Issue 16 / Clouds

March 27, 2022
Old-school browser window with an image of an forest, partially obscured by misty clouds

Decarbonization as a Service

Holly Jean Buck

Net zero is becoming a lucrative problem for software to solve.

Net zero has gone viral. Everyone is announcing their net-zero greenhouse gas emissions targets, from Saudi Arabia to Australia. By the final day of COP26—the annual United Nations Climate Change Conference held in Glasgow in 2021—more than 130 countries had made net-zero pledges of one kind or another, representing about 70 percent of the world’s UN-recognized nations. The private sector has also rushed in: more than 30 percent of the world’s 2,000 biggest public companies have committed to net-zero targets, including Amazon, Walmart, and ExxonMobil. 

A global ambition has coalesced. It’s an achievement of sorts. Limiting warming to 1.5°C requires reaching net zero by midcentury, according to the Intergovernmental Panel on Climate Change (IPCC). This doesn’t mean zero emissions; rather, it means that some remaining amount of positive emissions would be canceled out by “negative emissions”—that is, by carbon removals. Negative emissions can be generated by using ecosystems like forests, farms, and oceans to store more carbon, or by using industrial technologies like carbon capture and storage (CCS) to pull carbon from the atmosphere and store it in rock formations. When these negative emissions balance out the positive emissions—when the amount of carbon being taken out of the atmosphere equals the amount of carbon being put into the atmosphere—then net zero is reached.

Under such an arrangement, countries are free to continue burning fossil fuels, so long as they offset their emissions. For this reason, many climate advocates have been critical of net zero as a goal. One common critique is that net-zero pledges won’t stop the continued extraction and combustion of fossil fuels. Further, they’re aimed at a future that’s far enough away that present-day leaders won’t be accountable for what happens. “Net zero by 2050. Blah, blah, blah,” Greta Thunberg told a summit of young organizers just prior to COP26. 

It’s true that net zero is woefully insufficient. We also need to be talking about immediately phasing out fossil fuel production. But net zero is still a worthwhile transitional goal, because we don’t yet have all the technologies we need at scale for true climate repair. Climate change is a problem of stocks, not flows: we need not only to stop emitting carbon by switching to renewables, but to reduce the existing levels of carbon in our atmosphere. Net zero alone won’t get us there. Still, it gives us a way to buy time while we develop and deploy the technologies required for full decarbonization.

Yet net zero is harder than it looks. The difficulty isn’t just political—compelling countries and companies to make promises and abide by them—but epistemological. At the center of net zero is a knowledge problem: How do we know when we’ve gotten to net zero? Answering this question is surprisingly hard. 

There are two main challenges. First, there is immense technical complexity involved in accurately measuring both positive and negative emissions. Take positive emissions: how many are embodied in the manufacture of a car? You could measure how much carbon is produced by a single car factory, but a car has around 30,000 parts. Those parts might be sourced from suppliers around the world, each with their own carbon footprint. Further, the parts use different raw materials, and the extraction and transport of those materials carry emissions of their own. And that’s only one factory; there are some 300,000 car manufacturing facilities in the US alone.

Measuring negative emissions presents headaches of its own. For example, there are emergent methods for measuring how much carbon is stored in soil that hinge on spectroscopy, satellite-based sensors, and machine learning. But knowing what’s actually going on in the soil is complex, as conditions can vary even across a single farm. Knowing what’s going on with carbon in the deep ocean, as required for some carbon removal approaches, is even more challenging. Industrial technologies like CCS that store carbon geologically may seem easier to measure than biological systems, but, even then, understanding what happens to carbon that is stored thousands of feet underground isn’t easy.

Even if these difficulties are overcome, however, there is still the problem of carbon deception. This is the second main challenge involved in knowing when we’ve gotten to net zero: the presence of dishonest actors in the system. Think about Volkswagen rigging vehicles to cheat on emissions tests, or gas station chain owner Lev Dermen, who was found guilty of stealing $1 billion from US taxpayers through claiming tax credits for renewable biodiesel that didn’t exist. Behind the office-park facade of net zero lies a seedy hotbed of white-collar crime.

The knowability of net zero can’t be taken for granted, then. All those policymakers and executives are making net-zero pledges without a clue for how to see them through. How can these pledges possibly be fulfilled? How can the knowledge problem of net zero be solved?

The answer, apparently, is software. Companies large and small have begun to build “platforms” that promise to help organizations meet their net-zero promises. These platforms claim to be able to deliver “decarbonization-as-a-service” through the use of digital technologies that effectively track carbon. They are racing ahead of law and policy and performing de facto governance, creating new proprietary infrastructures for knowing and managing our planet.

If this new regime fails to reliably measure and monitor carbon—and it very well may—that will be bad for the climate. If it succeeds, that may also be bad for the climate, not to mention ecosystems more broadly as well as human communities. But right now, before the new computational systems are locked in, we may have a chance to intervene and shape them for the better.

Initial Tree Offering

In 2012, Derrick Emsley cofounded Tentree, a Vancouver-based clothing company that plants ten trees for every product sold. That sounds simple enough, but managing the planting projects turned out to be a challenge. “The hardest part was monitoring and verifying the work and claims we were making,” Emsley told The Hill’s Saul Elbein. Confirming that the trees they paid to plant were actually planted required costly in-person trips and lots of managerial overhead—it meant “traveling there, auditing them, making sure those trees were in the ground having an impact,” according to Emsley.

So Tentree started developing software to help streamline the process. That software became the basis of Veritree, a publicly available “planting management platform” unveiled in the fall of 2021. When a company that wants to offset their emissions signs up for Veritree, they get access to a user portal that lets them place orders for new tree plantings and displays metrics on their current tree holdings. These metrics are collected by the planters in the developing countries, who use custom-made “collect devices”—essentially modified smartphones designed to work in internet-limited environments—to take geotagged photographs of the trees. The trees get a unique digital token, and become digital inventory, hosted on the Cardano blockchain. Using a blockchain is supposed to reduce fraud by ensuring that trees are only counted once, so that the same forest can’t be claimed by multiple entities—a well-known problem in the world of carbon credits. 

Veritree is in its early days, but it hopes to become “an operating system for the restoration economy,” in Emsley’s words. An “Initial Tree Offering” was used to raise money for a “First Edition Forest,” which features trees in Madagascar, Indonesia, Nepal, Kenya, Senegal, and Haiti. The appeal for corporations hoping to make good on their net-zero pledges is obvious: they can invest in a reliable carbon offset program and obtain a real-time picture of precisely how much carbon is being offset, and share that information with the public. 

Veritree is far from the only platform hoping to dominate the net-zero space. A wide range of digital services is being developed by companies large and small. Established “Big Four” accounting firms like KPMG and Deloitte, as well as tech giants like Salesforce, are creating tools for Environmental, Social, and Governance (ESG) accounting that help measure the carbon footprint of firms, among other things. Some companies combine carbon measurement with a portal for purchasing offsets, such as the Atlanta-based startup Cloverly. Still others try to track negative emissions, like Veritree. Agreena, for example, is a Dutch startup that monitors changes in farmers’ fields after they switch to regenerative agriculture, and issues them e-certificates. Companies can sponsor these carbon reductions, and use Agreena’s platform to track them.

Atoms into Commodities

The creators of these new platforms believe that they can solve the knowledge problem of net zero with software. Given the proliferation of net-zero pledges, this is a profitable problem to solve. But the platforms all approach this problem in a particular way: they turn carbon into a tradable commodity. This points to a broader point: net zero, in its current configuration, is a market-based project. It requires creating a global market where offsets can be freely bought and sold. This market already exists, but it has much room to grow; Mark Carney, the former governor of the Bank of England, says it could be worth $100 billion. 

Creating that value, though, hinges on turning lively carbon atoms into a smooth commodity. And that task, in turn, hinges on code. The new platforms aim to improve and expand carbon markets by packaging carbon into a reliable product that can be easily bought and sold online. Their value proposition isn’t just about using digital tools to do superior carbon monitoring and accounting, but about disrupting traditional carbon markets by disintermediating them.

In traditional carbon markets, supply and demand is linked by retail traders who purchase carbon credits from suppliers and bundle them into portfolios, to be sold on to brokers or end buyers. This is an inefficient system, with too many middlemen; it is also rife with fraud. The new platforms want to cut the knot by connecting buyers and sellers—that is, the producers of positive emissions with the producers of negative emissions. Think of Veritree: corporations can purchase offsets by directly sponsoring planting projects in the developing world, without having to navigate a tangle of traders and brokers.

The platforms don’t just want to revolutionize existing carbon markets, however. They also want to create new ones. Climate Impact X is a Singapore-based carbon exchange that plans to build a “forest carbon marketplace” that uses remote sensing, artificial intelligence, and blockchain to “open participation to forests that were previously left out of the climate solution.” The idea is that a data-driven approach will lower the barriers of entry for the producers of negative emissions and enable more of them to participate. And, if given the chance, they probably will: there is a lot of money to be made. After languishing for many years, carbon offsets are trading at record highs.

Risk Factors

The rise of a decarbonization-as-a-service sector may make it easier to commodify and exchange carbon. But what if it doesn’t actually help address climate change? There is always the risk that the platforms could be algorithmically flawed—that their software isn’t accurately quantifying emissions, whether positive or negative. But there are also deeper risks, ones that can’t be mitigated by tweaks to the code because they’re inherent in a market-based system.

First, ask yourself what any of these platforms want. Continued emissions; more exchange. The platform becomes a vested actor. The company is not oriented toward the phaseout of fossil fuels. Rather, it wants to maintain its own life, its own revenue stream. If you were designing a net-zero platform with the goal of working towards full decarbonization, it would be a time-limited project. But that would be incompatible with market imperatives: nobody wants to start a business with an expiration date. This is why the companies running these platforms will have an incentive to encourage a version of net zero with continued residual emissions—not a version of net zero oriented toward a future that would make them obsolete.

More broadly, a market-based system cultivates a tendency toward cheapness that’s hard to square with the costs of carbon removals. Why buy a carbon offset for $50 or $100 a ton when you can buy one for $10 a ton? How is a sustainability manager at a company going to justify that to higher-ups or shareholders? Yet that $10 offset is much less likely to be trustworthy, since proper monitoring and verification costs money. Moreover, the most permanent kind of carbon removals, performed through industrial technologies like CCS (Carbon Capture and Storage), still cost hundreds of dollars per ton. Yes, some platforms will attempt to introduce boutique or premium removals. But no matter how sophisticated the platforms for carbon management become, market logic will inevitably push companies away from these more expensive and more robust offsetting techniques, and toward cheaper and less effective ones.

Another issue with a market-based system is that offsets can be purchased by any market actor. But talk to any climate scientist and they’ll tell you that to reach net zero, removals should only be used to compensate for emissions from sectors that are hard to transition to renewables, like aviation or agriculture. In an open market, offsets can be bought up by anyone, including entities that have no future in a post-carbon economy such as fossil fuel companies. That’s bad, because we need to allocate removal capacity to sectors that we want to preserve and whose emissions are truly hard to abate. Platforms can make carbon markets more efficient, but they can’t perform the sorting function needed to prioritize the offsetting of particular activities and industries.

A final problem with a market-based system is that it turns the platforms themselves into black boxes. As profit-seeking enterprises, they must protect their algorithms and data, as otherwise their competitors might gain an advantage. But this opacity obstructs the learning and experimentation process that’s required to combat climate change. We need a broad-based scientific effort to help figure out what works in terms of curbing emissions and removing carbon from the atmosphere. That can’t happen if all the data is locked away on proprietary platforms, hidden from view.

We wouldn’t just be entrusting the platforms with data, however. We would also be entrusting them to play a major role in the governance of ecosystems, with significant social consequences. Creating negative emissions is never a neutral, purely technical process—it always involves political choices. For example, communal land title is sometimes seen as a risk to the permanence of carbon credits.  This led to the exclusion of communally held land from one carbon credits project in Cambodia, as research by Sarah Milne and Sango Mahanty has shown. As these calculative practices scale up in digital realms, the repercussions will be turbocharged. 

Getting to net zero matters, but how we get there is just as important. Put another way, different attempts to quantify carbon produce different kinds of social relations. Ceding unilateral control over these choices to corporate platforms—letting them decide which kinds of net-zero social relations to make—would be a significant mistake. 

Public and Planetary

Software is essential for facing the challenge of climate change. We need to build a computational infrastructure for tracking carbon. The impact of humans on ecosystems has been so deep and so complex that digital monitoring and modeling are essential for ecological repair. As Benjamin Bratton writes, “planetary-scale computation” could “contribute to the comprehension, composition, and enforcement of a shared future that is more rich, diverse, and viable.” But right now, we are headed towards a world of proprietary platforms, repeating the same mistakes that we made with the development of the internet. 

Shouldn’t knowledge about the Earth’s carbon flows belong to communities and the commons? Shouldn’t the benefits of such data go to the people who are working to remove and sequester carbon—tending the trees, soil, seagrass meadows, and injection wells—and to society more broadly? Shouldn’t the political choices about how to quantify carbon—and, by extension, about what kind of social relations to create in pursuit of net zero—be made democratically, rather than by executives and shareholders? 

We need platforms that monitor our world for the sake of collective management of a shared commons—not platforms that measure our world for the sake of profiting from the data created from it, extracting value from the digital layer the same way that value is extracted from the mines, forests, and soils of our physical world. Plundering the earth and then repeating the plunder in the metaverse is dystopia squared.

We could imagine platforms that are so much better: better because they actually get us to net zero and address other sustainability and social challenges. The public sector must be central here. It alone can bear the costs of accurate monitoring of both positive and negative emissions—costs that disincentivize companies from buying high-quality carbon offsets, no matter how many advanced digital tools for carbon management emerge. 

Public platforms can also help dislodge another obstacle to climate action: community resistance. We will have to build an incredible amount of new transmission lines, solar panels, wind farms, geothermal facilities, factories, mines, and more to pull off a wholesale transformation of our energy system. But when people don’t understand the contours of the problem, or don’t benefit from the solution, they will oppose such initiatives. We’ve seen already that technocratic climate projects without public input are likely to face hostility from local communities. If such communities aren’t brought into the design process, the energy transition is imperiled. 

Better platforms are not a panacea for this challenge, but they could help. Imagine a platform for tracking carbon that isn’t just operating quietly in the background, but is highly visible, easy-to-use, open source, and closely integrated with community planning projects. Such platforms could enable people to collaborate on roadmaps for decarbonizing their town or region. They could integrate data around carbon flows and other ecosystem attributes with real-time visualization tools, in order to anticipate different scenarios and deliberate around possible tradeoffs. The platforms could facilitate a more participatory ecology, and thus generate community buy-in for decarbonization measures.

More broadly, public planetary data infrastructures would enable the inhabitants of Earth to know their world. Knowledge about one’s environment should be a basic human right. Imagine a world where you could know what’s in your air, what’s in your drinking water, what species are in the forest near you and how much carbon it’s storing, or what the health of your local lake is like. You could use this information not just to gauge the environmental risks you face, but also to discover which companies are degrading the ecological systems you live in and hold them accountable.

This isn’t to suggest that every piece of data needs to be public. As the Indigenous scholar Stephanie Carroll Rainie and her colleagues have explained, open data can be in tension with the rights of Indigenous peoples to govern their own data. In particular, it may cause tensions for communities that continue to experience data extraction under settler-colonial frameworks, and who are working to establish their own frameworks. When it comes to knowing net zero, the important thing is that the computational systems for monitoring emissions and carbon flows are publicly or community-owned, accessible to community members, and democratically governed. What this actually looks like on local, regional, and planetary scales needs to be established through participatory processes that will require time and trust.

Across the Binaries

How do we get there? The answer is not fancy. It involves assembling a coalition for public ecological data infrastructures, drawn from several existing communities. There’s the long-standing open source software movement that could participate. There’s also the open data movement in science, and the movement for Indigenous data sovereignty. Environmentalist NGOs are tracking data about emissions, from larger projects like the Environmental Defense Fund’s methane-tracking satellites to databases like the Carbon Disclosure Project. There are a number of people across these fields who might join a movement for open and public planetary data. So why doesn’t this movement exist in a more mainstream way, or even as a common demand at climate protests? 

One challenge is the fact that this is an anticipatory mobilization. We’re not mobilizing against something that’s already happened, we’re acting defensively based upon trends that are just beginning to emerge. Proprietary carbon management platforms are still in their infancy; their full impact might not be felt for many years. 

Then there’s the cultural divide between the worlds of climate activism and tech activism. Organizations concerned with climate change may be inclined to see the digital as outside their core mission; after all, many people became involved in this space because they loved being outside, not because they wanted to think about algorithms. Moreover, such people may be actively opposed to technological interventions, or understandably dismiss net zero as a narrow, technocratic goal. Meanwhile, people with expertise in algorithmic justice issues might not be tracking developments in the environmental sphere. Their attention is likely devoted to a multitude of other concerns.

But each of these communities has critical things to contribute. Climate activists can help us avoid the trap of “platform determinism”—that is, the risk of fetishizing the platforms as mythically powerful actors, instead of centering the choices made by the humans who design the platforms. Such activists also bring expertise in movement building: they know how to put pressure on investors and companies, and how to form relationships with policymakers. On the other hand, those who are working on the politics of data, whether from socialist, decolonial, or Indigenous perspectives, have valuable experience in identifying the problems with data appropriation, access, and use, as well as in creating just data frameworks. And technologists can put their knowledge of quantification and design to work building platforms that are genuinely usable and inspiring. 

The knowledge problem of net zero is difficult but not insurmountable. Solving it the right way will require a group effort. We must build data infrastructures that embody multiple ways of knowing and understanding our world, and that help us advance both ecological and social ends, before corporations conquer this space for themselves.

Holly Jean Buck is the author of Ending Fossil Fuels: Why Net Zero Is Not Enough and After Geoengineering: Climate Tragedy, Repair, and Restoration. She is an assistant professor of Environment and Sustainability at the University at Buffalo.

This piece appears in Logic's issue 16, "Clouds". To order the issue, head on over to our store. To receive future issues, subscribe.