The distortionary effects of unconstrained for-profit carbon dioxide removal and the need for early governance intervention

Abstract

Carbon Dioxide Removal (CDR) is an emerging activity with extremely limited deployment to date, but which is mathematically required to achieve net (rather than true) zero or negative anthropogenic contribution to climate change. The required scale of CDR, however, depends on decisions about what activities will be allowed to emit greenhouse gases – the “residual emissions” that must be compensated via CDR. Simultaneously, CDR’s availability is limited by resource needs and feasibility, much like conventional depletable resources. Governance and institutions, especially related to how CDR is allocated and paid for, will fundamentally shape CDR efforts, including by structurally incentivizing particular approaches and monitoring, reporting, and verification (MRV) objectives. We argue that the emerging tendency toward market-based, unconstrained, and for-profit CDR presents fundamental and predictable risks for climate and justice goals. Such a model incentivizes growth in profitable compensatory removal applications, effectively allocating limited resources based on ability to pay rather than public good, while also increasing the amount of CDR required to meet global climate targets. “Luxury” removals that could otherwise be mitigated not only displace, but actively disincentivize deployment for compensatory removals in high priority but low wealth applications, and for drawdown. Meeting these needs would likely become a socialized cost. Markets also establish unit-level property rights that require specific kinds of MRV that are misaligned with climate outcomes and face incentives for poor quality verification. We describe the need, development context, function, and resource limitations of CDR, then characterize the major challenges with the emerging unconstrained, for-profit governance model. We argue that instead implementing CDR as a centrally planned sector, with publicly deliberated and adaptable volumetric targets integrated with other climate action, could enable more just and effective outcomes.

Keywords:

• Governance
• markets
• public control
• risk
• uncertainty

Introduction

Despite well over a century of understanding that greenhouse gas (GHG) emissions would lead to climate change [1–3] and three decades of global policy talks [4], global GHG emissions continue to rise [5]. The recent global response to climate change increasingly includes carbon dioxide removal (CDR) as a cornerstone of planned action, a consequence of limited action to date [6]. In particular, pledges, targets, or binding commitments to reaching “net zero” emissions that balance positive and negative anthropogenic contributions to climate change [7–9] implicitly or explicitly presume the use of CDR for compliance. Such emissions targets without CDR would require a target of zero: as long as anthropogenic GHG emissions to the atmosphere continue, net zero is not mathematically achievable without compensatory negative emissions through CDR and potentially related greenhouse gas removal (GGR) approaches that offset ongoing emissions to the atmosphere. Such approaches are also required for reaching net negative radiative forcing through drawdown removals that address legacy emissions in the atmosphere, which might be but is not definitively necessary for reaching global temperature targets [10, 11]. As we will argue in this paper, resource limitations (e.g. land; energy; etc.) constrain the total amount of CDR that can be deployed collectively for compensatory and drawdown purposes. Because drawdown can only begin once net zero is reached, higher demand for compensatory CDR (e.g. to offset emissions that could otherwise be prevented) restricts global access to drawdown functionality. As such, governance and institutional structures for CDR should be able to both enable and encourage the sector to develop in a way that enables reaching net zero, then net negative, anthropogenic radiative forcing. Early development to date, which relies on unconstrained market-based deployment, might not meet these needs.

Here, we define CDR to mean intentional, additional actions taken to capture CO₂ from the atmosphere (either directly or via intermediaries like biomass or the ocean) and permanently store it such that the CO₂ will not return to the atmosphere on time scales that at least match the lifespan of its impacts on the atmosphere and ocean [5]. Commonly proposed approaches that are potentially capable of delivering CDR include (but are not limited to) direct air carbon capture and storage (DACCS); biomass carbon removal and storage (BiCRS);¹ direct ocean carbon capture and storage (DOCCS); enhanced rock weathering (ERW); forestry; and soil carbon management. Some storage mechanisms, particularly those that rely on biological sinks like forests and soils, are not permanent in the sense of matching the lifespan of CO₂’s impacts. As such, we distinguish between CDR-capable interventions (e.g. an afforestation project) and actual CDR, which might entail consistent rehabilitation or replacement for projects where CO₂ is stored for less than geologic time (and which necessarily imposes greater administrative burden for strategies requiring relatively short replacement intervals).

The amount of CDR necessary to reach international temperature stabilization targets is implicitly dependent on societal choices that dictate how robust and successful emissions reduction measures² will be. Thus far, estimates of how much CDR might be needed range from almost none up to more than 300 gigatonnes (Gt) cumulatively by 2050, or over 1,200 Gt cumulatively by 2100 [10] – that is, up to about 10–15 GtCO₂/year starting immediately, contingent on simultaneous rapid emissions mitigation, to meet 1.5° or 2 °C targets. Such estimates are purely mathematical, balancing positive with negative emissions: in theory, CDR could be used to counteract any emission (currently about 60 GtCO₂e/year [5]). As such, CDR requirements will be higher for less rapid and/or lower levels of emissions mitigation. To date, binding requirements for decarbonization that clearly articulate which emissions should be mitigated and which remain residual emissions to be addressed via CDR [12] are rare, and CDR remains voluntary, contributing to a lack of clarity on necessary scope, scale, pace, and degree of resource competition.

In practice, CDR is likely to be resource limited (e.g. due to energy, land, and water requirements) even if successful approaches are identified and matured [13]. One review-based estimate for CDR potential suggests maximum sustainable global deployment of 4 to 25 GtCO₂/year, of which 0.5 to 9 GtCO₂/year are from forest and soil approaches with high reversibility and highly constrained storage volumes [14]. The need for permanent storage means that like fossil fuels, CDR resources are depletable (e.g. through nondisplaceable land or pore space occupation).³ For example, the estimated 0.5 to 9 GtCO₂/year available for soil carbon storage, biochar, and forestry approaches is expected to saturate within 20 to 100 years of full scale deployment, and requires effectively permanent maintenance thereafter [14].

Potential demand will likely exceed supply, especially for compensatory CDR in marketized contexts where CDR might be chosen over mitigation based on cost rather than necessity [15], which we characterize as “luxury” CDR. Long-term expectations for CDR costs remain uncertain [16], but targets frequently reference expected mitigation costs in sectors where financial incentives alone might indicate CDR over mitigation, thus increasing total CDR demand versus higher mitigation scenarios. As such, how to allocate valuable and limited CDR resources presents a major governance question. The allocation question is particularly salient given that some of the emissions that are most difficult to mitigate and least desirable to eliminate (e.g. certain agricultural emissions or emissions supporting critical services for the poorest and most vulnerable people) are also those associated with the least surplus wealth that could be applied to CDR. The emerging model of CDR as a growth-oriented, for-profit industry, rather than as a centralized, coordinated waste management sector conscious of resource limitations, presents grave risks for the ability of CDR to enable net zero and net negative targets in general, and to facilitate global justice goals in particular.

Carbon dioxide removal: Goals and forces shaping an emerging sector

CDR does not yet exist at a meaningful scale, and its ability to fulfill ambitious targets across approaches – particularly for approaches that early research suggests people prefer [17, 18] – remains a question. Yet, this nascency also presents a crucial and urgent opportunity. The lack of existing structures, incumbents, and property rights (i.e. the authority through ownership to determine how a CDR resource is used) means that the governance and institutional structures for CDR are not yet fully defined [19]. Simultaneously, emerging marketization and structural incentives to align with fossil activities and the needs of CDR purchasers pose serious risks to the long-term success of CDR for meeting climate needs. It is increasingly probable that CDR will be necessary for climate stabilization, and it could serve a crucial enabling role for climate reparations [20]. As such, ensuring that governance and institutional mechanisms are fit for purpose in establishing global CDR as a sector free from the structural injustices of the private sector and as close to a just ideal as possible will be critical. We argue that implementing CDR as a sector that is accountable to societal priorities for resource allocation has the potential to more closely approximate justice ideals than an unconstrained, market based industry model, though such a structure is not sufficient to ensure desirable outcomes. Successful implementation will likely require extensive public deliberation emphasizing that both technical information and values are relevant for decisionmaking [21], and recognizing that people are capable of meaningful engagement on even very complex issues when given sufficient support and opportunity [22].

The need for CDR is a direct result of historical and ongoing failure to manage GHG pollution, a failure that has both caused and resulted from structural injustices of resource allocation. As such, justice-oriented governance and institutional structures for CDR should be predicated on a fundamental desire to avoid replicating these harms, which likely requires substantial engagement with and deliberation by civil society. Unconstrained for-profit CDR biases resource allocation from both the supply and demand sides by simultaneously encouraging investment in the most profitable applications and conceding decisions about allocation to the wealthiest. Implementing CDR as a for-profit industry that allocates this scarce and precious resource according to ability to pay, likely with wealth accumulated via the same structurally unjust logics of fossil capital [23, 24], mimics the structures that created the need for CDR in the first place. Further, a growth incentive ignores the point that CDR is a corrective waste management industry: higher levels of CDR required to meet climate goals implies higher levels of harm that need to be corrected, which in turn requires still faster growth.

Advocates frequently appeal to the promise of atmospheric stabilization, correcting the harms of the past (e.g. “restoring the climate”), making reparations to most affected people and areas (MAPAs), and other claims related to CDR’s potential drawdown function [25, 26]. As such, establishing structures that leverage public goodwill toward drawdown CDR and “legacy” efforts [27] to facilitate unbounded compensatory CDR is both dishonest and likely to eventually lead to backlash that might compromise societal ability to deploy CDR at all (see, e.g. [28] on framing CDR as climate mitigation versus remediation, and [29] on the influence of context on public attitudes). The influence of application on public attitudes is a potentially important area for future research, particularly as the limited literature on CDR and public attitudes has to date focused on differentiated responses across technologies rather than purpose.

Fundamentally, governance structures should encourage minimum rather than maximum levels of CDR needed to reach multicriteria goals. Those goals might include distributional, reparative, and other priorities that raise the level of desirable CDR above the biophysical minimum required to meet net zero anthropogenic GHG contributions: for example, prioritizing energy access for those burdened by poverty, climate impacts, and other injustices [30]; ensuring safe and reliable energy and industrial service provision during transition [31, 32]; capital infrastructure buildout for climate adaptation [33]; or reducing the total cost of transition [34]. Overall, though, structures that allocate limited CDR resources to the highest bidder rather than to needs that support public welfare threaten CDR’s capacity to support these goals. Put more simply, CDR cannot succeed at restorative and reparative goals if it is controlled by the same forces that created the problems it is trying to solve.

A major framing dynamic for the development of CDR governance is that an unconstrained, profit-driven CDR industry implicitly and explicitly benefits from the ongoing use of fossil fuels and fossil fuel infrastructure, both directly (for fuel and storage resources) and indirectly (to promote high-cost growth by compensating for avoidable fossil fuel emissions or to cross subsidize CDR with high value fossil-aligned utilization). CDR already shows signs of interdependency with the oil and natural gas industries, particularly related to reliance on natural gas, electricity, and liquid fuels for energy (relevant for DACCS, DOCCS, ERW, and BiCRS), use of EOR-based CO₂ storage, and broader expectations of dependence on oil and natural gas industry assets (e.g. pore space) and expertise (e.g. pipeline operations) for storage (particularly relevant for CDR involving fluid CO₂). In turn, CDR provides the oil and natural gas industries with both an opportunity to deploy their assets (largely for energy intensive or fluid CO₂-generating approaches) and a route to ongoing existence as providers of carbon-based fuels (for all approaches, e.g. Southern Company’s reference to forestry CDR as a component of its net-zero plans [35]). Although CDR/fossil interdependencies exist across CDR approaches, the dynamic is most obvious for DACCS, both for technical reasons and because preferential treatment for DAC in both public and private contexts (e.g. due to DAC-specific US and purchaser preference for long-duration storage via fluid CO₂ sequestration) has contributed to earlier investment. Occidental Petroleum’s August 2023 purchase of Carbon Engineering, a DAC company with a natural gas-reliant process, provides evidence of this growing interdependency [36]. In industry contexts, CDR is frequently referenced as a way to prolong the use of fossil fuels [35, 37, 38], potentially indefinitely, despite a focus in academic and climate contexts on emissions mitigation first [8, 39].

A related dynamic of CDR and fossil industry interdependency is that technologies developed and deployed in service of CDR (which involves both capture and storage of atmospheric CO₂) could also facilitate ongoing reliance on bulk carbon-based fuels via utilization rather than storage of atmospheric CO₂. For approaches that generate fluid CO₂ streams, including DAC, biomass combustion and/or gasification, and others, CO₂ can be used for applications like synfuels with potential to be low or no GHG on a life cycle basis [40, 41]. Although such approaches might appear to make sense from a GHG balance perspective, synfuels are massively resource intensive, combining the resource intensity of the capture stage of CDR approaches with additional energy, water, and other demands [42, 43], including for hydrogen production [44, 45] required to create hydrocarbons from CO₂. Perhaps more importantly from a governance perspective, synfuels designed to mimic oil and natural gas products provide an easy route to returning to fossil carbon by fundamentally preserving the institutional structures and infrastructure used for fossil systems. As such, the fossil carbon will always have an advantage in a system that was designed for its needs, presenting significant risk that during challenging times under the energy transition, the world will simply revert to using fossil fuels [46].

The window of opportunity to design CDR as a sector is rapidly closing. Private sector activity is already embedding values and standards in the evolution of the CDR sector in the absence of governance institutions explicitly prioritizing the public good, particularly by creating market structures [47–49]. Recent policy decisions in the United States and elsewhere embed profitability and a growth incentive (e.g. through emissions and removal fungibility [19]) as core features of CDR within conventional neoliberal and capitalist structures. For example, the US’ Infrastructure Investment and Jobs Act (IIJA), also known as the Bipartisan Infrastructure Law, allocated $3.5 billion for Direct Air Capture hubs [50], contingent on 50% cost share – a challenging hurdle to meet outside of for-profit contexts even though other institutions are technically eligible [51], especially given the short time frame of the grant proposal process. The US’ Inflation Reduction Act includes tax credits that flow to proponents in a manner that incentivizes growth and maximization of atmospherically sourced fluid CO₂ storage – including for use in enhanced oil recovery applications – not targeted, approach-neutral application of CDR to areas where it is most needed. Notably, both IIJA and IRA explicitly privilege the most oil and natural gas-intertwined forms of CDR, with exclusive demonstration support for DAC, a requirement that at least two of four DAC hubs be located in communities with high levels of fossil resources, and subsidies essentially requiring isolation of fluid CO₂ for storage or utilization. The primary US federal research and development target for CDR emphasizes reduction of marginal cost [52] to enable rapid and widespread deployment, rather than a volumetric target based on climate needs. Although deployment to prompt scaling and sociotechnical learning is important, the direction of scale matters. Early governance and incentive structures are likely to have outsized impacts on how the sector develops because they will prompt path dependent growth that constrains future development to these selected paths [53]. That is, positive returns to scale along the selected pathways will create technology and governance lock-in that will be difficult to change.

Models frame CDR governance

Models often enjoy considerable authority in policy processes, and in particular can function as framing devices for policy discussions that implicitly or explicitly select specific issues as “scientific” and thus not subject to political contestation [54]. Integrated Assessment Models (IAMs) frame climate change as a market failure, thereby commoditizing the atmosphere and instrumentalizing widely divergent types of values as exchangeable through markets [54, 55]. Already, models with both intended and actual influence on policy have made structural and sometimes explicit determinations that the cost of CDR approaches like DACCS or BECCS imposes a ceiling on acceptable mitigation costs [15].⁴ These determinations presume that compensatory removals are preferred to mitigation based on financial cost alone, and without recognition that CDR will be finitely, and maybe not even widely, available. Particularly in settings incentivizing poor verification of the quality of compensatory CDR, this equivalency poses climate risk. As Figure 1 of Browning et al. shows, models of a US net zero scenario with very large CDR deployment are also those with high residual emissions (largely from energy applications that could be mitigated instead) [15], reflecting the modeling goal of reaching net zero, rather than net negative, emissions. From a resource constraint perspective, this approach is particularly alarming even if one presumes CDR is not practically constrained by demand for compensatory removals in that it does not recognize the opportunity cost associated with displacing drawdown CDR.

Figure 1. Atmospheric impact of compensatory CDR, drawdown CDR, and conventional Avoidance-based offsetting.

A CDR thesis

The authors write this piece from a place of embeddedness in the US CDR policy ecosystem, though with a critical perspective.⁵ Our core thesis is that the development of CDR as an unconstrained for-profit industry that allocates CDR based on emitters’ willingness and ability to pay poses major and intractable challenges for reaching climate targets and justice ideals, largely because of the fundamental motivation to maximize profitable and minimize unprofitable activity. In particular, we highlight the tension that CDR as an unconstrained for-profit industry would face in balancing fulfillment of its two waste management functions: compensatory and drawdown removals [10, 29]. We also acknowledge that avoiding the for-profit model is likely necessary but not sufficient for facilitating more just and more effective CDR outcomes, and that the details of alternative governance models are crucially important. Simultaneously, we argue that navigating the complex value decisions and establishing alternative governance systems that are more capable of delivering more just and more effective CDR requires public deliberation and co-creation by diverse representatives. A key tension, then, is that the precise nature of an alternative system – and attendant details about allocation, standards, and other implementation details – depends on results of processes that have not yet taken place. Nonetheless, we argue that the for-profit model poses such great challenges to climate and justice goals that an alternative is necessary.

CDR is a limited resource [14]. For-profit goals inherently prioritize the activities for which some entity will pay the most, which are likely disproportionately related to compensatory removals in high wealth contexts. Allocation of more CDR to compensatory functions constrains availability for drawdown while increasing overall demand for CDR and CDR scaling. These incentives create a structural bias toward providing offsets to high-wealth emitters who can provide ongoing revenue streams, and away from offsets for low-wealth emitters or remedial drawdown activities. In effect, unconstrained for-profit governance of CDR allows for luxury consumption to colonize an emergent global commons. We argue that in the context of CDR as a limited resource necessary for providing specific waste management functions demanded by global climate targets, for-profit approaches will misappropriate resources by encouraging higher volumes of waste to manage in areas where those wastes provide the most financial value, irrespective of societal priorities for existing and stranded waste management needs. Further, the profit model incentivizes lower quality verification, exploitation of loopholes to maximize payments (as already observed with related mitigative carbon storage under the US’ 45Q credit [56]), and orientation toward protecting a property right, rather than to maximize societal welfare via climatically effective removals. Such misappropriation threatens attainment of both net zero and drawdown goals, while still ultimately relying on public intervention to deliver necessary but unprofitable removals. Nonetheless, there remains a narrow opportunity to build a CDR sector that prioritizes societal welfare and global climate justice through centralized public governance that treats CDR as an allocable public good.

The function of CDR

CDR definitionally provides a very specific atmospheric function: namely, removing carbon dioxide from the atmosphere. This function is distinct from emissions mitigation functions that prevent carbon dioxide or other GHGs from reaching the atmosphere, for example by avoiding the production of those emissions in the first place (e.g. substituting coal with wind power) or by capturing and storing the emissions after production but before release into the well-mixed atmosphere (e.g. mitigative carbon capture and storage on a coal-fired power plant’s stack). In a sociotechnical sense, though, CDR has two very different functions: compensatory and drawdown removals (Figure 1).

A compensatory removal is one where CDR approaches are used to balance the atmospheric impact of an emission, resulting in “net zero” impact by adding a negative emission impact to a positive one (i.e. 1 - 1 = 0). Balancing atmospheric impact is not the same as balancing tonnage: because of dynamics like ocean CO₂ absorption/releases, more than 1 tonne of CO₂ removal will be necessary to counteract 1 tonne of CO₂ emission (medium confidence; [5]). Although this balance can be understood as an offset (i.e. CDR offsets emissions), this piece uses the language of compensatory removal to distinguish from the currently larger, more formalized, and challenged carbon offsets industry, which largely relies on avoidance rather than removal [57–60]. Avoidance offsets are fundamentally incapable of enabling net zero emissions, as they effectively change atmospheric impact math from (1 + 1 = 2) to (1 + 0 = 1), reducing but not eliminating the impact (Figure 1).

CDR’s other sociotechnical function is atmospheric drawdown, which removes CO₂ from the atmosphere without being directly balanced by impacts from a new emission (i.e. 0 - 1 = −1) (Figure 1). What we call “drawdown” CDR is sometimes referred to as “net negative” CDR (e.g. in IPCC’s SR1.5 [10]). We prefer the term “drawdown” because of the increasing colloquial conflation of CDR as a function with CDR as a class of technologies and processes potentially capable of removing atmospheric CO₂ (e.g. DACCS, BiCRS, ERW, forestry, soil carbon management, etc.). Actual atmospheric removal is not guaranteed to result from application of the technology or process. As a result, “net negative” is commonly used to indicate whether a given CDR-capable approach is actually resulting in CDR as a function based on an evaluation of life cycle GHG emissions (see, e.g. [61–64]).

In practice, a process resulting in functional CDR is compensatory or drawdown depending on atmospheric and allocation conditions, not technological ones. A DACCS facility could look identical whether it was providing compensatory or drawdown CDR. In a market-based context with unit-level property rights, the difference is essentially whether the CDR function is being claimed for a new or legacy unit of GHG impact in the atmosphere. From a global climate perspective, though, no CDR can be considered “drawdown” until anthropogenic climate impact is neutral [8], even if commodification in a market context creates a paper property right that the owner can allocate as desired. Ultimately, a property rights model of CDR encourages allocation and verification based on unit-level CO₂ storage, while a sector level model of CDR could encourage more atmospherically relevant measures of net GHG fluxes. Such a shift would fundamentally change the way that standards, verification, and other technical details of CDR would be implemented, at the extreme end potentially eliminating the need for unit-level measurement, reporting, and verification (MRV) and complex standards creation necessary for establishing ownership and private allocation.

The importance of MRV is one of the major sociotechnical dynamics related to both compensatory and drawdown CDR. MRV essentially seeks to confirm that a CDR approach actually results in removal at the stated level. To date, MRV has largely developed in the context of an emerging voluntary market, which encourages unit-level MRV consistent with concepts of private property rights. Within that context, two critical attributes that MRV seeks to validate are permanence and additionality, both of which are required for a specific flow of CO₂ to result in CDR.

As the name implies, permanence refers to whether CO₂ is permanently sequestered in a manner that eliminates atmospheric impact. In practice, permanence is not a binary attribute, but rather expressed as a length of time that the CO₂ can be guaranteed to be sequestered (and after which CO₂ would need to be removed again in order to retain the climate benefit) – for example, 100 [52] or 1000 years [47]. Some CDR approaches have much higher permanence than others. In particular, sequestration that involves mineralization of CO₂ (e.g. into carbonate rock) is expected to be permanent in geologic time, while sequestration in bio-based reservoirs (e.g. trees) is permanent only in biologic time. For example, CO₂ storage in trees is highly reversible via wildfires [57, 65].

Additionality refers to whether CO₂ would have continued to contribute to climate change absent the specific CDR intervention (that is, whether the removal is in addition to what would have happened in the counterfactual). Additionality is binary (i.e. yes or no), but can be ambiguous, especially in cases where CDR efforts engage processes with existing CO₂ flows. Demonstrating additionality requires establishing both a baseline (i.e. what CO₂ flows already exist?) and a rigorous counterfactual (i.e. what CO₂ flows would have occurred absent intervention?), both of which rely on both technical and political decisions [59]. Establishing a baseline requires establishing a time frame and system boundary. One well known example of how this process can be politicized is the UNFCCC’s Kyoto Protocol’s use of 1990 as the baseline for determining nations’ emissions reduction targets under international climate goals, except for countries undergoing transition to a market economy [66], which advantaged Russia and other former Soviet states from an emissions accounting standpoint (although not a human welfare standpoint) because of the collapse in emissions associated with the fall of the Soviet Union.

As with permanence, additionality is easier to demonstrate in some cases than others. In all cases, quantifying additionality of net removals requires accounting for additional GHG releases throughout the life cycle, for example associated with energy inputs and land disturbances [67]. Such life cycle accounting is particularly challenging for systems like DAC that use electricity before the grid is fully decarbonized (e.g. see [68] for discussion of marginal source evaluation in the context of atmospheric methane removal) and for BECCS/BiCRS mechanisms that use biological systems as intermediary repositories for atmospheric carbon, particularly due to complex temporal mismatches associated with CO₂ uptake [69]. For storage, baseline fluxes matter. There are effectively no natural fluxes of atmospheric CO₂ to deep saline aquifers, so all CO₂ sequestered in this way is fairly clearly additional CO₂ storage and more easily validated at a unit-level. By contrast, atmospherically significant quantities of CO₂ flow from the atmosphere into forests regardless of CDR interventions [70, 71], so demonstrating additionality of CO₂ storage for forestry projects is more challenging.⁶ Further, anthropogenic forest and soil degradation has contributed to climate change. As such, some restoration of these CO₂ sinks is required as emissions mitigation before storage can be truly additional, but there has been limited attention to establishing a base year or other metrics for differentiating land-based mitigation versus removal.

A third issue related to CDR MRV – sufficiency – is not widely acknowledged or addressed through existing voluntary markets. The climate impact of removals and emissions is asymmetrical on a mass basis, suggesting with medium confidence that more than 1 tonne of CO₂ will need to be removed per tonne emitted [5]. The magnitude of this asymmetry is both uncertain and dynamic, posing especially large MRV challenges for CDR in a property rights context where an emitter must purchase a removal credit. To date, voluntary markets effectively assume asymmetry is 0, resulting in a climate impact gap when CDR is used to compensate emissions.

Although the principles of permanence, additionality, and sufficiency are common between compensatory and drawdown CDR, the importance of certainty and accuracy differ. For compensatory CDR, overestimating CDR quality results in further GHG emissions to the atmosphere that contribute to climate change (i.e. because they are believed to have been canceled out, but are not). By contrast, for drawdown CDR, the impact of deploying CDR that is less permanent, additional, and/or sufficient than expected reduces the benefits of drawdown but does not actually make climate change worse by enabling more emissions. Emerging market-based governance tends to encourage development of MRV on a per-unit storage basis, which is only a proxy for climate impact; sector-level governance could conceivably conduct MRV on a net flux basis or other more direct measures of climate effectiveness that could reconfigure the metrics and relevance of current MRV. Although reaching determined climate goals (e.g. net zero radiative forcing; a specific atmospheric CO₂ level in the atmosphere) still requires accurate MRV, sector-level governance enables use of approaches that do not require high resolution (e.g. per-unit) evaluation. Scholarship on CDR recognizes the substantial burden and MRV risks associated with enforcing per-unit evaluation, with some writers highlighting the potential role of a “contribution claim” model of flexible CDR financing that supports CDR but is not explicitly tied to ownership of a GHG credit or specific GHG neutrality claims [73–75]. To date, this model is discussed largely in the context of markets and private CDR financing, but highlights recognition of the major challenges associated with managing CDR claims at the unit level.

CDR as a limited allocable resource

In this section, we expand the point that CDR is a limited resource [13, 14] that will be allocated between two fundamentally different sociotechnical functions of compensatory and drawdown removals (Figure 1), and therefore that applications must be prioritized in contexts where potential applications exceed supply.⁷ Such prioritization will present technical, political, and ethical challenges that could look very different for an unconstrained for-profit industry versus a sector focused on meeting societal needs based on public deliberation. In particular, we draw on concepts associated with depletable mineral resources (including fossil fuels) to describe CDR and relevant limits. In general, CDR discourse should be conscious of resource competition in two major categories. Within CDR, competition exists between compensatory and drawdown removals, and within compensatory CDR, between more or less “necessary” ongoing emissions, described by Lund and colleagues as contested “politics of residual emissions” [77]. Drawdown CDR could similarly become heavily contested in potential future contexts where countries or organizations are required or expected to compensate for their own legacy emissions – essentially, any situation where CDR is specifically allocated to a liability “owned” by some entity will necessarily result in claims on resources toward that entity’s benefit. Competition also exists between CDR and other resource users, including existing users, high-priority new users (e.g. for transportation or heating electrification), and other resource intensive activities emerging in response to climate and decarbonization pressures (e.g. like hydrogen and Fischer-Tropsch processes for synthetic fuels).

As with many mineral resources, although the physical supply of anthropogenic CO₂ molecules and places to put them is finite, it might not be the near-term limiting constraint on access to the CDR resource. (Major exceptions include the land-based forestry, biochar, and soil CDR approaches that Fuss et al. refer to as “twenty first century NETs” due to expected storage constraints within decades to a century [14].) Again as with many mineral resources, access to the end product (here, CO₂ permanently sequestered out of the atmosphere) is constrained by the intensity of inputs required to process the physical resource into the end product. In turn, this intensity is jointly determined by resource quality and technology.

In other extractive contexts, the concept of “ore grade” refers to the concentration of the resource targeted for extraction. This concept translates easily to CO₂ for CDR. As with other extractable resources, higher concentrations (or grades) are generally easier to extract. For example, approaches that can create concentrations of CO₂ that are much higher than ambient (e.g. by burning biomass) are attractive CDR targets because of the relative ease of extracting high purity CO₂ of recent atmospheric origin. Also as with other extractable resources, the resource intensity of extraction, whether financial or otherwise, is a determinant of viability. This is particularly true when inputs are in the same category as the product (e.g. financial or energy return on investment, which illustrates the need to spend money (or energy) to make money (or energy) [78, 79]). In the case of CDR efforts, inputs with GHG emissions are one major aspect of this resource intensity, as projects must remove CO₂-equivalent climate impact on a net basis [80]. Given widespread demand for zero carbon energy, arguably the highest marginal emissions energy resources are the ones that should be assigned to CDR efforts in evaluating overall climate impact [68].

In extractive contexts, these issues of resource intensity are often captured in concepts distinguishing “resources” from “reserves,” where a resource usually refers to the physical quantity of an extractive target, and a reserve refers to the actually extractable portion of that resource base [81]. Extractability is frequently characterized in terms of whether a given target can be extracted at all with existing technologies (e.g. whether net removals are possible in a BECCS context) and, if so, how expensive that extraction is. One commonly used representation is the McKelvey diagram (Figure 2), which categorizes the resource base according to geologic certainty (i.e. does the deposit exist) and economic feasibility (i.e. can it be extracted, and if so, at what costs).

Figure 2. Illustrative McKelvey diagram showing CDR availability based on permanence and economic feasibility (adapted from [82]).

In the case of CDR, identifying a resource is not difficult, given that the atmosphere is the main target reservoir and CO₂ is a well-mixed gas. Broadly, access to the CO₂ resource is essentially evenly and uniformly distributed – leading to the often-repeated point that CDR facilities capturing CO₂ from the atmosphere can hypothetically be located anywhere [83, 84]. Practically, though, what defines a CO₂ “reserve” for the purpose of CDR is heavily dependent on processing costs (to process the CO₂ “ore” into CDR via capture, purification, and storage). Given the nascency of CDR, processing approaches, costs, and resource inputs are not yet fully understood, and substantial innovation is expected to accompany scaling – both in the form of improvements to existing (nascent) technologies and in the form of entirely new approaches. Nonetheless, given the fundamental challenges associated with extracting material at low concentration – such as the ∼420 ppm of atmospheric CO₂ – energy intensity is expected to be, and remain, one of the major expected drivers of processing costs [16, 61, 85]. CDR approaches use energy in different forms (e.g. heat and electricity) and in different quantities, e.g. for processes like grinding in enhanced rock weathering applications; for fans and media stripping for direct air capture; and for compression of the CO₂ to pressures suitable for transport and storage in most applications involving fluid CO₂.

In the context of overall decarbonization, the energy demand of CDR, particularly in the form of zero carbon electricity, is likely to pose a practical constraint on how much CDR can be deployed (i.e. a depletable “proved reserve” base). Emissions mitigation is expected to be largely conducted via a combination of decarbonized electricity and end use electrification, in some cases indirectly through electrofuels, chemicals, and other feedstocks [86–88]. Based on current modeling, reaching net-zero emissions in the US suggests an electricity system with 5–8 times the capacity of the current system (∼1 terawatt (TW) of installed capacity) [15], to be built over the next 25 years if the US target of net-zero by 2050 is met [89]. At the 2023 pace of generation capacity buildout of about 50 GW/year [90] and recognizing the need to replace generators on a roughly 30–50 year interval [91], reaching these capacity levels would take well over 100 years – suggesting a need for rapid acceleration. The scale of electricity demand for decarbonization means that it is likely incorrect to assume that electricity will not be a limiting factor on CDR deployment at scale.

Arguments that highly capital intensive infrastructure like direct air capture units for CDR or electrolyzers for hydrogen production will be able to use “excess” electricity without dedicated buildout, competing with arguably higher priority demands for decarbonized electricity like buildings, transportation, and industry, are likely misguided [45]. For example, Figure 3 shows estimated 2050 US electricity demand for the Net Zero America Project E- Scenario (see Supplemental Information for scenario and calculation details), which includes about 0.7 Gt CO₂/year of DAC in 2050 [92]. In this scenario, electricity consumption for direct air capture exceeds consumption in any other sector that currently exists, besides the industrial sector overall. Electricity consumption for electrolysis (to make hydrogen), which is also effectively 0 today, becomes the highest consumer. A multimodel comparison of net zero scenarios for the United States conducted through EMF37 includes multiple models with much higher DACCS use in 2050, with ADAGE and FECM-NEMS both suggesting about 2 GtCO₂/year removals via DACCS [15]. Using the NZAP electricity intensity for DACCS, such levels would imply demand of roughly 4,000 TWh/year for DACCS alone – essentially the entire 2022 US electricity demand – within less than 30 years. This DACCS usage is in addition to other CDR deployments, e.g. LULUCF and BECCS.

Figure 3. 2050 Electricity consumption by sector, net zero America project scenario E- (TWh).

Global models suggest similarly extreme electricity demand for DACCS under high deployment scenarios. Figure 4 shows that the least constrained CEMICS 2.0 CDR scenario from the Intergovernmental Panel on Climate Change’s Special Report on Global Warming of 1.5 °C (SR1.5) [10] (modeled using REMIND 1.7, [93]) includes over 9 GtCO₂/year of DACCS in 2100. Using projected 2050 electricity consumption intensity estimates, such levels of DACCS alone would require enough electricity to fulfil about 80% of global 2020 demand (see Supplemental Information for scenario and calculation details). Again, this DACCS usage is in addition to other CDR deployments, e.g. LULUCF and BECCS, that have their own energy, water, land, and other resource limitations.

Figure 4. Modeled DAC deployment and electricity demand in 2100, REMIND 1.7. (a) Gigatonnes per year of DAC deployed in 2100. (b) DAC electricity demand in 2100 as percent of global 2020 electricity demand.

Of course, electricity can be generated from fossil fuels, and some CDR approaches (notably DACCS) also use large amounts of heat. Processes based on the use of fossil fuels, particularly natural gas, have poor carbon removal-return-on-carbon invested, especially given the atmospheric burden of upstream methane emissions [67]. Simultaneously, such approaches lock CDR infrastructure in to dependence on networked, high hazard industries that are expected to steeply decline under deep decarbonization and could encourage misallocation of embodied impact to infrastructure that will be stranded before the end of its useful life [33, 46]. That is, natural gas-based CDR not only forfeits a large portion of its removal potential, but also commits CDR to a vision of the future that is still producing, transporting, and otherwise managing a substantial amount of natural gas. For context, using DAC to remove 1 GtCO₂-equivalent/year using natural gas for electricity and heat would require about one third of 2019 US natural gas production [67], suggesting that natural gas-based DAC would create strong incentives for the CDR sector to promote the ongoing use of natural gas in other settings to ensure economies of scale and availability. This association also has potentially significant implications for public reaction and the social license for CDR (see, e.g. early indications of skepticism that CDR addresses the root causes of climate change whether fossil-affiliated or not [28]), in addition to pollution and environmental justice considerations [94]).

The Distortionary Effects of Unconstrained for-profit CDR

Developing CDR as an unconstrained profit-driven industry has major pitfalls that can and should be avoided. Specifically, the profit motive creates meaningful distortionary effects related to the nature, size, and allocation of resources for the CDR sector relative to what we might expect as societally and climatically optimal – for example, small volumes of compensatory CDR used to balance high priority ongoing emissions (e.g. from activities with significant societal benefits where emissions mitigation is extremely difficult for technical or societal reasons), then drawdown, with effectiveness measured directly based on GHG fluxes to the atmosphere. Past the point of net zero anthropogenic climate contribution, how drawdown CDR is allocated has limited effects on the atmospheric outcome and becomes largely a question of which entities might be able to satisfy obligations at what resource intensities.⁸ Prior to net zero, however, the implicit or explicit determination of which ongoing emissions are worth an allocation of CDR could have profound implications for justice (e.g. through appropriation of CDR, through copollutants from the ongoing emissions source), scale, and climate risk. As such, exploring governance models that allow for public deliberation on the scope of allowable residual emissions and then enforce such decisions is timely.

The profit motive incentivizes the CDR sector to maximize compensatory removals oriented toward ongoing emissions that generate the most private wealth (and thus ability to pay). Further, the profit motivation inherently incentivizes bad verification (up to the point where credits can withstand any relevant legal challenges, which might not align with climate performance), due to the financial rewards associated with selling credits at lower cost (incentivizing less spend on MRV) and higher volume (incentivizing overstatement of climate benefits, including by failing to account for emissions and removal asymmetry) [95]. This bad verification then itself exacerbates climate change if stated removals do not successfully balance ongoing emissions. Based on historical experience of regulatory capture by industries, establishing standards for sufficient verification would likely be subject to pressure from the CDR industry itself.⁹ Although CDR remains too nascent for clear empirical examples of the verification problem associated with commodity CDR, this issue has been significant in the carbon offsets space. Some of the clearest problems have emerged in the forestry offsets context [57], including a high-profile verification failure associated with the Kariba project in Zimbabwe through the offsetting company South Pole [96], which is troubling for CDR given investigation of forestation as a CDR-capable approach. Offsets over-crediting is common and driven by multiple mechanisms, such as non-additionality, favorable counterfactuals, and perverse incentives, that are not solved in the current CDR context [59].

Potentially more seriously, the creation of property rights in a market context determines what needs to be measured, reported, and verified – so far, confirmation that a specific mass of atmospheric CO₂ has been captured and stored, usually with precision to the single tonne level, so that the property right can be held against an emission measured in similar ways. From a climate perspective, the relevant measure of CDR success is atmospheric GHG flux and concentration at megatonne and parts per million levels, not confirmed storage of CO₂ by the tonne. As such, markets as they are currently developing actively incentivize different CDR approaches (e.g. DACCS, with easy mass measurement), different objectives (e.g. tonnes permanently stored rather than atmospheric concentration or the sign of atmospheric flux), and different precision (e.g. to the tonne level for storage with temporal matching, but without attention to atmospheric asymmetry) than a sector evaluating progress at a megatonne and multidecadal scale.

CDR has both capital and operational costs, often high, and thus requires indefinite funding. Contrast this, for example, with capital investments in wind or solar generation equipment with high capital but essentially zero operational costs. As such, the only profitable CDR is CDR that some entity is willing to pay for on an ongoing basis – likely compensatory CDR in high wealth contexts where emissions mitigation is difficult for some reason, which might be lack of control over Scope 2 and 3 emissions rather than technical challenges. An unconstrained for-profit CDR industry does not have a clear phase-out pathway for such compensatory CDR as mitigation strategies mature because the profit motive encourages growth, which is only available when a supplier is able to expand its buyer pool. The route to expansion is increased use of compensatory CDR, which incentivizes sellers to represent CDR purchases as effectively equivalent to mitigation interventions. This incentive is also distortionary in the sense that it could redirect efforts and pressure from high wealth and climate-motivated entities away from mitigation interventions. For example, a company that is willing and able to buy down technological learning curves might choose to invest in CDR rather than engaging the suppliers creating its Scope 2 and 3 emissions on mitigation innovations.

Arguments that unconstrained purchases from sectors able to pay for CDR are necessary for scale-up are flawed in that they overlook the point that the necessary scale of CDR depends on how the CDR is deployed. More demand for CDR (related to higher residual emissions) means the sector needs to be larger, which means it needs to grow faster: faster growth is not helpful when it is fully offset by higher total demand. Overconsuming compensatory CDR increases the necessary scale of CDR deployment and, in a limited resource context, means less and/or lower quality CDR (Figure 2) is available for drawdown, reparations, or compensation for emissions with high societal value and few, if any, mitigation options. Further, CDR is associated with probable significant externalities [13, 14, 29, 97].

Paradoxically, by encouraging growth in high wealth contexts, the profit motive both encourages an oversized compensatory CDR sector and fundamentally limits the amount of CDR applied to drawdown and high priority but low wealth compensatory needs. That is, CDR suppliers will not supply CDR unless they make a profit, but only compensatory CDR for emissions with high financial value has long-term profit potential outside a strict regulatory setting. Figure 5 illustrates this issue. In general, a for-profit CDR industry where buyers have discretion over how CDR is allocated will tend to encourage more compensatory CDR than is physically necessary to reach net zero, while simultaneously neglecting low wealth residual emissions and appropriating limited CDR resources that could otherwise be deployed for drawdown or compensatory applications with higher societal value.

Figure 5. Unconstrained profit-driven CDR will tend to invest in profitable compensatory removal but not unprofitable compensatory or drawdown removal (gray: Unprofitable; black: Profitable; Illustrative).

We acknowledge that a for-profit industry could conceivably deliver compensatory and drawdown CDR at levels reflecting societal prioritization based on considerations other than ability to pay for compensatory CDR if those levels were externally set – for example, if regulatory mechanisms effectively restricted compensatory CDR to societally determined high priority residual emissions and required specific purchasers to take responsibility for a centrally decided amount of drawdown. This approach essentially matches cap-and-trade rather than taxation logic, setting demand rather than price, but remains subject to the significant structural challenges of market-based rather than regulatory policies [98]. For example, any penalty that does not involve atmospheric correction de facto allows for noncompliance (e.g. because an emitter might choose to pay a fine and continue to emit) unless there is a “remover of last resort” that applies such penalty to CDR projects – in which case the system implicitly requires a centralized authority with a volumetric target regardless. Addressing these emissions then likely becomes a government responsibility, but potentially after the highest quality/lowest cost resources (Figure 2) have been absorbed by purchasers who do not need them, thereby both increasing public costs of reaching net zero and limiting societal capacity for drawdown. It is unclear how an unconstrained for-profit model would cover compensatory CDR for unprofitable residual emissions, or drawdown emissions in general, without relying on the common outcome of socializing those cleanup costs (e.g. [99]). In practice, this suggests that the core, unprofitable CDR necessary for reaching net zero or net negative anthropogenic climate impact will be publicly funded in any case.

A vision for an alternative

This piece aims to illustrate predictable ways that a for-profit CDR industry with no size or allocation constraints would be problematic in the long run, particularly related to achieving global climate and justice goals. That is, this for-profit approach is unlikely to deliver the outcomes that motivate CDR (climate stabilization and repair) and likely to create harm, even if it is currently politically viable. What, then, is the alternative? Mathematically, there is some value of both compensatory and drawdown CDR required to reach a given target for net emissions rates, atmospheric CO₂ concentrations, temperature rise, or other climate metrics of global policy interest. The size of that need depends on societal choices, explicit or implicit, about ongoing emissions.¹⁰ As such, public deliberation about these choices is essential. Making those choices explicit through industrial planning or other mechanisms is one way to envision a CDR sector that does not leave decisions about size and allocation of the limited CDR resource up to “the market.” Crucially, without having engaged in public deliberation, we do not know what the best version of a justice-oriented CDR sector might look like: such a determination depends on inclusive discussion on both priorities and preferences for how decisions are made. Political viability is dynamic, and building coalitions and support for visions of a successful future often underpins political will.

Core components of a justice-oriented CDR sector likely include a public governance system to determine how CDR resources are allocated, and a focus on enforceable regulatory constraints on climate impact. For example, a publicly funded and publicly accountable CDR sector might have centrally determined volumetric targets, based on public determination of which residual emissions must be covered by CDR and commitments to drawdown. Such a structure would avoid many of the issues associated with allocating CDR to compensatory removals that might have high financial value but low societal priority, displacing needs from sectors with high public value but limited private ability to pay (e.g. farming; remote hospitals; etc.). Public funding could still incentivize cost minimization within strict standards, but with the potential for less regulatory capture of those standards. Accompanied by better overall public planning for the climate transition [46, 100], such an approach could likely more readily avoid moral hazard concerns of replacing possible emissions mitigation pathways with removals by also including regulation of residual emissions, appropriate emissions phase-out timelines, etc.

From a climate perspective, CDR must be robust in the long-term and at the global scale [8], suggesting that sector-level strategies are preferable to the project-level management, verification, and other governance elements that tend to accompany the creation of a private property right. Envisioning CDR as a centralized sector with planned volumetric targets tied to other climate action could enable this robustness. In particular, a centrally planned sector could be designed with the capacity to adapt such targets based on approaches and MRV designed around atmospherically relevant objectives rather than those designed to facilitate verification of legally sound property rights at small volumes.

In practice, the unconstrained for-profit model already relies substantially on public funding in the United States, including a $180/tonne tax credit for DACCS. Ensuring public funding is accountable to public priorities is a logical step in the development of the CDR sector. Notably, significant public intervention is commonly observed in resource-constrained contexts, often as a corrective to growth incentives when markets have been introduced. For example, electricity, water, and waste utilities, even when implemented as for-profit enterprises, are frequently subject to significant regulatory infrastructure and mechanisms (e.g. volume charges) intended to reduce consumption and counter the tendency to preferentially allocate resources to the wealthy. Such interventions are incompletely successful, as evidenced by ongoing massive and unjust disparities in energy burden [101, 102], clean water access [103, 104], and other services – reinforcing further the point that CDR should be developed as a publicly accountable sector from the beginning, rather than committing to a path that will require imperfect corrections later.

Scaling CDR under public governance

CDR approaches are likely to change significantly as they mature, facilitated by increased experience with deployment. Simultaneously, the amount of CDR needed to achieve international climate goals will far exceed the available supply in the near term. If there is a need to rapidly scale CDR, and if the small scale of CDR at present means that resource limitations are unlikely to be meaningful in the near term, why does it matter how this early scaling is accomplished? Particularly if there is no physical difference between CDR used to compensate for ongoing emissions and CDR used for drawdown (so deployment of either one advances learning and scale up), is it not preferable for early institutional and governance structures to incentivize maximum deployment? And, even if emissions mitigation is preferred to emissions removal, if every ton of CO₂ counts, why discourage the use of removals to compensate for avoidable emissions, particularly early on? We argue that very early decisions about industry structure are less likely to change over time than technological conditions, and that very near-term institutional and governance decisions will shape both how CDR develops as a sector and what is actually possible to achieve. Notably, slower deployment is not necessarily harmful to climate goals, particularly if faster deployment targets avoidable emissions and reduces pressure and/or incentives to avoid them.¹¹

At a basic level, institutions and governance create path dependencies that, coupled with increasing return to scale, will tend to be deterministic with regard to where investments and breakthroughs are made. For example, voluntary carbon markets with unit level mass-based property rights already display preference for closed-system approaches like DACCS and verification based on storage rather than atmospheric flux or concentrations. That is, a system designed primarily for luxury compensatory CDR might indeed look different than one designed to meet minimal waste management needs. More broadly, the “rules of the road” for an emerging sector create expectations to which actors respond. Over time, this tends to entrench interests that are well served by those expectations while disfavoring entrants that are not. As such, the time to make institutional and governance decisions is at the very beginning of a sector’s existence, even if progress might look broadly similar in the near term.

Structural signals, including through policy, will affect relative incentivization of removals versus emissions mitigation [8], establish property rights that could be difficult or impossible to alter in a timely manner if governance structures change (e.g. see the history of natural resource nationalization and privatization efforts [105–107]), and generally create power structures that are aligned with the status quo. Already, we see that the emergence of a for-profit CDR industry where demand is unconstrained is leading to alignments with fossil fuel incumbents that matured under similar conditions. For example, claims of net-zero oil [108] and significant participation by oil companies in CDR efforts [109, 110], including by establishing fossil fuel dependencies in the form of inputs (e.g. natural gas [15, 111]) or outputs (e.g. oil production via CO₂ flood enhanced oil recovery [112–114]), reveals some of the risks of accepting an extractive capital logic for CDR. Especially given the high political power of the fossil fuel industries, particularly oil [115, 116], allowing CDR to develop using the same logic that produced the existing fossil fuel industries likely fundamentally constrains what could be possible in the sector.

A call to action

The structure of the CDR sector is not yet final, though current trends suggest a strong bias toward an unconstrained for-profit market model. The nascency of the sector, including the lack of entrenched interests, widespread property claims, or legal liability means that there is still an opportunity to thoughtfully design a CDR sector that both protects the climate and structurally incentivizes more just outcomes. Although the need for CDR exists because of longstanding and ongoing injustices, the sector can be designed in ways that do not perpetuate the patterns that created the conditions that necessitate it. Particularly given the clear risk for significant interdependencies to develop between CDR and the fossil fuel industries, especially oil and natural gas, identifying and avoiding such patterns early will be necessary for long-term sustainability of CDR as an atmospheric function with high potential to provide substantial societal benefits, including by stabilizing and perhaps even repairing the climate, and by providing a pathway for some form of reparations by the most responsible. For now, the nascent CDR sector is reliant on public infrastructure and public funding, much of which has not even been disbursed as of this writing. This reliance suggests a clear pathway to public ownership and public management of CDR in the long term – but one that will quickly disappear as the sector matures. CDR has the potential to be both more successful and more just if it is not developed under an unconstrained for-profit regime. The time to act is now.

Data availability statement

All data are included by reference in the manuscript.

Disclosure statement

The authors declare no competing financial interests. EG currently serves as a Board Member of the Carbon Removal Institute as an academic representative to this 501(c)3 organization, which is affiliated with the 501(c)4 Carbon Removal Alliance industry group. ST currently serves as a member of Frontier Climate Impact Fund Advisory Board.

Notes

1 We present the acronym as commonly used, but note that carbon dioxide capture without storage is not removal: as such, saying “removal and storage” is redundant.

2 We use “mitigation” to mean emissions mitigation (i.e. emissions reductions) rather than climate change mitigation, under which the IPCC and other bodies categorize both avoidance and removal approaches.

3 Also like fossil fuels, being depletable does not imply a specific time scale over which this consideration becomes limiting: the time frame is a consequence of rate of use.

4 As Beck and Krueger write, embedded “intrinsic ethics” that are explicitly or implicitly written into models can become “extrinsic ethics” with meaningful societal impact when models inform policy [49]. For CDR, a liability-based, privatized model of emissions responsibility is a key example.

5 We (EG, ST) are both former members of the US Department of Energy’s Office of Fossil Energy and Carbon Management (DOE FECM): EG on an Intergovernmental Personnel Act assignment from academic duties as the Deputy Assistant Secretary of Carbon Management (7/21–6/22), and later as Senior Advisor for Energy Asset Transformation (12/22–5/23), and ST as the politically appointed Chief of Staff for DOE FECM (1/21–3/22). During our overlapping work with FECM, we were heavily involved with the creation of the Carbon Negative Shot, with R&D prioritization for CDR in the Biden-Harris Administration, and with initial discussions related to Direct Air Capture (DAC) Hub implementation after the passage of the Infrastructure Investment and Jobs Act (IIJA, also known as the Bipartisan Infrastructure Law, or BIL). ST previously worked at Carbon180 as Deputy Director of Technology Policy and later as a Senior Visiting Scholar (8/20–1/21 and 4/22–5/23), with a focus on just and sustainable policy mechanisms to scale CDR. EG currently serves as a Board Member of the Carbon Removal Institute as an academic representative to this 501(c)3 organization, which is affiliated with the 501(c)4 Carbon Removal Alliance industry group. ST currently serves as a member of Frontier Climate Impact Fund Advisory Board.

6 As Peng and colleagues point out, conventions about the carbon additionality of forest growth relative to a recent baseline where longer-industrialized countries are judged against their deforested context and industrializing countries are judged against a preindustrial forest level tend to suggest that longer-industrialized (and usually wealthier) countries produce climate benefits by harvesting, whereas industrializing (and usually poorer, often previously colonized) countries create climate harms by harvesting [72] – further reinforcing a dynamic where wealthy countries are able to exploit their natural resources while preventing poorer countries from doing so.

7 We assume this is true in a practical sense for many decades, given that GHGs are being emitted at a rate of about 60 GtCO₂-e/year [5] and that legacy anthropogenic emissions are roughly 3 trillion tonnes of CO₂ forcing equivalents [76]. Assuming that “potential” applications include fully compensating ongoing emissions and restoring the atmosphere to preindustrial CO₂ concentrations, potential demand for CDR far exceeds supply at decadal scales.

8 As with avoidance-based offsets, this question about who can access the lowest-harm CDR (in the form of financial and other costs, for example) to satisfy legal and other obligations is extremely important but poorly explored in the literature. Returning to the mineral resource analogy explored in the McKelvey diagram above, consider that CDR resources have an “ore grade” – some are high quality and low cost; others might require substantially more cost to produce a usable resource (analogous to a vein versus porphyry copper deposit, for example). In a fully financialized, commoditized context, the highest resource entities arrive early and can “high grade” CDR resources, leaving more difficult and/or more expensive resources for lower resource entities that might not move to CDR as quickly. Although this dynamic seems distant in a moment where CDR is largely pre-commercial (so the earliest movers may pay a premium for technology development), once approaches are largely mature, dynamics like lower cost land, access to suppliers, better quality storage sites, etc. could become relevant. As an example in the avoidance offsets context, consider that a low-wealth country might sell the climate benefits of its new wind farm (a common project type in the Clean Development Mechanism context) to a high-wealth country that continues to operate its coal plants. When the low-wealth country later needs to show compliance with an emissions schema, the only offsets available for it to purchase are associated with high-cost mitigation activities that the wealthy country was unwilling to pursue in the first place – i.e. the low-wealth country sold its lowest cost opportunities for mitigation and now must find alternative ways to comply.

9 In early stages, we might expect to – and in fact do – observe pressure from some CDR companies to increase MRV stringency to preserve the integrity of markets and marketed claims, particularly because the widely variable certainty (and time scales) of claimed removals mean that some companies have an incentive to align policy with their own technologies’ capabilities. As the industry matures, we would likely expect lobbying efforts to focus on reduced stringency up to the point that the most powerful actors are protected from competition from those selling lower quality products. For example, DACCS companies that claim 1,000 year geologic CO₂ storage have an incentive to prevent forestry companies that can only claim 10-100 year biologic CO₂ storage from entering the market, but do not have an incentive to require a standard of 1,000,000 year storage, 0 geologic leakage, responsibility for compensating for any leaked CO₂, etc.

10 Notably, it is not guaranteed that CDR deployments are possible for any envisioned scale that might be mathematically necessary to hit a given target, which means that prioritization of how to allocate limited CDR resources is especially important at times of great uncertainty.

11 Of course, it is reasonable to ask whether applying compensatory CDR to avoidable emissions is preferable to continuing to emit (i.e. not avoid) avoidable emissions. The answer depends on the counterfactual. If CDR is applied to avoidable emissions for a couple of years while the mitigation mechanism is set up, and then transitioned to either high priority compensation or drawdown purposes, particularly if the avoidable emissions emitter pays much of the life cycle cost of the CDR, this approach could be beneficial. At this very early stage of negotiating what net zero and net negative emissions might be, though, another (and possibly more likely) scenario is that an emitter of avoidable emissions who is able to access CDR at an acceptable cost will view their task as complete and continue to use CDR for these avoidable emissions indefinitely if not forced or strongly incentivized to transition to mitigation approaches. Even if mitigation might ultimately be less costly, transaction costs associated with changing approaches could be very high, especially if CDR prices are considered acceptable and might even have the potential to fall. A strong regulatory framework could potentially encourage the first scenario.