Insurance Will Restructure Europe’s Food System Before Regulation Does.

On the 8th of May 2026, a closed-door conversation in Zürich brought together food-system investors, founders, and operators to map the vulnerabilities in Europe’s food supply, and to be honest with each other about how close several of them are to breaking.

The conversation was hosted by Claire Smith of Beyond Impact, the venture firm investing in alternative-protein and sustainable food companies. It produced one of the most concrete and least comfortable diagnoses of Edition I, an event that brought together more than 160 decision-makers representing over $100 billion in assets under management.

This essay starts from that conversation, then builds outward. The diagnosis and the core argument are what the room produced. The supporting detail, the named companies, the figures, the policy context, comes from research conducted afterward to test and extend what was said. Where a claim is attributed to “a participant” or “the room,” it reflects something said in the conversation. Everything else is verified independently. It is for investors, founders, and policymakers who still treat food-system resilience as a future problem rather than a present one.

It argues three things.

One. Insurance is a binary forcing mechanism, and it will move the European food system faster than regulation can. Two. The system has seven acute vulnerabilities that are already being tested, and they are precisely the risks insurers price. Three. Consumers say they want sustainability but do not pay for it, so the winning products lead with price and protein.

The forcing mechanism nobody is watching

The room’s most original argument, and the one this essay is built around, concerned how the European food system will actually be forced to change. The answer was not regulation. It was insurance.

The reasoning is that insurance is binary in a way regulation is not. Regulation is negotiated, phased, delayed, and softened; a compliance cost can be absorbed, lobbied down, or deferred. Insurance is different. No serious farm, processor, or food business operates without it. And when an insurer decides a risk is no longer acceptable, or reprices it sharply, the effect is immediate and commercial. There is no phase-in and no consultation period. The cover is there or it is not, and if it is not, the operation stops, because you cannot run a regulated food business, or borrow against it, without insurance.

This makes the insurance industry a faster and blunter forcing mechanism than any regulator. As climate volatility raises the risk attached to certain farming practices, certain crops, and certain geographies, insurers reprice or withdraw. The moment coverage for a degraded or high-risk operation becomes unaffordable or unavailable, the economics of that operation collapse, regardless of what the regulation says. An insurer does not need a parliamentary majority, a transition period, or a consultation with stakeholders. It needs an actuary to change a number.

And the numbers are already changing. Agroseguro, the pool that insures Spanish farming, paid out €1.24 billion in 2023, a 56% rise on the year before and the worst year since the system was founded in 1978, with the drought alone producing the largest claim in its history. Agroseguro’s own analysis notes that droughts which used to return every seven to ten years, 1995, 2005, 2012, now arrive back to back: 2017, 2019, 2022, 2023. In 2025 the system paid out €804 million, the second highest in a decade, and the rating agency Morningstar DBRS expects claims to rise again in 2026. That is what a risk table being rewritten looks like, and it lands on a system with little slack. A study commissioned by the European Commission and the EIB put EU agricultural losses at €28.3 billion a year, with only 20 to 30% insured and farmers carrying the rest themselves. When cover reprices, the farmer’s choice is a higher premium, a thinner policy, or none.

The room’s insight was that anyone trying to understand where the food system will be forced to change first should watch the insurers, not the legislators. Insurance turns a slow-moving environmental risk into an immediate commercial one, and that is what actually moves operators.

If that is right, then the question that matters is: what will the insurers price? The answer is the seven cracks below. Read them the way an underwriter would. Each one is not just a vulnerability. It is a claim event waiting to be priced, a premium waiting to rise, or a class of cover waiting to be withdrawn.

What the insurers will price: seven cracks

The room laid out, without drama, the specific dependencies that make European food fragile. Start with the one that has already been tested, because it shows exactly how the mechanism works.

Disease. On 26 November 2025, two wild boar were found dead in Bellaterra, on the edge of Barcelona. Testing confirmed African swine fever, the first cases in Spain since 1994.

The suspected origin was mundane to the point of absurdity: a discarded meat sandwich, dropped near a motorway rest area, eaten by a boar. That explanation was later challenged. In December the Agriculture Ministry noted the strain resembled a lineage widely used in laboratories, opened an investigation into a possible leak from a nearby research centre, and police searched the facility. In February 2026 the Ministry’s report ruled the laboratory out on genetic sequencing and pointed investigators back toward contaminated food entering the country. Eight months on, the cause is still argued over.

The commercial damage never waited for the answer. Within days the case count rose past a dozen, the Spanish army was deployed, and a €8.8 billion export industry was in trouble because of an animal that had eaten someone’s lunch. Spain is the EU’s largest pork producer, roughly a quarter of all European output. Of its 400 export certificates covering 104 countries, about a third were blocked within a week. China, Japan and Mexico suspended imports. Live pig prices fell 16% in a single week, piglet prices 28%.

No farm was infected. Not one, then or since. Every positive case has stayed inside the containment zone, in wild boar. And the trade machinery is still grinding back eight months later: China kept buying by applying regionalisation to the province of Barcelona, while Spain reopens the rest market by market, certificate by certificate, the Philippines in May 2026, Honduras and Ecuador in June.

The biology was contained in weeks. The commercial system it triggered is taking the better part of a year to unwind. That is a loss event whose financial consequences ran ahead of its physical ones, and it is exactly the profile that reprices a market.

The other six cracks have not been tested like that yet. They are the same shape.

Animal feed. Europe imports the overwhelming majority of the soy that feeds its livestock, on the order of 97% of its soybean-meal protein. The EU Deforestation Regulation, enforced from the end of 2026, adds cost and complexity to that supply at the point of maximum dependence.

Micro-ingredients. Europe depends on China for roughly 70% of the vitamins and a similar or higher share of the amino acids that go into animal feed and processed food. For some individual vitamins, China’s share of world production runs well above 90%. These are not substitutable overnight, and the dependency sits at a chokepoint few consumers have ever heard of.

Fertiliser. Europe produces only a fraction of its own nitrogen fertiliser needs and leaned heavily on Russia and Belarus for imports, a dependency uncomfortable enough that the EU introduced escalating tariffs in 2025 to force a shift, at the cost of higher input prices in the meantime.

The cold chain. The refrigeration and processing equipment European food relies on depends on rare-earth elements and components processed overwhelmingly in China, which controls roughly 90% of global rare-earth refining, and which demonstrated during 2025 how readily that dominance becomes an export-control lever.

Energy. Food production is energy-intensive across fertiliser, processing, refrigeration and transport. The energy-price shock fed directly into food inflation, which across the euro area peaked near 15% in 2023.

War. Ukraine, one of the world’s great grain baskets, has seen wheat and other staple crop areas fall by roughly a third against 2021, removing a source of stability from European and global grain supply.

Now sort them the way an underwriter would, because the sorting is where the essay’s argument actually bites.

Three of these are already inside the policy schedule. Disease, extreme weather and energy volatility are priced, claimed against, and repriced every renewal cycle. This is where movement happens first, and where it has already started.

Two are priced indirectly, through business interruption and contingent supply-chain cover. Feed and fertiliser dependency do not appear on a policy as risks in their own right, but when they fail, the claim lands anyway. This is the layer where exclusions get quietly written, because insurers can see the correlation before they can price it.

Two are barely insurable at all. Micro-ingredient concentration and rare-earth dependency are political risks wearing industrial clothes. No actuarial table prices an export-control decision made in Beijing. The exposure does not transfer. It sits on the operator’s balance sheet permanently, which means the only available hedge is structural: another source, another supplier, or another input.

That is the map. Three cracks the insurers will price and are already pricing, two they will price at one remove, two they cannot price and will simply decline to carry. No single one is necessarily catastrophic alone. The room’s point was that they are correlated, they are simultaneous, and several are being tested in real time. Regulators respond to that with consultations and phase-ins measured in years. Insurers respond with premiums and exclusions measured in renewal cycles.

Consumers do not pay for sustainability

The third finding was a hard commercial truth that the room, several of whose members build and back sustainable food companies, stated plainly. Consumers say they want sustainability. Their purchasing behaviour says otherwise.

This is one of the most consistent findings in consumer research: a large majority express a preference for sustainable products, and a far smaller number actually pay a premium for them when they reach the shelf, especially under cost-of-living pressure. The stated preference is real, but it does not survive contact with the price tag for most shoppers most of the time.

The commercial implication is decisive for anyone building in this space. A product that leads with its sustainability credentials, and asks the consumer to pay more for them, is building on the weakest part of the consumer’s actual behaviour. The products that win lead with something the consumer will reliably pay for, price, protein, taste, convenience, and treat the sustainability as a built-in benefit rather than the headline ask. This is the same lesson that recurred across the food and materials rooms at Edition I: the sustainable proposition has to be delivered as a better or cheaper version of what the consumer already buys, not as a premium they are asked to pay for their conscience. Lead with protein and price. Let the sustainability ride along.

The position

The implication, for any investor, founder, or policymaker in European food, is that the fastest-moving force in the system is not the one everyone is watching. Brussels will consult, phase, and negotiate. The insurers will simply reprice. When cover for a high-risk crop, a degraded soil, or an exposed geography becomes unaffordable or unavailable, the operating model attached to it ends, whatever the regulation says. The system’s real transition timetable is being written in renewal cycles, not legislative calendars.

The rest follows from that. The seven dependencies are real, correlated, and already being tested, and they are exactly what underwriters price. The consumer will not pay a premium for sustainability, so the winning products have to lead with price and protein.

The alternative is to act on the system as it actually behaves. Watch the insurers for the earliest signal of where change will be forced. Treat the seven dependencies as current risks to be hedged now, not future risks to be studied. Build products that lead with what consumers actually buy.

The room was not unanimous on every point. There were disagreements on which dependencies are most urgent, and on how quickly the insurance mechanism will bite. There was no disagreement on the core diagnosis. Europe’s food system has multiple cracks, several are already showing, and the institution most likely to force the response is the one nobody is lobbying.

The next gathering is Edition II, on 11th of May 2027 in Zürich, and the meetup series opens this question long before then. The work between now and then is not the event. It is the architecture being built around it.

If your work intersects with what the room described, the door is open.

Ways in.

Join the meetup series. Launching this fall: focused gatherings on each of our key clusters, food-system resilience among them, ahead of Edition II. If you are addressing one of the seven cracks in your business or portfolio, reach out to Camille at camille@thedecisionroom.net and follow the calendar so you catch the food room when it lands. → Subscribe to our calendar

Apply for Edition II. This essay is one of eight from Edition I. Edition II is where the diagnosis becomes work. Founding access is open now, at the lowest rate we’ll offer. The full event is capped at 400. Each roundtable is capped at 20. → Apply for founding access

Partner the room. Partners shape the question. The room owns the answer. Co-host a roundtable, run the workshop that follows, or anchor a cluster across editions. → Explore partnership

Reply to this email if any of the workstreams above are yours. We will be in touch directly.

Aaron C. Leaman

Co-founder, The Decision Room aaron@thedecisionroom.net

The Decision Room is an invite-led climate and frontier-tech capital convening. Once a year. Zürich. This essay is the third in the series, one per closed-door conversation from Edition I. The full record is Before Capital Flows, essay written by Aaron C. Leaman.

You have sat in this meeting. The deck was strong. The data held. Everyone around the table agreed, and some of them even meant it. Then the meeting ended the way the last one did: warm words, no commitment, and a follow-up that will feel exactly like this one. Nobody was wrong. Nothing moved.

I have spent years inside that meeting, in climate, frontier tech and mission-led organisations, and I have stopped accepting the standard explanation for it. The standard explanation is that the room has not yet seen enough: not enough proof, not enough polish, not enough pages. So the team goes away and produces more of all three, and the next meeting ends the same way.

The problem is almost never the evidence. My work is finding the first point where belief breaks, and the decision sitting underneath that point that nobody has made yet. This essay is about that break, because the same one is failing at every scale that matters, up to and including the planet.

The pattern nobody names

Here is the pattern, stated plainly.

Climate science is arguably the largest evidence base ever assembled behind a single conclusion. The IPCC's most recent assessment of the physical science alone drew on more than 14,000 studies. A 2021 review of 88,125 climate papers found that the consensus among publishing scientists now exceeds 99 per cent. If evidence moved the world, this would be the most-moved-on subject in history.

It is not. And the distance is measurable: surveys by Yale's climate communication programme find that only around one in five Americans know the scientific consensus is even above 90 per cent. The evidence is settled. The belief never arrived. The world runs on belief, not just ideas, and belief is the one thing all that evidence has failed to produce at scale.

I started calling this the belief gap in my newsletter years ago. Communication scholars use the same term, and their core finding is blunt: on contested issues, ideology predicts what people believe better than education or information does. The same pattern runs through every room I work in. A founder with peer-reviewed results and a stalled raise. A first-of-a-kind facility that every expert agrees should exist and no one will finance first. A coalition where funders, leadership and partners all nod at the same deck and nothing signs. A ministry with the data, the mandate and a campaign that informs everyone and moves no one.

In each case, the people involved respond the same way. They assume the problem is that others have not yet seen enough. So they produce more: another report, another model, another appendix, another campaign.

And in each case the same uncomfortable truth applies.

Evidence is not belief. Closing that distance is the work.

Why more evidence rarely closes the gap

The instinct to add evidence is natural. Past a point, it is also remarkably ineffective, for two reasons.

The first is order. In most of the material I read, the number arrives before the case. The reader meets the figure that disqualifies you, the capex, the timeline, the ask, before they meet anything that would change their mind. The answer is usually in the material. It is in the wrong place.

The second is deeper. A list of facts engages the brain's language regions and largely stops there. A story recruits more of the machine: read the word cinnamon and the regions that process smell respond; read kick and the areas that move your leg fire. To the brain, a story is closer to an experience than to information about one. And there is a scaling problem underneath. As the numbers grow, the feeling shrinks: the psychologist Paul Slovic calls it psychic numbing, the reason one person in trouble moves us and a million becomes a statistic.

People cannot believe what they cannot first recognise as theirs. Recognition comes before persuasion, always. A wall of evidence built for a general audience recognises no one in particular, and so persuades no one in particular. This is why the most evidenced causes often have the most alienating communications: everything is proven and nothing is felt. And felt is what gets funded. That is not a slogan. A study of early-stage investors found that their gut read on the founder predicted returns better than their formal analysis did.

Governments run into this constantly. Public communication is usually built around one question: what should people know? It asks people to care before it gives them a reason. Communication research has a name for that assumption, the information deficit model, and it has been failing on the record since the early 1990s. The operative question in any room that decides anything is different: what would this specific room need to believe in order to move first? Those are not the same discipline. One informs. The other closes a gap.

The oldest version of this story

In 1847, Ignaz Semmelweis showed that when doctors washed their hands in chlorinated lime, deaths of new mothers in his Vienna clinic collapsed, from roughly one in five towards one in fifty. The evidence was about as clean as medicine ever gets. It was rejected for decades.

Not because the data was weak. Because of what believing it required: doctors accepting that they themselves had been carrying the infection. The evidence asked its audience to make a decision they could not face, so the audience declined the evidence. Medicine still calls the pattern the Semmelweis reflex.

I meet its smaller cousins every week. Which brings us to what is actually underneath.

What is actually underneath

In nearly every engagement, when I trace a communications problem back far enough, I find the same thing: a decision nobody has made yet.

It shows up as fragmentation. Three true descriptions of the same organisation, none of them the one, each held by a different person. Every deck, speech and page becomes a negotiation between versions. From the outside it reads as incoherence. From the inside it feels like a delivery problem, something a better writer or a better deck could fix. It is not. It is a depth problem: an unmade decision leaking out as messaging.

Organisations in this state have usually built the shallow version of their story six times instead of the deep version once. The shallow version is assembled outside-in, for each audience, under each deadline. It works until someone asks a real question. The deep version is built inside-out, from what is actually true and actually decided, and because of that it survives pressure. It is the same house with different doors, not a different house for every room.

The work, then, is rarely to write something new. It is to find the first point where what you know to be true stops being believed by the person who has to act on it, name the decision sitting underneath that point, and make it. Most of what is needed already exists. It is in the documents, in the wrong order, or in someone's own words and never written down.

The deadlock at full scale

The hardest version of the belief gap is not one sceptical reader. It is a deadlock: several parties want the outcome, each needs another to commit first, nobody is wrong, and nothing happens.

Capital waits for proof. Buyers wait for a reference. Partners wait for a commitment. Funders wait for alignment. No precedent means nobody wants to be the precedent. Meanwhile the clock runs anyway: budgets cycle, windows close, and the people who believed first begin to hedge. The money shows the same gap the surveys do: the world now deploys roughly 1.5 trillion dollars a year of climate finance against an annual need estimated at nearly six trillion.

This is the structure of the climate transition itself. I wrote last year that innovation is no longer the constraint; coordination is. The technology exists. The capital exists. The demand is emerging. What is missing, room after room, is the case that lets one specific party move first without feeling alone, because it was built around what they, specifically, need to believe.

We talk about collective action problems as if they were economics. Mostly they are narrative problems wearing an economics costume. A Nobel prize has already been won for a version of this argument: Robert Shiller's narrative economics holds that contagious stories, not fundamentals alone, drive booms, crashes and the decisions in between. And on the one occasion a global environmental deadlock genuinely broke, the ozone layer, it broke narrow: a vivid frame, a named set of gases, a specific industry, and first movers whose case for moving was concrete. Game theory says everyone waits. A story, held clearly enough, is how someone stops waiting. Every first mover in history moved on belief before the spreadsheet fully agreed, because a case existed that made moving first survivable.

What closing the gap looks like

I will give you the small, documented version, because small and documented beats grand and vague.

eeden is a German company building chemical recycling for textiles. Real science, real team, and the classic first-mover problem: fashion brands wanted the outcome and nobody wanted to be the precedent. We rebuilt the case around the believer rather than the evidence: what a brand partner would need to see first, in what order, to make moving first defensible inside their own building. Belief is built incrementally, not theatrically. eeden signed a letter of intent with one of the world's leading fashion brands. The work directly contributed to that outcome; the claim is exactly that, and no more.

The pattern transfers to any scale you care to name. Find where belief breaks. Name the decision underneath. Build the one story that survives every room it enters. Then put it in the room and watch what changes.

And what changes is observable. The meeting gets shorter, not longer. The questions move from scepticism to logistics. And the sentence you have been hearing for a year, who else is in, quietly becomes a different one: how do we move.

The work

If you hold evidence that is not being believed, the temptation will always be to add. More proof, more pages, more polish. Resist it long enough to ask the harder question first: where, exactly, does belief break, and what decision is sitting under that point, unmade? You can never ask a room to believe something you have not yet given it the reason to believe.

That question scales. It is the same question for a seed round and a sovereign transition plan. The most important communications work of the next decade, corporate, institutional and governmental, is not information. It is closing the distance between what we can prove and what rooms full of capable people actually believe, one findable break at a time.

Evidence is not belief. Closing that distance is the work. And it is work: findable, nameable, buildable.

The next version of that meeting can end differently.

Aaron C. Leaman works on complex communications problems in climate, frontier tech and mission-led organisations. The Belief Gap Diagnostic is where that work begins.