A wastestream that was probably useful

The problem is that the established hydrocolloid industry (the gel and thickener producers working with species like gracilaria and kappaphycus across Southeast Asia) has been solving the same narrow question of how you get the highest volume of extract at the best concentration out of your raw material for decades. That’s it, that's the whole optimisation target. Everything about how these operations are set up, where they're located relative to processing, how they dry their material, follows from that one question.

It shows in the economics too. The farmers and growers, the people actually doing the cultivation, capture almost none of the value. Nearly all of it sits with whoever's doing the extraction. That gap between growing and processing was one of the first things that told me there was a real problem worth solving here, separate from anything about the biology.

But the part that actually changed how I thought about the business was watching what happens to the material after extraction. Once you've pulled the gel out, you're left with a solid biomass byproduct. Everyone in the industry already knows this product has potential. Plenty of operators spread it on their own local land, informally, because on some level they knew it had some value. But it was never designed to be sold. It was never a viable product. It was a waste stream that people had a vague, non-monetised sense was probably useful.

The gap we built the second half of our business around

Instead of accepting the waste stream as a fixed cost of doing the extraction the traditional way, we redesigned the processing itself. This means zero chemical additives and low energy intensity so that both what comes out the other end are both product streams, not one product and one problem to dispose of. No harsh alkaline treatment, no perlite, nothing that would make the solid material difficult or unsafe to sell as a soil health product. We get a bit less of the traditional high-purity gel extract as a result. But we get two saleable outputs instead of one, and the second one turns out to solve a genuinely large, genuinely underserved problem: degraded soil, particularly around mine site rehabilitation, where synthetics don't actually move the needle on recovery rates and the cost of trucking anything in is often prohibitive anyway.

That reframe, stop asking "how do we get more extract" and start asking "what value is hiding in everything this process touches" is, I think, the actual transferable lesson here, more than anything specific to seaweed. If you're building anything bio-based or resource-processing, there's a good chance somebody upstream of you has already decided part of your raw material is waste, purely because nobody built a process around treating it otherwise. That decision is usually inherited, not re-examined. It was worth re-examining for us.

Engineering is the easy part

Which brings me back to engineering versus biology, because the two are more connected than they look. Designing that dual-output processing system was, honestly, the easy part. It's maths. It's controllable variables; energy inputs, chemical inputs, yield ratios, the kind of problem I was trained to solve and genuinely enjoy solving. Once you know what outcome you want from both streams, you can engineer backwards to get there with a reasonable degree of confidence.

The biology is a different animal entirely. We're cultivating native seaweed species in open water in Australia, which nobody has done successfully before at any scale. Our species are close relatives of the ones grown commercially in Southeast Asia, not identical, so a lot of the "established" knowledge doesn't transfer cleanly. And even where cultivation is established internationally, it's mostly built on decades of accumulated growing skill rather than a deep mechanistic understanding of why a given crop germinates or fails. There's no equation for that the way there is for extraction chemistry.

So we've built the business to absorb that uncertainty rather than pretend we can engineer it away. We plan for six harvests and budget as though only four will actually deliver. We're putting a meaningful share of early profit straight back into marine biology R&D — seasonal growth rates, direct seeding, what other species exist locally that our processing could also handle — because that's the area where the unknowns are still genuinely unknown, not just under-resourced.

If there's one thing I'd want another nature tech founder to take from my experience, it's to be honest with yourself about which parts of your business are engineering problems and which are biology problems, because they need completely different kinds of confidence. Engineering problems you can model, derisk, and largely solve with enough iteration and enough maths. 

Biology problems you plan around, you build buffer into, and you accept that some of what you're doing has genuinely never been done before. Conflating the two, assuming that because you've solved the hard technical process, the living system underneath it will just cooperate, I suspect is where a lot of nature tech companies get into trouble. We try very hard not to make that mistake twice.

It’s only byproduct until someone bothers to design around it

We didn't discover some novel compound or invent a new species of seaweed. We just refused to accept that half of what came out of our own process had to be a cost rather than a line of revenue. It's a fairly unglamorous insight as far as deep tech goes, but it's doubled the number of ways we can make this business work, and it means our soil product and our gel extract now fund each other rather than one quietly subsidising the disposal of the other. If you're sitting on a process with an output nobody's bothered to sell, that's usually not a sign there's no market for it. It's a sign nobody's redesigned the process with that market in mind yet.