Microsoft has committed to buying millions of tonnes of carbon removal credits across multiple record-setting deals, positioning itself as the largest corporate buyer in a market reshaped by the energy demands of artificial intelligence. The company’s purchases span permanent underground storage in Sweden and soil-based sequestration from U.S. farmland, together covering more than 6 million credits. These agreements signal that Big Tech’s AI ambitions are now directly financing the scale-up of carbon removal technologies, but whether those technologies can deliver verified, lasting results at the volumes promised is far from settled.
Stockholm Exergi and the Largest Permanent Removal Deal
The single biggest contract in Microsoft’s carbon credit portfolio is a deal with Stockholm Exergi, the Swedish utility that operates the district heating system for Sweden’s capital. The agreement covers 3.33 million tonnes of permanent carbon removals, making it the world’s largest such contract to date. The removals will be carried out through bioenergy with carbon capture and storage, or BECCS, at Stockholm Exergi’s facility in the Vartan district of Stockholm.
BECCS works by burning biomass for energy, then capturing the CO2 produced during combustion and injecting it into deep geological formations where it is intended to remain permanently. The technology has attracted attention because, unlike forest-based offsets that can be reversed by wildfire or logging, geological storage is designed to lock carbon away for thousands of years. Stockholm Exergi’s plant already generates heat and electricity for the city, so adding carbon capture to its existing operations could, in theory, turn a conventional power station into a net-negative emissions facility.
The scale of the contract matters beyond the headline number. By committing to 3.33 million tonnes, Microsoft is effectively prepurchasing output from a facility that has not yet reached full operational capacity for carbon capture. That kind of advance purchase agreement gives Stockholm Exergi the revenue certainty it needs to invest in completing the BECCS infrastructure, a dynamic that turns the tech company into a de facto financier of industrial decarbonization hardware. The risk, of course, is that the technology encounters delays, cost overruns, or lower-than-expected capture rates, leaving Microsoft holding credits that may not materialize on schedule.
For Stockholm Exergi and similar developers, the deal also acts as a signal to investors and policymakers. Having a marquee buyer like Microsoft on the books can help unlock additional capital and regulatory support, particularly in Europe, where policymakers are looking for scalable ways to meet climate targets. Press-release distributors such as PR Newswire have amplified the announcement, underscoring how corporate demand for removals is becoming a mainstream part of climate communications rather than a niche technical topic.
Soil Carbon Credits and the U.S. Farm Connection
Microsoft’s appetite for carbon credits extends well beyond Scandinavian engineering projects. The company struck a separate record agreement for soil carbon credits, agreeing to buy 2.85 million credits over 12 years. These credits are tied to agricultural practices on U.S. farms that sequester carbon in soil, such as cover cropping and reduced tillage.
Soil-based sequestration operates on fundamentally different principles than BECCS. Rather than capturing emissions from a smokestack, it relies on biological processes in farmland to pull CO2 from the atmosphere and store it in organic matter below ground. The appeal for buyers is that it channels money directly to farmers, creating a financial incentive for land management practices that also reduce erosion, improve water retention, and build long-term soil health. According to reporting on the program, such practices can both cut carbon emissions and help soils retain more water, a co-benefit in drought-prone regions.
But soil carbon is notoriously difficult to measure and verify. Storage permanence depends on farmers maintaining specific practices year after year; if a participating farm switches back to conventional tillage, the stored carbon can be released. That makes monitoring and enforcement complex and potentially costly. It also raises questions about whether soil credits should be treated as equivalent to permanent removals from geological storage, given the very different risk profiles.
The 12-year timeframe of the deal is notable. It locks Microsoft into a long purchasing commitment, which gives agricultural credit programs a stable demand signal that can justify investments in data collection, modeling, and farmer outreach. Yet it also means the company is betting that measurement, reporting, and verification standards for soil carbon will improve enough over the next decade to withstand growing scrutiny from regulators, investors, and environmental groups. If verification methodologies tighten faster than anticipated, some early credits could be deemed less robust than later vintages, complicating claims about climate impact.
Why AI Is Driving the Buying Spree
The timing of these deals is not coincidental. Microsoft and its peers have seen their operational emissions climb as they build out vast networks of data centers to power AI workloads. Training and running large language models requires enormous amounts of electricity, and even as companies invest in renewable energy procurement, the sheer pace of data center construction has outstripped the available supply of clean power in many regions. The gap between corporate net-zero pledges and actual emissions trajectories has widened, and carbon credits have become the primary tool for closing it on paper.
Microsoft ranks among the largest buyers of credits in the voluntary carbon market. That buying pressure, concentrated among a handful of deep-pocketed tech firms, has created a supply crunch. Carbon removal credits now command far higher prices than traditional offsets pegged to forest preservation projects, according to the same reporting. The price premium reflects both the perceived higher quality of removal credits and the simple economics of scarcity: there are not yet enough verified removal projects operating at scale to satisfy the demand that Big Tech is generating.
This dynamic creates an unusual feedback loop. AI development increases energy consumption, which increases emissions, which increases demand for carbon removal credits, which drives up prices, which channels more capital into building removal infrastructure. In the best case, that capital accelerates the deployment of technologies like BECCS and soil sequestration to the point where they become cost-effective tools for economy-wide decarbonization. In the worst case, the credits serve as accounting entries that allow tech companies to claim progress on climate goals while their actual emissions continue to rise.
The Gap Between Accounting and Atmosphere
The distinction between what happens in corporate ledgers and what happens in the atmosphere is at the heart of the debate over Microsoft’s strategy. On paper, a tonne of CO2 removed in Sweden or sequestered in American soil can balance out a tonne emitted by a data center almost anywhere in the world. This logic underpins the entire voluntary carbon market. But in practice, the timing, durability, and risk profile of those tonnes vary widely.
Permanent geological storage, if executed as designed, offers a clear climate benefit: carbon that would otherwise be in the atmosphere is locked away for centuries or longer. Yet even here, uncertainties remain. Long-term monitoring obligations, liability for potential leakage, and the social license to operate underground storage sites are all evolving issues. For Microsoft, relying heavily on such projects means accepting that some of the climate benefit will not be fully demonstrable for decades.
Soil carbon, by contrast, delivers benefits that are more immediate but potentially less durable. A farmer who adopts cover crops this year can improve soil structure and biodiversity quickly, but the associated carbon gains can be reversed if market pressures or policy changes push them back toward more intensive practices. That makes the climate value of soil credits contingent not just on current behavior but on a web of future economic and regulatory conditions.
Another layer of complexity is additionality, the idea that a credited activity should represent a change that would not have happened without carbon finance. For Stockholm Exergi, Microsoft’s advance purchase is clearly tied to building out BECCS capacity that does not yet exist, strengthening the case that the removals are additional. For U.S. farms, the picture is murkier: some practices may have been adopted anyway for agronomic reasons, while others truly depend on carbon revenue. Distinguishing between the two is difficult, yet central to the integrity of the credits Microsoft is buying.
Transparency will be critical to maintaining trust. Platforms that distribute corporate climate announcements, such as newswire services, can amplify bold claims about “world’s largest” deals and “record” purchases. But investors, watchdog groups, and the public are increasingly demanding granular data: project-level performance, verification reports, and clear accounting of how credits are used against specific emissions sources.
Can Big Tech’s Bets Reshape Carbon Removal?
Microsoft’s multi-million-tonne commitments demonstrate how quickly demand for carbon removal can scale when a small group of companies is motivated by both climate pledges and the reputational stakes of AI-driven energy use. The company is effectively stress-testing two very different approaches: industrial-scale BECCS that aims for permanence, and distributed soil projects that tie climate finance to everyday land management.
Whether these bets ultimately help stabilize the climate depends on factors that extend far beyond Microsoft’s balance sheet. Regulators will need to set clearer rules for what counts as a high-quality credit. Project developers must prove that they can deliver removals reliably and transparently. And tech firms themselves face a more fundamental question: how quickly they can curb the growth of their own emissions through efficiency gains, grid decarbonization, and demand management, rather than relying indefinitely on offsets.
For now, Microsoft’s record-setting purchases are accelerating the build-out of carbon removal infrastructure and drawing attention to both its promise and its pitfalls. As AI continues to drive energy demand, the company’s approach may become a template, or a cautionary tale, for how digital innovation and climate responsibility intersect in practice.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
