The Credit Sells Before the Carbon Settles

Week 35  |  August 2026

A carbon credit is a promise about one specific tonne. It says that tonne of carbon dioxide was removed from the atmosphere, that it will stay removed, and that it would not have come out without the money that paid for it. Real, permanent, additional. Sell a credit before you can stand behind all three and you have sold a promise you cannot keep.

Kelp carbon has been selling that promise for years. Grow macroalgae, sink it to the deep ocean, book the drawdown, issue the credit. The trouble is that the science underneath each of those three claims has been quietly falling apart while the credits kept trading. The gap between the carbon a kelp project claims to remove and the carbon anyone can verify it removed is the arbitrage. This week we map that gap, and then we do the more useful thing: follow the buyers who already worked it out and moved their money somewhere else.

We wrote the post-mortem on Running Tide in May (Week 23). Running Tide was the sector's best-funded operator. It folded and told everyone the voluntary market had simply stopped buying. The collapse was the symptom. What follows is the diagnosis, and the trade that the surviving buyers made in response.

Start with the removal itself, because that is where the first leak is. A 2024 review in Science of the Total Environment pulled together what the field actually knows about seaweed farming as carbon removal. The finding that matters: the emissions from running a kelp farm, the vessels, the nursery, the moorings, the harvesting, can offset somewhere between 61 and 90 per cent of the carbon the farm sequesters. Even on farms designed specifically to maximise drawdown. You are not capturing a tonne. You are capturing a fraction of a tonne and burning most of the rest to do it.

A techno-economic study in Frontiers in Marine Science put a sharper number on it earlier still: a baseline kelp farm cleared an additionality rate of just 39 per cent, with nursery electricity the largest single deduction. So, before anyone argues about permanence, more than half the headline number is already gone to operations.

Then permanence, which is meant to be the whole point of sinking biomass into deep water. A blue carbon study out of Namibia, published in Deep Sea Research Part II in March 2025, looked at exactly this. Its estimate for how much macroalgal carbon actually gets buried and locked away from the atmosphere was between 1 and 11 per cent. The authors were blunt about the rest, recommending research into the ultimate fate and permanence of the carbon because the values and specifics remain debated. That is not a foundation you build a thirty-year offset on.

Additionality has a side effect nobody was pricing, too. Modelling by Wu and colleagues in Geophysical Research Letters last year found that harvesting macroalgae at scale strips nutrients the open ocean was using, suppressing phytoplankton production offshore. Some of the carbon you book nearshore is not new, it is relocated from carbon the ocean was already fixing, and the disturbance could persist for centuries.

Here’s the part that should settle the argument. In 2026 a team of industry and academic specialists set out to write a rigorous macroalgae carbon removal methodology, the accounting rulebook a registry needs before it can issue credible credits. They published their results in Frontiers in Climate. They did not finish it. Their own conclusion was that few field-scale studies exist, that no accepted crediting framework exists, and that getting one will likely need public money rather than voluntary-market revenue. The people trying hardest to build the rails said the rails cannot yet be built. That is the institutional gap, stated by the practitioners themselves.

Watch what the most disciplined buyer in carbon removal did while all of this was landing. Frontier, the advance-market-commitment vehicle backed by Stripe, Alphabet, Shopify, Meta and McKinsey, spells out its marine position in plain language: it buys ocean carbon removal only on abiotic pathways that use the ocean's inorganic bicarbonate system. Ocean alkalinity enhancement, direct ocean capture. Not biological seaweed sinking. Frontier has, in effect, drawn a line straight through the biotic kelp model and stepped over it.

The number attached to that line is the one to remember. In August 2025 Frontier's buyers committed USD31.3 million to Planetary Technologies for 115,211 tonnes of removal between 2026 and 2030, at roughly USD270 a tonne. That is what serious money looks like when it believes the measurement holds: high price, long tenor, abiotic chemistry. Microsoft made the same kind of move earlier, signing Ebb Carbon in October 2024 for an alkalinity-enhancement deal with options running to hundreds of thousands of tonnes over the course of a decade. The demand for durable ocean carbon did not disappear when Running Tide fell over. It walked across the room to the pathway you can actually meter.

The registries tell the same story if you read them as a credibility ladder. At the top of the biotic ladder sits nothing you can buy: Verra's framework for seaweed carbon projects remains under consideration, not approved, years after the concept note went in. One rung across, Puro.earth approved three marine methodologies in 2025, but they cover anoxic-basin storage and microalgae sinking, not the grow-and-sink kelp cultivation model. And where credits are actually being issued and verified, Isometric booked the world's first verified alkalinity-enhancement tonnes. The pattern is hard to miss. The credit-issuing edge of this market is abiotic. Biotic kelp is stuck one rung below the point where a buyer can transact with confidence.

The survivors are the tell. Cascadia Seaweed, once North America's largest kelp cultivator for agricultural products, has launched OceanNexus, an AI monitoring platform that turns underwater imagery into audit-ready ecological records. Read the positioning carefully and it is a pivot into measurement infrastructure, not a carbon-credit play. Kelp Blue, farming giant kelp off Namibia, states flatly that no formally recognised carbon credit methodology for kelp exists. Its sequestration has been independently measured, but the Gold Standard methodology it is co-developing with the Kelp Forest Foundation is not expected to be approved before 2027. Meanwhile it sells product, not offsets. 

Further out, the credit claims get shakier, not firmer. Gigablue sold 200,000 ocean-removal credits, more than half of everything transacted in the category in 2024, then spent 2025 fielding questions about what its sinking particles are actually made of. Carboniferous is cleared to run a permitted pilot sinking crop waste into an anoxic Gulf basin and drawing formal objections over ocean-dumping law. Seafields has repositioned toward turning Sargassum into biostimulants, bioplastics and biochar rather than banking its sequestration. The direction of travel among the credible operators is consistent. They are backing away from selling the sink and toward selling the biomass, the monitoring, or the product.

This points at the redirect. The sequestration literature itself keeps arriving at the same conclusion: kelp is worth more as a material than as a carbon sink. Chopin and colleagues put it in a 2024 title, calling deep-ocean seaweed dumping questionable, risky, and not the best use of valuable biomass. The short version for anyone tracking procurement: the defence primes carrying binding SBTi commitments do not need kelp offsets they cannot verify. They need bio-based materials they can put in a supply chain. Both the science and the capital point the same way, and it is not toward the sink.

The constraint on ocean carbon was never photosynthesis. Kelp grows. It always did. The constraint is a measurement-and-certification pipeline that does not exist, which is precisely why credits keep selling before the carbon settles and precisely why the disciplined buyers priced the risk, walked, and put their tonnes into abiotic chemistry they can meter. Kelp's climate value is real. It has just been sold into the wrong market, at a price the accounting cannot support. The operators working that out first are the ones quietly rebuilding as materials and monitoring companies. The ones still selling the sink are selling the arbitrage, whether they mean to or not.

Five separate acquisitions of undersea and autonomy specialists in roughly a month, on both sides of the Atlantic, with Lockheed and Ultra the newest. The bet is that autonomy scales by owning the whole stack under one roof: sensors, navigation, neutralisation, hull. The counter-model says stay asset-light and partner. Neither technology nor capital is the constraint here. Industrial structure is. The open question is whether acquired engineering speed survives contact with a prime. 

Since you have been, thanks for reading.

Cheers,

Mick

Ocean Tech Intelligence provides informational analysis only. Nothing in this publication constitutes financial, investment, legal, or strategic advice. Readers act on this content at their own risk. For full details see our Disclaimer.