Beyond Carbon Accounting
The relevant question for the climate is not whether a company can achieve net-zero on its own books, or how to tighten the standards such that the company’s accounting claims are credible; rather, it is whether the company causes more clean energy and climate-related infrastructure to get built than would otherwise exist, and what actions it is taking to make that happen. The first approach has consumed a decade of corporate sustainability efforts and accounting debates, while the second has too often been ignored and undervalued.
This paper argues that investment additionality and transparency about desired outcomes and instrumental actions – not company-level accounting precision – is the right way to evaluate corporate climate ambition. It explains why the current framework systematically fails to reward either approach, uses the data center buildout as a consequential illustration of what is at stake, and proposes a contribution-based alternative that asks companies to state clearly what outcomes they are trying to achieve, demonstrate that their actions cause clean energy and infrastructure to exist that would not otherwise, and report on what they are actually doing to get there: financing first-of-a-kind projects, funding transmission, and advocating for the policy changes that markets alone cannot deliver. These are claims that can be verified against observable actions and would connect corporate behavior to actual global decarbonization outcomes in ways the current framework cannot.
The Problem
- Corporate net-zero frameworks assign emissions liability to individual companies and require them to eliminate or offset it. This property rights logic works for simple, localized pollution problems, but will fail at addressing a global infrastructure challenge.
- Carbon offsets, which allow liability to be transferred on paper, mostly fail to solve the physical problem. Rigorous empirical research now finds that less than 16% of offset credits represent real emissions reductions.
- The data center buildout makes the current framework’s issues apparent. Hundreds of gigawatts of new demand for electricity is coming online, but the existing framework is guiding companies to buy renewable certificates from projects that likely would have been built anyway and offset the rest. This does nothing to bring clean firm power down the cost curve, build transmission capacity, address interconnection backlogs, or advance climate policy.
The Alternative Proposal
- Evaluate corporate climate ambition based on whether a company’s aggregate behavior is making decarbonization more or less likely, instead of asking whether its purchase caused a specific clean energy project to exist. The individual counterfactual approach assumes that corporate procurement decisions and decarbonization are connectable by a traceable causal chain, but they are not. Rather, decarbonization is driven by system-level variables that no bilateral purchases control or meaningfully influence.
- Require companies to state what they are actually trying to accomplish and report on the instrumental actions they are taking to achieve it. This is a more honest measure of climate ambition than ton-denominated accounting, although it does not solve the underlying problem with corporate climate commitments, which is that they are voluntary.
- Measure impact against observable action and sector-level outcomes, not against accounting positions that can be satisfied by doing the lowest-cost thing without meaningfully influencing the system. The goal should not be a new form of individual credit-claiming, but an honest account of whether a company’s capital and influence are directed at the issues that actually determine the pace and scale of decarbonization.
The Error at the Heart of Corporate Net-Zero
Unlike many conventional pollution problems, climate change is inherently global. The atmosphere does not differentiate between a ton of carbon emitted in Ohio and a ton emitted in Beijing. The atmosphere responds to the total flux of greenhouse gases across the entire global economy, and to whether the systems that produce those emissions are transforming – or not.
The corporate net-zero framework treats this complex, global, sociotechnical problem as if it were a matter of property rights. The logic is Coasian: Assign ownership of an externality to the party that causes it, require that party to eliminate or transfer the liability, and trust that the aggregate of individual compliance will produce system-level change. Applied to voluntary climate action, this means assigning emissions to individual companies, requiring them to reduce or offset those emissions, and evaluating ambition based on the quality of their ledgers.
This logic works for simple, localized externalities where the assignment of ownership is clear and compliance is enforceable, but it fails for voluntary corporate climate commitments for reasons that are structural, and cannot be fixed by tightening accounting standards. First, the assignment of ownership is ambiguous. Whether a ton of carbon dioxide belongs on the books of a company that burned the fuel, the company that extracted it, or the company that manufactured the product it powered is an accounting convention, not a physical or moral fact. Scope 1, 2, and 3 boundaries do not correspond to causal responsibility in any clean way. Second, the transfer mechanism that allows liability to be shifted (carbon offsets) allows the liability to be removed on paper without solving the actual physical problem. If ownership of emissions can be transferred to a forest offset that turns out to not represent real, permanent carbon removal, it may “net” on paper, but the atmosphere is worse off. Third, the property rights framework only works to solve climate change if the sum of individual compliance actions produces system-level transformation. But the truth is that the fundamental barriers to decarbonization (modernizing the grid, bringing clean firm power down the cost curve, changing the policy and regulatory environment) can’t be solved by companies optimizing for their own accounting books.
Importantly, the property rights approach might be defensible in mandatory compliance contexts. If every significant emitter were required to participate and if the commitments were legally enforceable as property rights typically are, then one could assign liability, require remediation, and let the aggregate of individual compliance produce system change. But in a voluntary system, the property rights approach produces an outcome in which companies optimize their own accounting positions while the rest of the economy is unaffected. In a voluntary system, the property rights approach produces an outcome in which companies optimize their own accounting positions, while companies that do not participate have no incentive to change their behavior. The result is not system-level transformation, even if it can be marginally impactful.
Moreover, even if participation were universal, this approach assumes that the sum of individual emissions reductions actions would result in the change the world needs to limit the worst impacts of climate change. The energy transition is not just, or even primarily, an accounting problem – it is a complex, sociopolitical infrastructure challenge. First, making these systems work in the real world has proven enormously challenging. But even if firms did reduce emissions, this approach would not have nearly as much impact on reducing carbon as would solving coordination failures in infrastructure investment. Corporations reducing their emissions footprint are not incentivized to do coordinated, long-term investment across multiple jurisdictions and regulatory regimes, nor are they incentivized to invest in new climate technologies the world desperately needs to manage climate risk. They are incentivized to buy the cheapest clean energy certificate and offset what they cannot match. Addressing infrastructure barriers requires a different kind of corporate action that is not just about responding to accounting incentives, but about actively financing, contracting, and advocating for the infrastructure that markets alone cannot yet deliver.
Another reason for the persistence of the company-specific, ton-denominated framework is that companies are motivated by proprietary benefits, but many of the interventions that are most useful for global decarbonization are public goods (e.g., transmission investment, clean firm power, policy change). A company that advocates for permitting reform or invests in advanced geothermal often creates benefits that accrue to every user of the grid, including its competitors. There are very few incentives for a company to finance shared infrastructure or advocate for policy changes that benefit the whole system. Any new framework for voluntary action has to grapple with this directly.
This is why the question of corporate climate ambition cannot be reduced to accounting precision. The most valuable thing a large energy buyer can do right now is use its scale, balance sheet, contracting power, and advocacy capacity to help solve the coordination problems that markets and price signals have not. That is a different theory of corporate climate contribution, and it requires a different framework.
The Voluntary Framework: What It Got Right and Where it Went Wrong
Large companies with net-zero commitments now cover more than $36 trillion in annual revenue. It’s important to note that early corporate clean energy procurement was catalytic, because the primary barrier to renewable energy deployment was commercial, not infrastructural. Large buyers willing to sign long-term power purchase agreements (PPAs) at relatively high prices helped drive solar and wind down the cost curve at a time when cost was the binding constraint on deployment. Between 2014 and 2024, corporate buyers signed over 100 gigawatts’ worth of clean energy deals, representing 41% of all clean energy capacity added to the U.S. grid over that period. And there may still be cases where commercial barriers are significant. For instance, long-term offtake agreements can provide the revenue certainty to enable project financing: According to the Corporate Energy Buyers Association (CEBA), voluntary PPAs reduced rates of financial distress among renewable projects by more than 75% compared with purely merchant scenarios. These were real contributions to deployment.
However, the major barriers facing climate action have changed dramatically in the last few years and the framework has not evolved. Wind and solar are now the cheapest sources of new electricity generation in most of the country, and the constraint on their deployment is no longer commercial viability, but infrastructure. More than 2,290 gigawatts of proposed clean energy projects are waiting in interconnection queues, most of them wind, solar, and storage, not because they lack buyers or financing, but because the transmission system cannot accommodate them. In this environment, particularly in the U.S., purchasing unbundled renewable energy certificates (RECs) does not cause more clean energy to get built. Long-term PPAs can stabilize revenues, which does provide some real value, particularly as federal tax credits expire, but even well-structured procurement does nothing to address the interconnection backlogs, transmission gaps, and grid infrastructure deficits that now determine whether clean energy can scale. A company can purchase enough RECs to claim 100% clean electricity, hit every sustainability target it has made for itself, and contribute very little to solving the problems that are actually limiting clean energy deployment.
The existing framework systematically misprices the most valuable corporate contributions to decarbonization (the ones that shift system trajectories rather than improve individual accounting positions) while rewarding contributions that are legible to auditors, but may be marginal in climate terms. Sophisticated corporate procurement today has already moved beyond the procurement-only model, an implicit acknowledgment of the inadequacy of the status quo. CEBA now highlights tariff innovation, grid infrastructure investment, nuclear restarts, and demand flexibility as core activities alongside traditional PPAs, with members like Google, Microsoft, and Oracle financing transmission upgrades, battery storage projects, and utility planning processes directly. These contributions are real and consequential, and they are also largely invisible to the frameworks that evaluate corporate climate ambition. A company can fund a $50 million battery storage program, negotiate a novel tariff structure that brings two gigawatts of new clean generation onto the system, and advocate for grid-enhancing technologies, but none of these contributions have a natural home in standard corporate greenhouse gas reporting. Scope 2, which governs matching clean electrons to corporate demand, is not measuring grid investment or infrastructure outcomes, and Scope 3, while broader, has no methodology for capturing the system-level benefits of tariff innovation, transmission financing, or grid advocacy, and very few companies attempt to do so.
An emerging alternative to standard GHG reporting, the Task Force for Corporate Action Transparency’s Mitigation Action Accounting and Reporting Guidance, distinguishes between activities that change a company’s accounting position and activities that change actual atmospheric outcomes. But even this framework explicitly excludes actions whose climate impact cannot be expressed in metric tons of CO2e, leaving transmission financing, interconnection cost sharing, and grid advocacy – the contributions that matter most for large electricity buyers – largely illegible. Consequently, some of the most important actions a large energy buyer can take to accelerate global decarbonization are not legible to the frameworks that are meant to evaluate climate ambition.
This is not an argument against ton-denominated accounting, as such. For product-level emissions disclosure, carbon border adjustments, and emissions trading regimes, the ability to assign comparable, auditable quantities of emissions to specific products and transactions is genuinely valuable, and debates about how to do that well are worth having. The argument here is narrower: that ton-denominated accounting is the wrong tool for evaluating corporate climate ambition, because the question of whether a company is accelerating decarbonization is not reducible to the question of internal emissions accounting. A ton-denominated framework that is reasonably well-suited to its origins in regulatory and trade contexts can also be poorly suited to the task of directing voluntary corporate action toward the highest-impact investments.
Importantly, the underperformance of corporate climate commitments is not mostly due to bad actors or “greenwashing” (although this is certainly real). The issue is that the framework guides even rational, well-intended companies to the wrong activities: buying RECs that aren’t driving additional clean energy onto the grid and offsetting residual emissions with the cheapest tons. Corporate net-zero commitments emerged and proliferated during a period of relative political consensus on climate ambition (2015-2022), when near-zero interest rates made clean energy investments easier to justify financially, ESG was ascendant, institutional investors were pushing companies toward stronger commitments, and reputational pressure from employees, customers, and civil society created genuine incentives to participate in these efforts. Eventually, the voluntary regime began to morph into mandatory disclosure frameworks. The U.S. SEC proposed this, California passed a law requiring large companies to disclose their emissions footprints, and the EU’s Corporate Sustainability Reporting Directive (CSRD) came into force in 2023.
The context has changed dramatically: The SEC climate disclosure rule has been withdrawn, California’s implementation has been delayed, and the EU’s CSRD has been dramatically weakened. As for voluntary action, ESG backlash accelerated sharply, with major asset managers (e.g., BlackRock, Vanguard, State Street) pulling back from or softening their positions, and several large banks and insurers leaving the Net-Zero Banking Alliance and Net-Zero Insurance Alliance. Republican state AGs are pursuing aggressive legal strategies to undermine ESG actions, and several major companies have been quietly walking back or revising their commitments. The retreat from voluntary climate commitments also reflects a problem of the framework’s own design: The reputational logic that drove corporate climate ambition created powerful incentives to announce aggressive targets (e.g., net-zero by 2030, carbon negative by 2040, 100% clean electricity by yesterday) with little scrutiny of whether those targets were achievable or what achieving them would actually require. (There are parallels with public policy targets, as well.) The result was a decade of commitments that outpaced operational reality by a wide margin. A 2025 analysis found that none of the 55 major global companies reviewed demonstrated high- or reasonable-integrity climate strategies, not because all of these companies were acting in bad faith, but because the framework rewarded ambition signaling over verified delivery.
Two Illustrations: Carbon Offsetting and the Data Center Buildout
The problems with carbon offsetting – the mechanism the current framework relies on to address residual emissions – are not new. The core claim in every offset transaction is fossil-fuel equivalence: that a ton of carbon avoided in a forest or a cookstove project is climatically identical to a ton not emitted from a coal plant. If that equivalence holds, the company’s net position is zero; if it doesn’t – if the offset credit does not represent a real, additional, permanent reduction in carbon dioxide – then the company has emitted a pulse of carbon dioxide into the atmosphere in one place and purchased an accounting object with no physical correlate. The evidence is now clear that this equivalence fails to hold in the vast majority of cases. Researchers have documented that overcrediting is not an exception, but a structural feature of how offset programs are designed. A 2024 meta-study in Nature Communications synthesized more than 60 rigorous empirical studies covering 2,346 carbon mitigation and nearly 1 billion tons of carbon dioxide in credits, finding that less than 16% of those credits represented real emission reductions. A separate 2025 analysis of the 47 largest offset projects found that 80% of retired credits from those projects were unlikely to deliver promised reductions, and that 93% of those problematic projects were in the Global South.
Some companies have already implicitly acknowledged the equivalence problem by investing in permanent carbon dioxide removal techniques. Frontier, an advanced market commitment for permanent carbon removal launched in 2022 by Stripe, Alphabet, Shopify, Meta, and McKinsey, is premised on the recognition that durable, geologically permanent carbon removal – and not avoided emissions or temporary sequestration – is the only equivalence to a pulse of fossil fuel emissions. Notably, some of these companies were using contribution-based logic, focusing on what the world needs to manage climate risk, regardless of what it meant for their reported emissions positions.
The debate about accounting for the emissions impacts of the data center buildout is another example of the weaknesses of our current voluntary corporate climate regime. Hundreds of gigawatts of new load are coming online over the next decade, and yet the corporate climate frameworks guiding those decisions are almost entirely focused on the wrong question: Which procurement strategy best supports a company’s net-zero claim? Jesse Jenkins and colleagues at Princeton have argued that only hourly temporal matching — procuring clean energy at the same time as consumption, around the clock — consistently produces measurable, systemwide emissions reductions. Annual volumetric matching, the dominant corporate strategy, produces emissions reductions that are effectively zero, because the solar and wind projects it supports would largely have been built anyway given their cost competitiveness.
Today, the dominant corporate clean energy strategy scores relatively low on additionality. For a standalone renewable energy certificate purchased from an existing installation in a mature market, where the project likely would have been built even without a company purchasing a certificate, there is a very weak physical relationship to additional clean energy deployment. Long-term PPAs fall somewhere in the middle, because they can stabilize project revenues and reduce financing risk, though their additionality has declined as wind and solar have become cost competitive without corporate offtakers. Again, a company can make a net-zero claim and have done very little to address the major barriers to decarbonization.
The problem is that even perfect accounting, achieved by individual actors optimizing for their own footprint, is not the same as actually reducing emissions on the grid or accelerating global decarbonization. The world needs cheap, clean firm power that can outcompete fossil energy, not just in wealthy grids with deep capital markets, but everywhere. Solar and batteries have already come down the cost curve globally, and are being deployed at massive scale without corporate PPAs from U.S. hyperscalers. The technologies that have not yet come down the cost curve, including advanced geothermal and nuclear, are the ones that need deployment at scale, which drives down costs through learning rates and supply chain development. A company that achieves net-zero in the U.S. through demand flexibility or solar and batteries may be making a claim that is technically accurate, but it does not move the cost curves that determine whether firm clean power becomes a viable option, particularly in non-OECD economies where emissions are projected to grow.
With respect to data center electricity demand, the important fact is that the grid is a shared system and electrons are fungible. The right response is not to tighten accounting standards so that companies must work harder to make the same kind of claim. This approach cannot solve the underlying counterfactual problem, and the political conditions for enforcing tighter voluntary standards do not currently exist. Asking companies to meet a tighter standard voluntarily, at a moment when the incentives to meet even the current standard are eroding, will not work. It is time to move to different claims altogether.
This is not an argument for wholesale abandonment of existing commitments. Companies with targets or net-zero pledges should maintain them, not least because ton-denominated, accounting-based targets serve as a backstop against the cases where contribution-based actions somehow fail to translate into emissions reductions. But ton-denominated accounting should not be the primary basis on which corporate climate ambition is evaluated. Instead, contribution-based approaches focused on investment additionality, transparent theory of change, and advocacy should be the primary basis on which that ambition is judged.
The Right Question: Additionality and Outcomes
Corporate climate commitments should be evaluated against two questions: 1) Does this action cause more clean energy and climate infrastructure to exist than would otherwise (investment additionality)?; and 2) What climate outcomes is the company actually trying to achieve, and what instrumental actions is it taking to get there?
The real shift that companies should make is from “accounting for our footprint accurately” to “taking actions that change the trajectory of the global energy system.” A company serious about global decarbonization and climate should say clearly: We want firm clean power to get cheap enough to outcompete fossil fuels globally, and here is what we are doing about it – financing first-of-a-kind projects, signing long-term offtake agreements, funding transmission infrastructure, and advocating for regulatory and policy change. This is verifiable against observable actions rather than contested accounting conventions, and it would connect a company’s behavior to actual outcomes.
Under a contribution-based framework, impact can and should be measured across both dimensions: 1) the amount of dollars deployed into high-leverage investments, megawatts of new clean generation financed, miles of transmission enabled, and policy proceedings influenced; and 2) evaluation of what the company’s explicit theory of how global energy systems decarbonize is, and whether the company is taking actions consistent with that theory – not just optimizing for its own footprint, but working to shift the conditions that determine whether clean energy wins globally. These are imperfect metrics, but so, too, are the metrics we are using today, which convey a false precision that may be worse than grappling with honest and irreducible uncertainty.
A contribution-based framework would have to contend with real questions about attribution (if multiple companies co-finance a project or achieve a policy win, who gets to claim credit?) and operationalization (without a quantified target, how does a company budget for and justify action?). On the former, one feature of a contribution-based framework is that it moves away from clean individual accounting positions that are often falsely precise, and toward something more like the logic of corporate philanthropy or shared research and development. On the latter, the current framework (a need to think in terms of auditable individual targets) has driven this anxiety. In reality, companies budget for research without being able to attribute specific revenue outcomes to specific investment very frequently, even without precise attribution. Moreover, binding financial commitments, like support for a fund or advance market commitment, give companies something concrete to budget against more than the current framework does.
A contribution-based framework would reorient these incentives entirely. Consider the American Grid Infrastructure Fund (AGIF) model: a voluntary consortium mechanism through which large loads — hyperscalers, primarily — make binding financial commitments to grid modernization as a condition of favorable interconnection terms and policy treatment. Participants finance transmission infrastructure, bear a share of interconnection costs on a cost-causal basis, and commit to verified demand flexibility that improves grid operations. Under a ton-denominated accounting framework, none of this has a natural home, because it doesn’t reduce a company’s Scope 2 emissions directly, it doesn’t generate verifiable offsets, and it isn’t legible to standard auditing frameworks. Under a contribution-based framework, it is precisely the kind of action that could be recognized and rewarded: dollars deployed into high-leverage infrastructure, megawatts of new clean generation made possible by transmission capacity that would not otherwise have been built, and grid flexibility that reduces the need for fossil backup. A company participating in AGIF should be asked to state plainly what it is trying to accomplish: cheaper clean firm power, transmission capacity, and a grid that can power electrification. It should also report on the advocacy it is doing to achieve those outcomes in regulatory and legislative proceedings.
The truth is that a contribution-based framework still has to grapple with the fact that the most important benefits associated with the desired actions are public, not private. There are several ways to address this concern. The most honest answer is that most important public goods required for decarbonization require binding policy, and no voluntary framework can substitute for it. Still, a contribution-based framework could reward advocacy that produces binding policy, making lobbying for permitting reform or interconnection standards a visible and credible contribution to climate action. In addition, there are some areas where public and private interests really do align (e.g., transmission investment), and companies can capture enough upside to make the investment worthwhile. Lastly, infrastructure investments and policy advocacy can be justified on non-climate grounds, making them potentially more politically sustainable than a framework focused exclusively on emissions accounting.
The point is not that AGIF is the only model or that transmission finance is the only contribution worth making. It is that the current framework systematically excludes this entire category of action, while directing corporate dollars toward cheaper, more auditable, and less impactful alternatives. Fixing the framework means making room for the contributions that matter most, and the data center buildout is the most important near-term test of whether we can get this right.
Conclusion
The corporate net-zero framework emerged as a genuine attempt to apply rigorous accounting discipline to a real problem, and in its early years it produced real results, including corporate procurement that drove solar and wind down the cost curve at a moment when commercial barriers were the binding constraint on deployment in the U.S. and abroad. But the context has changed, and frameworks must evolve. The alternative proposed here is not a rejection of accountability; rather, it asserts that companies should be held accountable for whether their actions cause more clean infrastructure to exist than would otherwise, and whether they are transparent about what they are actually trying to accomplish and what they are doing to get there. This is in some ways a harder standard to meet than buying RECs or nature-based offsets, but it is also more honest. It requires companies to have a theory of how global energy systems decarbonize, not just to clean up their own books, and to take actions consistent with that theory: signing offtakes to derisk investments the market won’t make on its own, financing technologies that are not yet cost competitive, and advocating in policy proceedings for the changes that markets alone cannot solve.
Now is the moment to get this right. The data center buildout is the largest demand-side event in a generation, and the companies driving it are making decisions now that will shape the grid and energy systems for decades. These companies have the capital, contracting power, expertise, and regulatory and policy presence to move markets. They are also, at the moment, being guided by a framework that asks them to account for their emissions rather than using those assets for the good of the global climate system. The gap between what the current framework incentivizes and what the situation requires is large and consequential. The framework in this paper offers a way to realign the enormous financial and political resources of the largest electricity buyers in the world with the infrastructure investment, technology innovation and deployment, and policy change that actually determines whether the world decarbonizes. It’s time for the framework to evolve.