Catalytic carbon removal purchasing: Milkywire’s strategy

Introduction

Milkywire’s strategy for purchasing durable carbon removal enabled by our corporate partners like Klarna, Spotify and Salesforce is designed to act as a catalytic driver for the CDR ecosystem. By prioritising neglected opportunities and supporting innovation, we seek to maximize the long-term impact of CDR as a pillar in global climate solutions. 

That means looking at what has potential in the long run, and what the sector needs to develop as opposed to just trying to remove as much carbon as possible today. Although the long-term need for CDR is immense, and the ecosystem has made a promising start, the sector is likely to face challenging years ahead. Despite rapid early growth and investment, voluntary demand is lagging, and many startups are likely not going to be able to raise more venture capital to make up for the lack of sales. 

As an ecosystem, we must avoid: 1) Leaving promising CDR ideas untested, 2) Losing talented individuals who have gained valuable experience in carbon removal, and 3) Burning investors. We don’t want a repeat of Cleantech 1.0, where widespread losses led to a "dead decade" for investment. Catalytic buyers can play a crucial role in mitigating these risks, particularly the first.

The CDR ecosystem still needs altruistic buyers focused on innovation to ensure promising ideas are not overlooked. Such buyers can complement mainstream purchasers, and be unconstrained by metrics such as total tonnes purchased, price per tonne, or delivery timelines. For example, in Enhanced Rock Weathering, buyers need to prioritize supporting companies that advance understanding of ERW and gather data from diverse geographies and crops, rather than fixating solely on the lowest cost per tonne. Similarly, with technical methods like DAC, cost-insensitive buyers can help ensure extraordinarily promising new ideas are explored. Additionally, such buyers can help elevate approaches that provide the highest co-benefits for people and nature in the long run and ensure geographical diversity. While some of these bets may fail, for catalytically focused buyers, the knowledge gained and long-term potential can justify the risk.

We believe the 2020s is the "CDR Startup Decade," a time to identify what works best and provide the conditions for growth. This is when altruistic, catalytic buyers are most critical. Commercial scale-up of the most promising approaches will largely take place in the 2030s, with the 2040s focused on achieving massive "soak-up" at the gigatonne scale.

We see our role in strengthening three layers of the CDR ecosystem: the entire sector, startups, and specific projects. At the sector level, we aim to accelerate knowledge-building and overall growth. At the startup level, we help promising ventures reach critical milestones for further funding or scaling. At the project level, we catalyze real-world demonstrations that can influence the entire ecosystem.

We want our catalytic CTF purchases to directly help the CDR supplier to:

• For new CDR companies that have not sold significant amounts of tonnes yet: 

• Either to develop and prove a particularly promising new, previously untested CDR method in the real world (including promising variations of existing methods), or

• Scale up an existing method or practice in a more sustainable or cheaper way or with higher co-benefits than what already exists, and/or 

• Contribute to ecosystem-wide positive effects. Such as enhancing the scientific understanding of a method, creating new business models or financing pathways of CDR, or unlocking new geographies and groups of buyers. 

• For established CDR companies (for some sales or deliveries, even if our main focus is on new approaches):

  • Help them reach a significant milestone of learning, cost decreases or co-benefits

  • Require a track record of at least a pilot-scale removal and credible initial results.

  • Prioritize solutions that show high future potential and clear evidence of additional benefits or breakthroughs.

We do not want to fund 

• Projects where our purchase doesn’t make a direct difference in outcomes. If our funding won’t affect whether a pilot is built (for example, if the company already has sufficient financing, or if the project is so large that our share won’t move the needle), then the purchase is unlikely to be catalytic.

• Straight-forward scale-up of well-established CDR methods if it does not yield significant new benefits or learnings. For example, running an existing DAC plant longer, or spreading more rock in ERW beyond what’s required for meaningful learning.

These purchases are not meant as mere awards or recognition for being “best-in-class.” We look for additional impacts beyond simply acknowledging excellence, such as accelerating a project’s timeline, helping secure follow-on funding, or enabling meaningful R&D breakthroughs. 

However, we remain open to allocating a limited portion of the budget to “confidence-driven purchases,” i.e., supporting very promising CDR startups at risk of shutting down, even if doesn’t directly contribute to new learnings, provided we have reasonable confidence in their ability to deliver. This would be limited targeted support to prevent promising innovation from stalling and losing talented staff and signal enough market confidence to attract further investment or partnerships.

This strategy refers to our catalytic new CDR purchases. Milkywire has already supported 27 different CDR startups and can also help companies who are primarily interested in near-term deliveries from established methods to purchase tonnes from our mature CDR portfolio companies. 

Here is a summary of how we are thinking about different CDR methods for our catalytic purchases

• Direct air capture (DAC) 

  • It is a crowded space with over 140 different DAC companies and counting. We don’t expect more than a handful of actors to reach the “soak-up” phase, but it is still too early to pick a dominant design to start scaling, we need to figure out what has the highest potential first.

  • Currently, 85% of all DAC volume has been sold by only three suppliers. Frontier, Milkywire, and Terraset are almost the only buyers of long-tail, early-stage DAC. (Data from CDR.fyi)

  • Projected near-term demand for DAC credits are unlikely to enable multiple megatonne deployments. That means few startups can rely on economies of scale to drive cost decreases. Those who can drop costs in the lab and have iterative, modular designs have a better shot than those who require capex-heavy large-scale deployments. 

  • There is a bit less need to broaden the DAC ecosystem compared to 2-3 years ago though. For any new purchases, we would need to see either completely new promising methods, or strong innovation on resource use, integration, co-benefits or in other ways have very innovative ideas.. 

  • We may repurchase from our existing suppliers if a) first machine is deployed removing carbon, b) the potential still looks very high compared to the competition, and c) our purchase would make a big difference. 

  • These points goes to some extent for Direct ocean removal DOR, but it is not as crowded and a bit more open to innovation still. DOR companies have very high electricity use in general, so we would look for innovation here.

  • So far we purchased from the following DAC companies: Heirloom, Climeworks, Terrafixing, Parallel Carbon, Mission Zero, Octavia, Yama, Holocene, Phlair, Ucaneo and Gaia refinery.

• Biochar

  • A relatively mature technology and the most popular choice among new carbon removal buyers. Many biochar companies are ready to produce, whether starting from zero, or expanding current operations, but hold off until getting credit purchases. The field ultimately scales with purchases in tandem with a market for the physical biochar. We expect there to exist many biochar providers in the “soak-up” phase due to the distributed nature of the method. The logic is not to trial a lot of different ideas to see which works best, but rather to seed uniquely strong and sustainable biochar providers in different markets. 

  • Our biochar purchasing approach focuses on identifying companies that are uniquely well–suited to scale effectively based on sustainability, co-benefits, viable business models (valorizing the physical biochar and other co-products is key), and cost-efficiency. 

  • The sector as a whole does not need altruistic purchases as much as other methods. Almost all buyers of durable CDR is purchasing biochar. However, altruistic buyers can give an extra push behind suppliers with the highest co-benefits and or best outlooks, to help them get started.

  • So far we have purchased from Husk, Mash Makes, Takachar, Biosorra, PyroCCS and Planboo/Solidaridad.

• Enhanced Rock Weathering 

  • Shows promise for large-scale carbon removal and soil benefits, but reaction rates, measurement challenges, and soil variability remain unresolved. 

  • Pioneering companies deploying basalt are generating crucial data but currently struggle to reliably and consistently detect bicarbonate formation in soil water, proving the removal. 

  • Carbonate rock weathering (e.g. lime) shows potential for faster and more measurable removal, but may be constrained by suitable soils.

  • More research, improved MRV, and data sharing are needed. 

  • We will only support projects that enhance the scientific understanding of ERW and require data sharing with researchers in the Cascade Data Quarry. 

  • We will prioritize companies that use new methods, new types of rocks, new soils or similar innovation improvements, as well as those that have the largest co-benefits. 

  • We would not buy larger volumes than what is necessary for learning, and we are exploring research-focused pay-for-practice agreements rather than traditional credit pre-purchases.

  • So far, we purchased from Silicate, Mati Carbon, Inplanet and Flux. 

• Direct biomass storage 

  • Direct biomass storage is a low-tech CDR method with the potential to store carbon very cheaply for centuries or millennia if managed properly. Approaches range from above-ground dry storage, to buried biomass, deep injection, and ocean sinking.

  • There are remaining uncertainties around permanence, environmental impact, and best use of biomass. More research, standardized MRV, and clear regulations are needed. 

  • The method is cheap, so should do well on its own without catalytic buyers, but there is uncertainty among the wider buyers pool about the method. 

  • Milkywire can support innovative projects that help reduce uncertainty and demonstrate long-term sequestration and safety. We will require data sharing with researchers. 

  • We will look for serious operations by strong teams who take permanence, long-term monitoring, co-benefits and community buy-in into account. 

  • We have so far purchased from InterEarth.

• Bioenergy with Carbon Capture and Storage (BECCS) 

  • Combines biomass energy production with CO₂ capture and storage, enabling net-negative emissions. It spans power plants, biogas, ethanol, and pulp industries. Challenges include high capital costs and large size of projects that are difficult to pull off, storage infrastructure, policy support, and sustainable biomass sourcing.

  • We are unlikely to purchase from large-scale projects as our purchases would not be likely to make much difference. Smaller-scale projects are a possibility, but the opportunities we evaluated in previous years have not been catalytic enough (aside from the purchase we made from Carbon Removers, prev Carbon Capture Scotland).

• Ocean-based approaches

  • Includes alkalinity enhancement, Direct Ocean Removal, macroalgae cultivation, and biomass sinking. Offer scalable potential but face challenges in MRV, environmental impact, and regulatory clarity.

  • The field needs research on ecological effects, MRV improvements, and stakeholder collaboration. 

  • We will look for opportunities to support promising innovation. 

  • We have purchased from Vycarb, Aquarry and SeaO2.

• Mineralization and Industrial Waste Approaches

  • Includes methods, such as ex-situ and microbial mineralization, waste mineral utilization, and other novel technologies. The field is in its early stages but has large potential. 

  • Key needs include R&D and identifying sources of waste materials suitable for CDR, demonstration projects.

  • We will look for opportunities to support promising innovation.

  • We have purchased from Alkali Earth.

• We are also always open for other ideas for durable carbon removal that may not fit in with any of the above methods. 

Data Sharing and Transparency Requirements We will also start asking companies to share data and learnings with the wider community.  Milkywire will mandate data sharing from supported companies to advance the carbon removal field. Exactly how the data sharing will happen can differ between companies. We encourage collaboration with researchers and alignment with ecosystem initiatives to enhance MRV and best practices. We will not ask companies to share proprietary business secrets that could hurt their viability though.

Credit purchases as a support mechanism Prepurchasing carbon removal credits has been our modus operandi for supporting the CDR ecosystem, but it has some limitations. Credit sales are highly sought after by startups, and they provide critical funding that can be used directly for capital expenditures. For instance, $500,000 may fund the capex of a 100-500t/yr DAC machine. Sales also help startups raise venture capital, often this is the biggest effect of a purchase. However, taking on the responsibility to deliver credits carries some risk of diverting early-stage companies from their core focus. For example, requiring early-stage DAC companies to secure external storage for their CO₂ to deliver removal credits risks shifting their priorities away from R&D. Extending delivery timelines or offering flexibility in purchase agreements can mitigate some of this risk. (This is particularly a DAC/DOR-specific risk, as other methods like BECCS and Biochar are less about R&D and more about execution, and others such as ERW have storage as an integral part of the method.)

Separate selection criteria Note that this CDR purchasing strategy complements our Milkywire CTF selection criteria. The strategy sets the broader framework to help find and shortlist projects. The criteria are then used as a rigorous tool to assess already shortlisted projects' integrity, potential, and the added effect of our support.