Responsible Carbon Removal Means Putting Science First

A glimpse into the diligence behind Microsoft’s first, and the world’s largest, carbon removal procurement 

When Microsoft launched its ambitious plan to decarbonize all of its historic and future emissions by 2050 the world took notice.

Microsoft was among the first large-scale buyers willing to look critically at all the available options for carbon management. Moving from vision to action was a complex process requiring deep expertise across disparate scientific and business disciplines. We created Carbon Direct to help both companies like Microsoft and governments navigate this challenging landscape, and together create an economy that removes more carbon than it emits. 

Carbon Direct faced critical questions: How can we help clients like Microsoft navigate a complicated ecosystem of both existing carbon markets and new technologies to remove more than a million tons of CO2 from the atmosphere? How can we achieve this objective in a scientifically robust manner that does not fall victim to historical challenges of the carbon offset market? 

When we began our partnership with Microsoft the available carbon removal options were limited. Most of these issues persist:

  1. Most offsets are avoided emissions, not removals. Partners like Microsoft who want to exclusively source carbon removal face a supply crunch. 
  2. Where carbon removal options are available, they are predominantly nature-based solutions like forests or soils. Very high-quality nature-based removals do exist, but distinguishing quality and durability requires expert diligence. 
  3. New engineered carbon removal pathways that offer clearer additionality and higher durability are emerging, but these too require expert review to verify their claims and ensure the integrity of carbon storage. Where engineered solutions exist, supply is limited and prices are higher than for more traditional offsets.  

Microsoft was our firm’s first client but far from our first experience in carbon management. Before partnering with Microsoft our science team, composed of over thirty leaders in academia, research and policy, had already mapped the landscape. We had reviewed every company, project, and organization we could identify that was capable of performing carbon removal as a service, a set of roughly 350 potential project developers at varying levels of commercial readiness. Partnering with Microsoft was a unique opportunity to put our knowledge into practice. 

Our interdisciplinary team brought expertise from across the many disciplines relevant to carbon removal: soils, forestry, direct air capture, biomass carbon capture and storage, blue carbon, and mineralization, as well as a number of promising nascent techniques. A science-first approach to analyzing carbon removal projects is critical. There were 79 organizations that proposed carbon removal services to Microsoft and we spent hundreds of hours analyzing a subset of those against a comprehensive framework for high-quality removal. 

Microsoft took a five-step approach that Carbon Direct was instrumental in supporting: 

  1. Create clear criteria for high-quality removals, covering additionality and baselines, carbon accounting, durability, avoidance of leakage, and a “do no harm” principle. 
  2. Develop a clear, user-friendly RFP with the flexibility to accept applications from across the breadth of natural and engineered carbon removal solutions. 
  3. Develop a scientifically rigorous process with our team of scientists to review documentation, and conduct full life-cycle emissions analysis and additional due diligence to identify promising suppliers and projects. 
  4. Work with Carbon Direct to identify the key risks to the integrity of the carbon removal purchases and a cost analysis that accounts for quality. 
  5. Make suggestions for ongoing monitoring and verification for the delivery of purchased carbon. 

Microsoft selected 1.3 million tons of CO2 from 26 carbon removal projects that span a range of techniques, including forestry, soils, bioenergy, and direct air capture. It’s been encouraging for our team to see this blended approach to carbon removal adopted more widely throughout the market. 

Interest in carbon removal and carbon management has grown rapidly since Microsoft’s initial announcement. Today, Carbon Direct’s client base is looking to avoid or remove over 700 million tons per annum of CO2, a carbon output the size of Germany’s. 

Through this close partnership with the Microsoft team, we uncovered several lessons for the industry that continue to inform our work across carbon abatement and removal. 

  • Carbon removal is difficult but attainable. Yes, finding high-quality removal – actual elimination of CO2 from the atmosphere coupled with highly durable storage – is challenging. We’ve joined in critiquing the dismal state of most offsets on the voluntary carbon market today. High quality supply does exist, however, if not yet at the scale that is needed or promised. We are committed as a firm to working with both existing carbon removal providers as well as emerging firms in the ecosystem. 
  • Project evaluation is more science than art, but it requires a flexible mindset. Our scientists have spent their careers exploring the ins-and-outs of carbon removal and storage within their scientific disciplines. However, value judgments are still required when analyzing carbon removal options. Longer duration carbon storage is more valuable than short-term duration of carbon storage, for example, but to what degree? Baselines and counterfactuals can also be incredibly complex to determine with confidence. We observed some project submissions with clearly overstated carbon baselines, and others where the evaluation was qualitatively challenging and actively debated within the team. 
  • While atmospheric carbon is fungible, the options for removing it are diverse from kelp farmers to air miners to turning CO2 into mineralized forms. Given the high degree of rigor that both Microsoft and Carbon Direct demand, and the fact that many removal techniques are not yet incorporated into mainstream certification schemes, we needed to conduct in-depth, project-by-project evaluations. 
  • There are passionate stakeholders in all verticals of carbon removal. Passionate supporters would often argue for the superiority of one form of carbon removal versus another. But the Intergovernmental Panel on Climate Change (IPCC) is clear in calling for large scale carbon removal by mid-century, and that effective carbon removal requires a portfolio of techniques. The enormity of the challenge is too large for a single vertical. The diversity of approaches on carbon removal techniques is present within our own team as well. Though we divided our advisory team into forest, soil, marine and engineered carbon removal verticals, the need to track all of the projects against a common set of criteria forced the team to leverage each other’s unique skills and disciplinary expertise.
  • These types of project and protocol-level reviews are still necessary, but the more we learn, the closer we’ll be to standardization and scaling the market for high-quality removal. Significant barriers to standardization remain. Our work with Microsoft underscored two major hurdles: (1) many carbon removal techniques are too new and insufficiently scaled to have protocols, and (2) where protocols exist, such as for mature pathways like reforestation, they are often too flawed to be trusted at face value without a second layer of review.

The next ten million tons.

Carbon removal needs to scale. Our expanded client base now represents an emissions footprint of more than 700 MtCO2/year. The vast majority of these emissions can and will be addressed via emissions avoidance. The best strategy for removing carbon is to not emit in the first place. But as Microsoft has described, not all emissions are easily avoided and even as the world moves towards a lower rate of emissions, we will still need to address over a trillion tons of historic emissions. While the challenge is enormous, the Microsoft RFP represents clear progress. We hope this work will inspire many other companies, institutions, and governments.