Captain Drawdown’s daily logbook on every CDR story, paper, and expert voice — so you don’t have to read them all.

The CDR industry doesn’t have a demand problem. It has a customer concentration problem. And Microsoft just made that painfully clear.

Robinson Meyer broke the news this week: Microsoft Is Pausing Carbon Removal Purchases, with the tech giant accounting for more than 90% of industry volume last year. Read that number again. More than ninety percent. When one buyer represents that much of your total market, you don’t have a market. You have a dependency. James Temple put it plainly: “MSFT is the carbon removal market, so if this is anything more than a brief pause, it’s a v. big deal & v. bad news for an already shaky sector.” - James Temple (@jtemple.bsky.social)

What makes this week so revealing is not just the Microsoft news. It’s what happened alongside it. In the same stretch of days, biochar and mineralization companies announced expansions, partnerships, and executive hires. The contrast is stark. And it tells us something the CDR community has been reluctant to say out loud: not all carbon removal business models are equally fragile. The ones selling a product people already want are pulling ahead. The ones selling only a tonne of CO2 removed are in trouble.

The patronage system cracks

Let’s be honest about what the voluntary carbon removal market looked like before this week. Microsoft’s massive advance purchase commitments, through its climate fund and Frontier-adjacent deals, created the illusion of a functioning market. Startups could raise venture capital by pointing to Microsoft as a credible buyer. Investors could model future revenue against the assumption that more Microsofts would follow. The whole theory of change was: one anchor buyer proves the concept, others pile in, prices come down, the market matures.

That theory just hit a wall. As Dirk Paessler noted, the CDR community is collectively processing this as a sector-defining event, not a routine business decision (@dpaessler.bsky.social). And the most exposed pathway is direct air capture, or DAC. The reason is structural. DAC’s primary output is a verified tonne of CO2 removed from the atmosphere. That’s it. There is no soil amendment. No building material. No agricultural product. The tonne itself is the product, and its value depends entirely on someone’s willingness to pay a premium for it.

When that someone pauses buying, the floor disappears.

Knowledge walks out the door

The downstream effects are already visible. Heatmap reported on a startup literally trying to salvage carbon removal know-how before it’s lost forever. The initiative, called Ctrl-S, is archiving institutional knowledge from DAC companies before they fold. Think about what that signals. The sector itself recognizes that a wave of DAC venture deaths is coming, and the priority has shifted from scaling to preservation.

This is the kind of knowledge-loss crisis that hits young industries hard. DAC engineering talent is scarce. The operational lessons learned from running pilot plants, from understanding sorbent degradation rates, from optimizing air contactors in real weather conditions, that knowledge lives in people’s heads and in company Slack channels. When a startup shuts down, that knowledge scatters. As we’ve explored in our analysis of how DAC can scale, the path to affordable direct air capture depends on accumulated learning-by-doing. Losing companies means losing the learning curve itself.

Even the DAC companies that survive are pivoting their message. Sustaera announced a pathway to 3x more affordable DAC technology, framing radical cost reduction as the new imperative. That’s an implicit acknowledgment: the old model of selling premium voluntary credits to a single megabuyer is broken. The next chapter for DAC, if there is one, requires costs low enough to attract compliance markets or industrial integration. Not another patronage deal.

Co-benefits as a lifeline

Now look at the other side of the ledger. Exomad Green and Beston Group deepened their biochar production partnership this week, expanding capacity. Biochar has revenue streams that exist independent of any carbon credit buyer. Soil remediation. Water treatment. Concrete additives. Agricultural yield improvement. The carbon removal is real and measurable, but it’s bundled with something a customer already needs. That changes the economics entirely.

Paebbl hired a former Tesla and Novo Energy executive to lead commercial expansion of its mineralization technology, which locks CO2 into building materials. The product is a construction input with a carbon removal co-benefit. You don’t need Microsoft to buy your credits when your customer is a concrete company. As we’ve covered in our enhanced weathering primer, mineralization pathways that produce useful materials have a fundamentally different risk profile than pathways that produce only removal credits.

And then there’s the JPMorgan deal. JPMorganChase signed a long-term strategic carbon removal partnership with Graphyte, a biomass carbon removal company. Even as Microsoft steps back, finance capital is stepping toward pathways with tangible co-products. JPMorgan is not doing this out of charity. They see a more bankable business model. One where the offtake risk is distributed across multiple revenue streams, not concentrated in a single buyer’s climate budget.

This pattern, biochar expanding, mineralization hiring, biomass locking in bank partnerships, is not coincidental. It’s Darwinian. The Microsoft pause is acting as a selection pressure, and the pathways with diversified revenue are the ones surviving it.

The uncomfortable truth

The CDR community has spent years arguing that all tonnes are equal. A tonne removed by DAC is the same as a tonne removed by biochar is the same as a tonne removed by enhanced rock weathering. From an atmospheric perspective, that’s true. From a business model perspective, it’s dangerously misleading.

A tonne that comes bundled with a product someone already buys has pricing power. A tonne that exists only as a credit has whatever price the last willing buyer will pay. When that buyer is Microsoft, the price is generous. When that buyer pauses, the price is uncertain. When that buyer disappears, the price is zero.

This is not an argument against DAC. We need direct air capture. The climate math demands it, as we’ve noted in our broader analysis of why carbon removal needs more than trees. But the business model for DAC cannot remain tethered to voluntary corporate purchases from a handful of tech companies. That was always a bridge. The bridge is now shaking.

CDR entrepreneurs and investors should be stress-testing every business model against a simple question: if no tech giant writes the next big check, does this company survive? If the answer depends on a single buyer category, the model needs to change. Co-benefits are not a nice-to-have. They are a survival strategy.

One important caveat. None of this changes the fundamental principle: CDR exists to address hard-to-abate residual emissions after aggressive decarbonization. It is not a permission slip to delay fossil fuel phase-out. Microsoft pausing purchases does not change the atmospheric need. It changes the commercial pathway. Those are different problems.

What to watch

Keep your eyes on whether the Ctrl-S knowledge-preservation effort becomes a broader pattern. If established energy or materials companies start acquiring distressed DAC startups for their engineering talent and intellectual property, rather than their credit pipelines, it will confirm something important. DAC’s future may lie not in standalone credit generation but in integration with industrial processes. Cement plants. Steel mills. Chemical facilities. Places where captured CO2 or the heat from capture has a use. That would be a very different industry than the one Microsoft’s purchases built. It might also be a more durable one.