Every factory has a back door. Out of it flows the stuff nobody ordered: warm water, steam, sludge, ash and gases. For most of industrial history, that back door led to a smokestack, a river or a landfill. Industry paid twice, once to buy raw materials and again to get rid of what it didn't use.

But what if one plant's back door opened onto another plant's front door? That's the idea behind industrial symbiosis, where neighbouring businesses trade waste heat, water and by-products as inputs. A small Danish harbour town has been doing this for over half a century. Its story shows how we might design industry differently.

Kalundborg sits on the west coast of Zealand, roughly 100 kilometres from Copenhagen. In the early 1960s, a new oil refinery there needed water. Groundwater was scarce, so the refinery piped in surface water from nearby Lake Tissø instead. Other companies soon tapped into the same supply. Nobody called it symbiosis. It was simply sensible.

Over the following decades, the deals multiplied. The Asnæs power station began selling steam to the refinery and to Novo Nordisk's pharmaceutical plant. Its surplus heat warmed homes through the town's district heating network. The station's flue-gas scrubbers produced gypsum, which went to a plasterboard factory next door. Fly ash went into cement. Yeast left over from insulin production was turned into pig feed.

What's striking is that no one drew this map in advance. Each link was a deal between two companies: one had a disposal problem, and the other saw a cheaper input. Only in the late 1980s did people step back and realise they had built a network. Today it involves dozens of exchanges, and the partners report steady savings in money, water and emissions.

Takeaway

Resilient systems often grow from many small, practical decisions rather than a grand design. The real skill is creating conditions where good connections are easy to make.

In 1989, two General Motors researchers, Robert Frosch and Nicholas Gallopoulos, published an article in Scientific American arguing that industry should behave more like an ecosystem. In a forest, there's no such thing as waste. A fallen leaf feeds fungi, fungi enrich the soil, and the soil feeds trees. Outputs loop back as inputs. This vision became known as industrial ecology.

Applying it is an engineering puzzle of matching. Every process has an output profile: what comes out, and at what temperature, purity, volume and timing. Steam that's too cool to spin a turbine efficiently can still heat fermentation tanks perfectly well. A by-product with trace impurities might be useless to one buyer and ideal for another. Engineers map these flows and look for pieces that fit.

This echoes Amory Lovins's idea of whole-system design: optimise the system, not just its parts. A single factory squeezing out efficiency gains is good. A cluster where each plant's leftovers feed another can reach savings no individual plant could manage alone. The question shifts from "How do we dispose of this?" to "Who could use this?"

Takeaway

Waste is often just a resource in the wrong place, or one that hasn't found its buyer yet.

Steam doesn't travel well, and neither does warm water or bulky sludge. Pipes are expensive, and heat leaks away with every kilometre. That's why symbiosis is fundamentally local. Kalundborg's partners sit within a few kilometres of each other, close enough that exchanges are cheap enough to be worthwhile.

Closeness alone isn't enough, though. If you build your process around a neighbour's by-product, you depend on them staying in business, running steadily and not changing their chemistry. That's a real risk. In Kalundborg, plant managers knew each other in a small town where reputations mattered. That trust made shared pipelines feel less like gambles. When the power station later moved from coal towards biomass, its partners had to adapt alongside it.

Can this be planned from scratch? Many eco-industrial parks have tried, with mixed results. A few lessons stand out. Choose tenants whose inputs and outputs complement each other. Build shared water and heat infrastructure early. Fund a neutral coordinator who helps partners keep talking. You can design the pipes. Trust takes longer.

Takeaway

Infrastructure can be engineered in a few years, but the relationships that make it work are built over decades. Good industrial design plans for both.

Kalundborg isn't a perfect model. Its network grew up around fossil fuels, and some links have faded as industries changed. But it proves something important: waste is a design choice, not an inevitability.

As economies decarbonise, new streams will appear, such as waste heat from data centres, captured carbon and by-products from hydrogen production. The towns and industrial zones that thrive may be the ones that ask early: who are our neighbours, and what could we trade?