Can Seaweed Really Remove Carbon?

Seaweed grows fast, needs no fertiliser, no freshwater and no land, and absorbs carbon dioxide while it does it. Few climate propositions have ever looked more intuitively appealing, and few have attracted investment faster. Over the past five years, companies have sold carbon credits for growing seaweed and sinking it, backed by some of the largest corporate buyers in the voluntary market.

The science has been moving in the opposite direction to the money. A 2024 commentary in One Earth, signed by twenty scientists across 23 institutions, called for a moratorium on sinking seaweed in the deep sea and described the practice as a dangerous fallacy amounting to blue washing. In May 2026 a modelling study in Nature Communications went further still, concluding that once iron limitation is accounted for, there is effectively nowhere in the ocean suitable for growing kelp for carbon dioxide removal.

That is an unusually blunt scientific position, and it deserves to be understood rather than simply reported. The carousel's framing is exactly right: absorbing carbon and storing it permanently are two entirely different things, and almost all of the difficulty lives in the gap between them.

 

Seaweed Is a Natural Carbon Absorber

 

The first step is not in dispute. Seaweed photosynthesises, drawing dissolved carbon dioxide out of seawater and building it into tissue. Kelp forests are among the most productive ecosystems on Earth, rivalling tropical rainforests in growth rate per unit area, and wild macroalgae globally fix an enormous quantity of carbon each year.

What happens next is where the question begins. Most of that carbon is eaten, respired and returned to the water within weeks or months, which is a normal and healthy part of a marine food web but is not carbon removal. A smaller fraction escapes: fragments detach and drift, dissolved organic carbon leaches from the tissue, and some of this material is carried off the continental shelf into deep water. Carbon that sinks below roughly 1,000 metres, beneath the permanent thermocline, is isolated from the atmosphere for a century or more, and that depth threshold is what the word sequestration actually means in this context.

The quantities involved in natural export are real but hard to measure, and crucially they are already happening. Wild seaweed has been doing this for millions of years. Any claim that farmed seaweed removes carbon must therefore show that it adds to this flux rather than substituting for it, which is the additionality question that haunts every nature-based carbon claim and is unusually hard to answer in a moving fluid.

 

Absorbing Carbon Is Not Enough

 

The carousel's third slide identifies the crux. Carbon in a living kelp frond is stored for as long as the frond lives, which is a single growing season. When the seaweed dies and decomposes, microbes respire most of that carbon back into dissolved inorganic form, and if that happens in surface or mid water it re-equilibrates with the atmosphere on a timescale of years to decades. Nothing has been removed. The carbon simply took a detour.

For a durable climate benefit the carbon has to end up somewhere that does not exchange with the atmosphere: deep ocean water, deep-sea sediment, or a long-lived product. Each route has its own problems. Burial in sediment is slow and localised. Long-lived products exist but the markets are small. Which leaves the deep ocean, and the proposal to put the seaweed there deliberately.

 

Can We Sink Seaweed for Carbon Storage?

 

The concept is simple enough to explain in a sentence: cultivate seaweed at scale, then sink the harvested biomass to the deep ocean, where the carbon is expected to remain for centuries.

Every word of that sentence has been contested. The One Earth authors argued that large-scale permanent sequestration on climatically relevant timescales is highly doubtful, that the fate of biomass during the different phases of its descent is poorly understood, and that the only plausible exception might be small areas of permanently low-oxygen water where decomposition is suppressed, assuming sinking to those depths were even technically feasible. Biomass that decomposes on the way down releases its carbon at intermediate depths, where it can return to the surface within decades rather than centuries, and nobody can currently verify where in the water column a given tonne of kelp actually breaks down.

The ecological objection is separate and arguably stronger. The deep sea is the largest habitat on the planet and among the least studied. Depositing millions of tonnes of fresh organic matter into an environment adapted to a slow, sparse rain of particles would consume oxygen, alter sediment chemistry and change benthic communities in ways nobody has characterised. The authors' point is that we would be conducting an uncontrolled experiment on ecosystems whose biodiversity and functions we cannot yet describe, in order to claim a carbon benefit we cannot yet measure.

The scale numbers make the trade worse rather than better. Harvesting one gigatonne of seaweed carbon a year would require more than a million square kilometres of highly productive exclusive economic zone waters in the equatorial Pacific. Earlier work in Nature Plants put the cost of sequestration by sinking at around 480 dollars per tonne of CO2, against roughly 50 dollars per tonne of CO2 equivalent in profit from using seaweed to displace land-based crops. Sinking is both the most expensive use of the biomass and the least verifiable.

 

The Nutrient Problem Nobody Priced In

 

The 2026 Nature Communications study is the most consequential development in this debate and it attacks the proposition at its root, before the question of sinking even arises.

Seaweed does not create nutrients. It takes nitrogen, phosphorus and iron from seawater, and in doing so it competes with phytoplankton, which are already using those nutrients to fix carbon. If farmed kelp simply takes nutrients that phytoplankton would otherwise have used, the net addition to carbon uptake is far smaller than the kelp's own growth suggests, and under some conditions it is negative.

The study modelled 25 years of kelp cultivation across global exclusive economic zones using a modified NEMO-PISCES ocean biogeochemical model, then ran 50 years beyond cessation, comparing an unconstrained case against nitrogen and phosphorus limitation and then against nitrogen, phosphorus and iron limitation together. The results narrow sharply at each step. Nitrogen and phosphorus limits alone reduced cultivation potential to roughly 15 percent of the unconstrained figure. Adding iron cut it roughly threefold again, to about 5 percent. The share of global exclusive economic zones classed as high potential fell from around 20 percent to 3 percent.

Peak removal flux fell from about 1.83 gigatonnes of CO2 a year unconstrained to 1.66 with nitrogen and phosphorus limits and 0.73 once iron was included. Removal efficiency, measured 25 years in, fell from 82 percent to 64 percent to 24 percent across the same three cases. Most strikingly, with iron limitation included, around 20 percent of the ocean became net negative, meaning cultivation there reduced total ocean carbon uptake rather than increasing it, with some eastern Pacific upwelling regions reaching roughly minus 50 grams of CO2 per square metre per year. Global phytoplankton primary production declined by up to about 8 percent.

The authors' conclusion is that seaweed only helps if it uses nutrients more efficiently than the phytoplankton it displaces, and that on their modelling it generally does not. Their summary judgement, that there is nowhere in the ocean suitable for growing kelp for carbon dioxide removal once iron is considered, has one small exception in a Senegalese upwelling area. They also note their estimates are probably optimistic, since the model does not account for remineralisation of the harvested biomass.

Two honest caveats belong here. This is a single modelling study at coarse resolution, and it does not explicitly represent seaweed biomass or fine-scale coastal processes. Its own sensitivity analysis found efficiency ranging from minus 43 percent to plus 78 percent depending on assumed nutrient demand and uptake kinetics, which is a very wide band. It is not the last word. But it is the most complete treatment of the nutrient question to date, and the direction of its correction, that previous estimates were too generous because they ignored a limiting nutrient, is the kind of finding that tends to survive.

 

What Happened When Someone Tried

 

The commercial history is instructive. Running Tide was the most prominent ocean carbon removal company in the world, backed by Shopify, Microsoft and Stripe, selling credits for sinking biomass in the North Atlantic. It shut down in June 2024 and laid off its entire staff, having been unable to build a business even with those buyers on its books.

The lesson is not that the founders were dishonest. It is that the combination of high operating cost, unverifiable sequestration and a buyer base that was starting to ask harder questions about measurement proved commercially fatal. A carbon credit is a claim that a tonne of carbon was removed, and in the marine environment that claim is extraordinarily difficult to substantiate. Dye tracing, sediment traps and biogeochemical models can estimate where carbon went. None of them can confirm it for a specific batch of biomass at a price the credit is worth.

 

The Potential Is Real, the Proof Matters

 

The carousel's closing slide lists three requirements, and each has a specific technical meaning.

Measuring carbon storage means demonstrating, for a particular operation, how much carbon reached below the permanent thermocline and how long it will stay there. That requires monitoring infrastructure in the deep ocean that largely does not exist, and modelling that can be independently audited.

Tracking environmental impacts means the nutrient competition question above, the deep-sea ecological question, and the effects of large cultivation arrays on light, currents and local productivity. The 2026 study shows this is not a secondary consideration appended to a carbon calculation. It changes the carbon calculation.

Accounting for farming emissions means the vessels, the rope and buoys, the seeded line, the hatchery energy and the sinking operation itself. These are the easiest to quantify and the easiest to overlook, and for a marginal carbon claim they can consume a large share of the benefit.

What survives all this is still substantial, and it is worth being clear that criticism of seaweed carbon credits is not criticism of seaweed. The One Earth authors, having called for a moratorium on sinking, argued in the same piece for expanding seaweed use on its merits: substituting seaweed-derived materials for higher-carbon terrestrial biomass, using it in food and feed, and deploying it to draw excess nutrients out of coastal waters and buffer local acidification. Those benefits are real, measurable and immediate. They do not require anyone to prove where a carbon atom went.

So, can seaweed really remove carbon? It removes carbon dioxide from seawater continuously and at enormous scale, and that is not in doubt. Whether any of that constitutes durable, additional, verifiable carbon removal that someone should be able to sell is a different question, and the current scientific answer is closer to no than the market has been pricing. The potential is real. The proof, so far, is not.

 

Note: This article reflects the state of the science as of October 2026, drawing on sources including Berger et al. in Nature Communications (2026) on nutrient limitation and seaweed carbon dioxide removal efficacy, the 2024 One Earth commentary calling for a moratorium on deep-ocean seaweed sinking, the 2022 Nature Plants assessment of economic and biophysical limits to seaweed farming, and reporting on the closure of Running Tide. Modelling results in this field carry wide uncertainty bands and should be read as directional rather than definitive. A companion piece in this series covers the commercial scaling of seaweed farming.