In mid-November, Microsoft signed a first-of-a-kind, tech agnostic deal to buy removal credits from direct air capture project developer Deep Sky: a deal designed to overcome many of the infrastructure hurdles that DAC often faces, and which will start delivering tons in 2025.
The deal reflects the tech sector’s growing urgency to sign deals that can deliver on shorter timelines. 2030, which is when many of the most ambitious corporate decarbonization targets start to kick in, is just around the corner — and the emissions resulting from the artificial intelligence boom complicate what was already an ambitious undertaking.
As Microsoft’s senior director of energy and carbon removal Brian Marrs put it, “my driver’s license renews in 2030!”
As part of the deal, Deep Sky will deliver an initial 10,000 tons of removal from eight DAC projects that it’s developing in Canada, and have access to up to a million additional tons.
Deep Sky does not itself create direct air capture technologies; rather it’s acting as a project developer, bringing in a range of tech solutions from which it will aggregate and sell removal credits. Among the DAC solutions from which Microsoft will be receiving credits are Irish startup NEG8 and British “plug and play” DAC company Mission Zero. Deep Sky, which vetted and selected the various technologies, will handle infrastructure elements including real estate, sourcing clean energy, and managing carbon transport and storage.
The purchase marks the start of a new commercial approach for Microsoft, one Marrs described to Latitude Media as “a portfolio in a portfolio.” Deep Sky is both sourcing demand for its vetted DAC partners, and solving for a host of logistical problems in an aggregated way, allowing the startups to “focus on what they do best,” Marrs added.
Deep Sky’s model is a good fit for buyers that need tonnage by 2030 deadlines, explained the company’s VP of carbon markets Charlie Renzoni. The idea is that by aggregating several technologies in one place, Deep Sky will both deliver more tons more quickly, and bring down costs.
“That model…allows Microsoft to have a high probability channel for access to the supply that they need in 2030,” Renzoni explained. Microsoft will be receiving carbon removal credits from Deep Sky as an aggregate credit, he added, rather than as individual credits from the eight DAC companies.
By using multiple DAC technologies, Deep Sky and Microsoft are avoiding some of the risks that a bilateral deal — like between a tech company and a single DAC company — could face.
“If one component in a direct air capture unit experiences a delay in the supply chain, that company or project might be delayed up to multiple years,” Renzoni said. “We’re de-risking deliveries for our customers by having optionality between technologies.”
The primary difference between Microsoft’s deal with Deep Sky and its purchases directly from customers, Marrs said, comes down to a clause in the contract that gives Microsoft access to Deep Sky’s lab data.
“Microsoft will receive the data from each carbon removal credit delivered to them on which capture unit delivered it, the time of day the carbon was removed, when it was stored, and the performance characteristics of that technology,” Renzoni explained.
Deep Sky’s approach is so far unique in carbon removal, but both Marrs and Renzoni describe the deal as similar to a power purchase agreement.
“If you look at the market today, the companies that deploy projects are dedicated renewable energy project developers, and they do so in partnership with OEMs,” Renzoni said. “We are scaling that exact same model in the carbon removal market and for direct air capture specifically, we have absolute confidence that the market will track in that direction.”
Microsoft, Marrs said, is hopeful that the model will scale and be replicated moving forward. “From my seat, CDR is the new PPA.”
Marrs declined to comment specifically on whether Microsoft would be able to meet its 2030 goals in light of the AI boom, or whether they’d be able to purchase enough removal credits to make up for those increased emissions. That said, he did say that there’s “no linear path” and that getting to the 2030 target will require both technical and commercial innovation.
“Business growth and sustainability progress are related, and they change year to year and they’re non-linear,” he said. “So you’re going to have lumpy progress on both sides.”
