vegan cheese isn’t
very nutritious. It’s high in saturated fat and contains almost no
protein. It also doesn’t taste of much, apart from coconut. In theory,
any taste can be mimicked with the right combination of synthetic
flavours, but recreating the notes of a ripened cheddar is prohibitively
expensive.
https://cdn.economistdatateam.com/videos/1843/vegan-cheese/4_1280.mp4
There
is an obvious solution: make casein without a cow. A handful of
scientists and entrepreneurs are using a process called precision
fermentation to do this. Fermentation is when a microorganism, like
bacteria or yeast, breaks down sugars into new compounds. It turns grape
juice into wine, cabbage into kimchi, and milk into cheese. Precision
fermentation involves genetically modifying that microorganism to make
it break down sugars in a specific way. (In the 1980s it was used to
make synthetic insulin—previously diabetics had to inject themselves
with insulin from animal pancreases.)
Last
year I visited Wageningen University in the Netherlands, where teams
are trying to perfect the method. Scientists there have isolated the DNA
sequence in cows that produces casein. They stick it into E. coli
bacteria, give the bacteria glucose to feed on, and caseins get left
behind. They are then mixed with oil and water to make a kind of
mozzarella. “It’s more similar to the animal product,” Julia Kepler, one
of the scientists, told me. “Though, of course, we haven’t tasted it
yet.”
Under European Union (EU)
law, precision-fermented cheese is classified as a “novel food” and
must receive approval before it can be marketed to consumers—a process
that can take years. The EU doesn’t explicitly ban
scientists from tasting the food, but most countries take a cautious
approach, treating any human consumption as off-limits until approval is
granted. (Since my visit, the Dutch government has relaxed its rules
and now allows tastings under strict conditions—the first EU country to do so. Researchers in Wageningen hope to organise their first tasting later this year.)
Some EU
countries are deeply resistant to alternative proteins. The Italian
parliament banned lab-grown meat in 2023, in defence of the country’s
culinary traditions. The move was pre-emptive: cultivated meat is not
yet approved for sale in the EU (it’s also not clear
whether the ban would survive a court challenge). But the vote
demonstrated strong feelings about what the country’s agriculture
minister called “the social and economic risks of synthetic food”.
Vegan
cheese makes up less than 1% of the American cheese market, and sales
were 10% lower at the end of 2025 than a year earlier
The
work at Wageningen is funded through government grants and partnerships
with food companies, which will have to confront the EU’s
regulatory hurdles before bringing the lab’s creations to market. Those
companies will face another challenge: precision fermentation is
expensive, and it’s unlikely cheese created in this way could ever
compete on cost with real cheese.
For
now, though, the researchers are trying only to prove the technology
works. They test the physical performance of samples against real
cheese, measuring density, melting temperature and stretch length.
Kepler insisted she’s never sneaked a bite. Previously, she spent four
years working on precision-fermented milk before she was allowed to try
some. That was at a conference in America, where regulations are more
relaxed. “People are consuming it there,” she said, “and they’ve
survived.”
It
is no surprise that California has become a breeding ground for
alternative proteins. Its tech, research and wellness industries are
fuelled by abundant venture capital and lack of red tape—at least
compared with Europe. Beyond Meat and Impossible Foods both emerged from
this ecosystem, alongside a lesser-known company called Climax Foods.
Its
founder, Oliver Zahn, had been a data scientist at Google, and before
that at SpaceX and Impossible Foods. He grew up in Germany and had come
to America for postgraduate studies in astrophysics. He wrote
influential papers on the origins of the universe and served as director
of the Centre of Cosmological Physics at the University of California,
Berkeley. But although he enjoyed decoding the mysteries of space, he
was increasingly distracted by another problem of cosmic proportions:
how to make vegan cheese taste good.
https://cdn.economistdatateam.com/videos/1843/vegan-cheese/2_1280.mp4
I
first spoke to him on a video call early last year. He is long and
lean, with a shaven head and dark eyebrows. “There was nothing even
borderline good as an alternative” for vegans like him, he told me, not
to mention the two-thirds of the global population that has some form of
lactose intolerance. In 2019 he founded Climax, which he said would fix
vegan cheese in the same way he’d approached astrophysics: by making
sense of complex data. The following year Zahn left Google to run Climax
full time. By then, Beyond Meat had gone public and Zahn wanted in on
the plant-based revolution.
Climax,
which was based in Berkeley, pitched itself as the kind of foodmaker
only Silicon Valley could produce. It was a “data-science company
innovating the future of food”, according to its website. The company
boasted that its scientists would develop plant-based cheeses through
“sophisticated machine intelligence tools”, including using artificial
intelligence (AI) to predict what would happen when a
recipe was tweaked. Climax’s name and logo, in which the “M” and “A”
formed mountain peaks, suggested a pinnacle—products “better in every
way”, as Zahn liked to say in interviews, than the cheeses they were
designed to imitate.
Zahn
raised $7.5m in seed funding and began recruiting through his industry
contacts. He and his new team would do the hard thing first: speciality
cheeses. This was a strategy Zahn had picked up from Elon Musk when he
worked at SpaceX. Zahn’s inspiration was the Tesla Roadster, a high-end
proof of concept that ushered in the mass adoption of electric vehicles.
“If you can make the fanciest, most difficult car to get right, do that
really well, you can hopefully do all the other things really well,”
Zahn told me.
Making
good on Climax’s promise to leverage machine learning would take time.
It would first require a data set, fed with the results of experiments
on a wide range of potential ingredients. In the meantime, the priority
was making a cheese to show to investors. Zahn’s team set to work on a
full slate of varieties, including brie and goat’s cheese. Instead of
relying on the usual oil and starch, they built their recipes around
seeds and legumes. These ingredients were far more nutritious but they
were also expensive—a handful of companies had made nut- and seed-based
cheeses, but their products were priced high and hard to find. Zahn,
however, was confident his team could perfect the taste first and bring
costs down later.
Although
he enjoyed decoding the mysteries of space, he was increasingly
distracted by another problem of cosmic proportions: how to make vegan
cheese taste good
Soon
they had something to show for their efforts: Climax Blue, an imitation
wheel of blue cheese. It looked impressive: ripples of mould ran
through a creamy base of blended seeds, beans and coconut cream. Zahn
spent more and more time on the road, showing off Climax Blue, while his
team worked back at the lab in Berkeley. He appeared at food expos and
met star chefs, promising “zero-compromise” vegan alternatives. At one
point, he flew to London for a private tasting with Paul McCartney, who
became an investor. Zahn sent me a photo of himself grinning and
grasping hands with the former Beatle next to a table strewn with cheese
samples. Zahn loved watching the delight of people who tried his cheese
for the first time. “I wish you could have filmed their faces,” he told
me.
The
company was looking for a big cash injection that would allow it to
make vegan cheese at scale. At the start of 2022 a tech fund in London
promised a $60m investment. Zahn and his team were on a high. “The
assumption was that we were all going to explode, like the best of the
food tech startups,” one former employee told me. The startup’s staff
grew to roughly 50 and Climax poured money into research and
development.
In
a way, blue cheese had been relatively easy to get right; its most
important quality, taste, could be replicated using the same mould. The
real challenge lay with more popular varieties, like mozzarella, where
physical properties, like melt and stretch, were paramount. To get those
right, Climax would need to confront the casein problem. They would
need to develop a process for modifying a seed protein to make it
function like a milk protein, without resorting to costly precision
fermentation.
https://cdn.economistdatateam.com/videos/1843/vegan-cheese/3_1280.mp4
In
April 2022 Zahn was in a meeting when he got a text from one of his
staff telling him to come to the lab right away. Zahn arrived to see
employees shooting video on their phones. One of the scientists had been
working on an analogue to mozzarella. He’d accidentally overheated the
sample, thinking he was setting the oven temperature in Fahrenheit when
it was programmed in Celsius. Under the high heat, his sample did
something that, up to that point, had seemed near impossible for food
made with plant proteins: it melted and stretched, just like real
cheese. Zahn became emotional. “We knew immediately we had hit jackpot.”
Climax
had cracked it. They called the ingredient Caseed. It would not only
make a stretchier cheese, but one higher in protein than existing vegan
cheeses. The process needed fine-tuning and was expensive. But with the
promised cash from the London fund they would surely have enough runway
to figure it out.
By
the middle of 2022, however, the heyday for plant-based foods appeared
to be ending. Over the previous year, Beyond Meat had lost nearly 85% of
its value. Investors had been treating food like it was software, which
can be scaled and sold with the click of a button. But increasing
supply of a food product means building more production capacity, and
margins were slim, especially given the need to compete with meat and
dairy on cost.
The
London fund pulled out. Despite the setback, things appeared to be
going well for Climax. The public got their first tastes of Climax Blue
at pop-up events in December 2022. By the end of the following year,
more than 15 restaurants, among them the renowned Atelier Crenn in San
Francisco and Eleven Madison Park in New York, had included the product
in their dishes. Zahn continued to pull together enough money from a
group of small investors to keep the company going. In April 2023 Bel
Group, the French makers of Babybel and Boursin, announced that it would
be working with Climax to make new vegan versions of its products using
Caseed. A joint press release talked of “leveraging the power of data
science and artificial intelligence” to create “nutritious and
affordable” cheeses that would be “indistinguishable from their dairy
counterparts”.
The
following April staff at Climax were excited to hear the company had
reached the finals of a prestigious competition, the Good Food Awards.
Their cheese would be competing against dairy cheeses in a blind taste
test. According to Zahn, the competition organisers emailed Climax to
say it had won. But before the awards ceremony, Climax was disqualified.
Someone had complained that its cheese included kokum butter, which
isn’t on the Food and Drug Administration’s list of foods “generally
regarded as safe”. Kokum butter is better known as an ingredient in
skincare products, but it is edible, if high in fat, and is sometimes
used in Indian cooking. At the time, Zahn claimed sabotage by an
“informant” dairy producer. “At the very least,” he told me when I asked
him about the late-stage disqualification, “it’s incredibly
unprofessional of them.”
Climax,
which was based in Berkeley, pitched itself as the kind of foodmaker
only Silicon Valley could produce. It was a “data-science company
innovating the future of food”
Behind
the scenes, the company was struggling to raise money and was laying
off staff. By the end of the year, Climax was down to a handful of
employees. The roll-out of Bel products using Caseed, which was promised
for the end of 2024, never happened. I spoke to some former employees
who told me that even as financing dwindled, it felt like Zahn was
trying to do too many things at once. Priorities seemed to shift week to
week, based on the latest advice from some adviser or star chef. One
former staffer said the team was in a “civil war” about which direction
to pursue. Fine cheeses, such as Blue, had impressed investors. But
crowd-pleasers such as mozzarella were more likely to sustain the
business. Back when raising money was easier, making several products
felt more feasible. Another former employee said it seemed like Zahn was
waiting for a “white knight” investor that would allow Climax to
continue doing everything at once.
There
was still some interest from investors. In October a fund called S2G
Ventures led a $6.5m round, but with one condition: that Zahn step down
as chief executive. As the market for plant-based foods shrank,
investors wanted someone focused on making products cheaply and
efficiently—rather than a brilliant inventor. Sandeep Patel, a former
investment banker who’d been advising Climax, would become CEO, with
Zahn shifting to a less business-facing role.
The
company rebranded. Climax Foods would now be known as Bettani Farms.
Bettani evoked bucolic Italian pastures and was, said Patel, a riff on
Zahn’s boast that their cheeses would be “better in every way”. When I
spoke to Patel over a video call in October I asked him why the name had
changed. “I think you can answer that question yourself,” he said. When
pressed, he admitted he thought the name sounded too sexual. “We’ve
been at conferences where people come up to us and say, ‘What are you
working on, a Viagra replacement?’” They’d also consulted a branding
expert who told them that Climax was a tech-forward name that might have
appealed to early investors, but meant little to shoppers. People
didn’t want to eat food served by someone in a lab coat, said Patel.
They want food served by Martha Stewart.
https://cdn.economistdatateam.com/videos/1843/vegan-cheese/5_1280.mp4
The
firm decided to stop production of Climax Blue and focus on more
popular cheeses such as feta and mozzarella. In a press release Patel
promised to “do for pizza what oat milk has done for coffee”. When I
spoke to him in October, Patel criticised “an inordinate amount of focus
on a very small area of the market” at Climax.
Gone were any mention of data science or AI, which had seemed so important to partners like Bel. Patel wouldn’t discuss how central AI
had been to the work at Climax. Former employees told me the issue had
caused some internal dispute, after journalists overstated the company’s
use of the technology following conversations with Zahn (Zahn denies
any misrepresentation on his part). In any case, Bettani Farms was now
emphasising its “naturally derived” ingredients.
Eventually
Zahn left the company completely. “I’m not at liberty to discuss,”
Patel said when I asked him why Zahn had left, “but it just wasn’t the
right fit.” He said the press—no offence—was too focused on the idea of a
hero entrepreneur. “It’s too much reliance, too much belief, that one
person is the saviour.” Innovation is ultimately a team effort, he
added. And even then, “just because something is a great idea and a
great product, doesn’t necessarily make it a great business,” Patel
said.
“You
have to have a compromise,” said Alex, a Climax employee who didn’t
want to give their real name. “I think [Zahn] never understood that.”
Alex suggested the company could have gone all-in on blue cheese,
contenting itself with being a small company for a small market. Another
former employee said the team grew overconfident when financing was
easy, and tried to do too many things at once. “I look back on it as a
missed opportunity,” they said. In an alternate reality, “we would have
this mozzarella cheese in Domino’s right about now.” Both agreed an
element of hubris was involved. “They all thought they would be the
kings of the world,” said Alex, “and it just never happened.”
Last
summer Zahn was on a cycling holiday in the Bavarian Alps with his son
when he flipped over his handlebars and landed on his head. He fractured
vertebrae in his neck but mercifully avoided paralysis. When I spoke to
him again in October he said he was focused on getting better and
planning his next steps, including a new AI-related venture. He wouldn’t tell me anything about it, only that it will enable “humans and AI
to discover together”. He was reluctant to talk about his departure
from Climax. But he maintained that his original strategy, the Tesla
Roadster approach, was the right one. “It was always about mass market,”
he said. He’s annoyed at recent statements from Bettani that suggest
otherwise, and he insisted that most resources had already been shifted
to mozzarella before he stepped down as chief executive.
People didn’t want to eat food served by someone in a lab coat, said Patel. They want food served by Martha Stewart
Vegan
cheese, meanwhile, is still waiting for its oat-milk moment. Cost
remains the biggest hurdle for companies experimenting with artificial
casein. Caseed’s patent was approved last year, but it is still
expensive to make, so using it as a substitute for milk protein would be
challenging. One workaround, according to a former Climax employee I
spoke to, may be to use small amounts of Caseed to give an edge to
conventional oil-and-starch formulations.
It’s
hard to escape the feeling that vegan cheese was marooned after the
plant-based food boom of the 2010s: capital was pouring in, but the
problem of how to make it affordable and delicious remained stubbornly
unsolved. Now, with funding harder to secure, that prospect looks even
more unlikely. You could argue that it offers a parable for the AI
bubble today. If the hardest technical problems, such as artificial
general intelligence, prove more challenging than investors expect, then
they too could be stranded if the bubble bursts.
Before
I left the Wageningen campus in the Netherlands, the lab team invited
me to see its not-quite-mozzarella for myself. They popped a few pieces
out of a silicon ice-cube tray and onto a table. It’s off-white, lightly
shiny. I picked it up, and it started to soften between my fingers. It
didn’t really smell of anything, but it felt decidedly squishy, just as
mozzarella should.■
Sam Colbert produces podcast series at The Economist
ILLUSTRATIONS BY ROB & ROBIN