The Cows on the Water

I'm lucky, in Sussex to be close to some innovative per-urban rewilding and growing projects - like Hackhurst farm in Polegate, Lovebrook farm and Greenlands. I've been looking recently for even MORE innovative peri-urban growing models, and this article covers one of my favourites (though its not strictly in line with the rest of me 'rewilding' writing - other than generally being about the reintegration of animal life into city centres) - its success and its downfall.

Among the cranes and container stacks of Europe's largest port, on a three-storey platform that rises and falls two metres with the tide, a small herd of red-and-white dairy cows has spent the last seven years being milked by robots above the brackish water of the Maas. The Floating Farm in Rotterdam is one of those images that lodges in the mind: cattle grazing a patch of quayside, then ambling up a gangway to a barn that floats. It has drawn film crews, academics and, this year, Jeremy Clarkson. And now it is closing, its 28 cows sold, its owners blaming red tape.

It is a story tailor-made for a certain kind of outrage: the green pioneer strangled by mindless bureaucracy. That version is doing the rounds, and there is real truth in it. But it is not the whole truth, and this series exists to tell the whole one. So this piece asks three honest questions. What was actually clever about the Floating Farm? Is it, as it is often called, a model for urban rewilding? And what does its closure really teach us, not about cows on water, but about why so many green experiments in our cities quietly die, and how the rules could be written so they do not.

What the Floating Farm got right

Give it its due, because it earned a good deal. Peter and Minke van Wingerden built the thing in 2019 for around four million euros to prove a simple provocation: that food could be produced in the heart of a city, close to the people who eat it, without swallowing farmland. The design is a genuinely elegant loop. The cows are fed largely on the city's leftovers, spent grain from Rotterdam breweries, wheat bran, potato peelings, grass cut from sports pitches and parks. A robot scrapes their manure within a few hours and separates it, which the owners say cuts emissions by around sixty per cent, and the dry fraction is pelleted into fertiliser that goes back onto city parks and playing fields, whose clippings then return as feed. Rainwater is harvested off the roof and port water desalinated using heat drawn from the manure. Milk becomes yoghurt, butter and gouda in the same building, sold a short walk away.

Two things here are worth carrying forward whatever happens to the farm. The first is circularity: a system designed so that one process's waste is the next process's input, the principle at the heart of the permaculture piece. The second is climate-adaptive design. In a country a third of which sits below sea level, a farm that floats is a farm that keeps producing when the water rises. That is not a gimmick. It is a serious answer to a serious question about food in a flooding world, and delta cities from Dubai to Singapore have been watching.

But is it rewilding? An honest answer

Rewilding means stepping back and letting wild ecological processes return: the grazing, the predation, the messy self-willed recovery of a system we stop micromanaging. The Floating Farm is close to the opposite. It is intensive, high-technology agriculture: 28 cows on a steel pontoon, milked and mucked by machines, fed a managed diet, producing a product. There is nothing wild about it, and there is not much habitat in it either. And the honest reckoning goes further, because dairy is one of the higher-impact foods there is. A floating cowshed is a wonderful piece of engineering and a poor emblem for lightening our footprint, when the lower-impact urban food is almost always the plants.

So hold two thoughts at once. The Floating Farm is not urban rewilding, and pretending otherwise would be wrong. But its life and its death still have a great deal to teach anyone who wants to bring nature back into a city, because the thing that is killing it is the same thing that kills community meadows, verge planting, beaver releases and rain gardens. That thing is almost never ecology. It is governance.

Why it is really closing

Two threads, and both deserve a fair hearing rather than a slogan.

The first is a floor. The barn's original low-emission flooring, the surface that made the fast manure separation possible, wore out after a few years. It began to bulge and lift, which was dangerous for the cows and stopped the cleaning robot running. The owners bought a replacement, manufactured in Germany, that was not certified for use in the Netherlands. They believed their own checks showed it kept emissions within the legal limits, and assumed that would be enough. The local environmental regulator disagreed and imposed sanctions. From the outside this looks like pure pedantry: a performing floor rejected on a paperwork technicality.

The second thread is time and place. The farm sits on a quay it rents from the Port of Rotterdam, and its right to stay is running out. The owners say they have little more than a year of permitted operation left, with the lease and municipal permits expiring by the end of the decade and, they say, no extension available for reasons they find unclear. Faced with only a short window remaining, they were not willing to spend tens of thousands of euros ripping out and relaying a certified floor for a farm they must soon vacate anyway. So the cows go.

Here the honest account requires the other side, which the outrage version omits. The Port of Rotterdam Authority says the Floating Farm was always meant to be temporary, granted a ten-year exemption, and that a dairy farm is not a sensible permanent feature of a working industrial port. Read that way, this is not a pioneer being forced out so much as a time-limited pilot reaching its planned end, with a worn-out floor arriving at an awkward moment. There is also political noise: the Party for the Animals has suggested the owners are closing because they want out of animal farming, a claim the van Wingerdens firmly reject. The truth sits somewhere in the overlap of a genuine regulatory obstacle, a lease that was always finite, and an expensive repair that stopped making financial sense.

The rule that looks stupid, and is not (quite)

The certification requirement is the part worth dwelling on, because it is where the easy story is most wrong and the real lesson is richest.

The Netherlands is in the middle of a nitrogen emergency. Ammonia and nitrogen oxides, much of it from livestock, have been degrading the country's protected nature for years, and in 2019 its highest court struck down the government's permitting scheme for breaching European habitat law, freezing thousands of building and farming projects overnight. In early 2025 a court in The Hague ordered the state to cut nitrogen emissions hard by 2030 or face fines. This is the backdrop against which a barn floor's emission performance is not a trivial detail but a legal and ecological one.

And here is the uncomfortable fact for the red-tape narrative. One of the main reasons the certification regime is strict is that low-emission barn floors have a poor record of delivering in practice what they promise on paper. Environmental groups and courts have repeatedly criticised them as a technological false hope, cutting emissions far less in the real world than their specifications claim. A rule that says "you cannot simply self-certify a floor's performance, it must be independently verified" is not bureaucratic spite. It exists precisely because "trust our own checks" is exactly the situation that has let unverified green claims through before.

So the honest verdict is a genuine both-and. The owners may well be right that their particular floor performs. But a system that has been burned by unverified floors cannot reasonably take that on trust, and they are caught by a rule with a real and defensible purpose. The failure is not that the rule exists. It is that the rule is built as a list of approved products rather than a test of actual outcomes, so a floor that genuinely works but is not on the list has no quick route to acceptance, and an innovator gets treated like a cheat. That distinction, list versus outcome, is the whole lesson.

What urban rewilding should take from this

Because it is the same trap, again and again. The obstacles to bringing nature back into British towns are rarely biological. The wildflowers will grow, the swifts will nest, the beavers will build. What stops them is administrative: the verge mown to stubble because the contract was written for tidiness and never updated; the community garden on a two-year meanwhile lease that no one dares plant a tree in; the sustainable drainage scheme that falls foul of a building rule written for a drier decade; the pilot that succeeds brilliantly and then expires because success was never given a path to permanence. The Floating Farm died of the temporary-pilot trap and the approved-list trap. Our urban rewilding dies of the same two things.

Policymaking could do better, and not by abandoning regulation, which in the nitrogen case is protecting genuinely threatened habitats. It could do better by getting smarter in a few specific ways.

Regulate for outcomes, not products. Set the standard as the result you want, the emission cut, the biodiversity gain, the water held, and let anything that measurably meets it qualify, rather than blessing a fixed list of approved solutions that innovation immediately outruns.

Build proper sandboxes. Give green experiments a defined space to operate under close monitoring, where a novel approach can prove itself without first satisfying rules designed for the thing it is trying to replace.

Give tenure that can last. Short, non-renewable leases on marginal urban land guarantee that nobody invests in the slow work nature needs. A pilot that proves its worth should have a route to staying, not a built-in eviction date.

Design pilots with an exit into permanence. Fund and permit demonstration projects with the question "and then what?" answered from the start, so a success becomes policy rather than a fondly remembered one-off.

Join the departments up. So much green ambition dies in the gap between highways and parks, planning and environment, the port and the city. A standing presumption in favour of nature on suitable land, shared across departments, would stop good projects falling down the cracks between them.

The same trap inside organisations

There is a version of this for anyone who runs anything. Organisations kill their own innovations the same way cities do: they set up a skunkworks or a pilot, it works, and then it dies, because there was never a path from experiment to core business, and because it was judged against the checklist for how the old thing was done rather than the outcome the new thing achieved. The Organisation as Ecosystem strand takes this on more fully, but the pattern is identical: a promising thing on a short lease, measured by the wrong yardstick, with no route from proof to permanence.

When the cows leave the water

The Floating Farm is not strictly rewilding. But we can learn from it precisely, which is more useful than mourning it loosely. It was a clever, circular, climate-adaptive experiment that ran into a wall built for a different problem, at a site it was only ever lent, with a repair that stopped adding up. None of the walls it hit were made of biology. They were made of rules: rules with real purposes, written in ways that could not tell a genuine innovation from a dodge, on land offered only for a while.

That is the quiet truth under most failed attempts to let nature back into our cities. The plants are willing. The animals are willing. What we have not yet built is a way of governing that can recognise a good experiment and let it live. Write those rules for outcomes, lend the land for longer, and give what works a way to stay. Then the next surprising thing on the water, or the verge, or the rooftop, might not have to leave.

References and further reading

• DairyReporter (2026). "Red tape sinks world's first floating dairy farm." 5 August 2026.

• Food & Agribusiness (2026). "Floating farm shuts down due to innovative barn floor." August 2026.

• Farmers Guide (2026). "First floating dairy farm put up for sale amid regulatory issues." August 2026.

• Agri-View / Chris McCullough (2026). Report on Floating Farm closure, emissions regulations and lease deadline.

• CityChangers.org; CleanTechnica (2023); Euronews (2024); VOA News (2021). Background on Floating Farm's design, circular systems and operations.

• Animal Science Cases (CABI, 2026). Case study framing Floating Farm as a complementary, experimental model rather than a replacement for land-based dairy.

• District Court of The Hague (January 2025), Greenpeace Nederland v. the State of the Netherlands: order to cut nitrogen emissions by 2030. Reporting via AP, Loyens & Loeff, Irish Examiner.

• FiscalNote (2026). "Farms to Courts: An Overview of the Dutch Nitrogen Crisis," including the 2019 collapse of the PAS scheme under EU Habitats Directive rulings.