Bridging the Grid Gap: Data Center Sustainability in Fossil-Fuel Dependent Regions

Calli O'Neal, Roberts Environmental Center at Claremont McKenna College

We invited students from Claremont McKenna College’s Roberts Environmental Center (REC) to provide their perspectives on key issues related to carbon markets. Here is a series of articles they have developed as part of a market research project in which Tradewater participated. These views are theirs and do not necessarily reflect Tradewater’s views — but in the interest of stimulating conversation we think they are valuable to share.

Key Points

  • AI Boom Drives Energy Demands: Surging AI technologies are straining data centers, with U.S. energy use projected to grow 12% annually

  • Midwest Faces Renewable Bottlenecks: Key data center regions like Illinois struggle with fossil-fuel dependence and slow renewable grid integration

  • Immediate Carbon Reductions Matter: Early carbon cuts via local offsets and infrastructure upgrades deliver more climate benefit than delayed, speculative solutions

The world is witnessing an unparalleled surge in AI-driven technologies, revolutionizing industries from healthcare to finance. However, this rapid growth comes with an equally significant challenge: increasing energy demand. Data centers, the backbone of AI and cloud computing, are increasing in size and creating skyrocketing energy demands. Meanwhile, the tech industry, which leads in AI product development and sales, has made some of the most ambitious net-zero commitments of any sector and faces mounting pressure to decrease its carbon footprint. 

While striving to stay at the forefront of innovation, tech companies must ask themselves: how can their AI coexist with their sustainability objectives? 

The AI Surge and the Energy Dilemma

AI applications require immense computational power, and with U.S. data center energy demand projected to grow by 12% annually until 2030, the strain on the grid is intensifying at an unprecedented rate. According to recent S&P Global estimates, this AI demand could increase power usage by 150–250 TWh, resulting in an additional 40–67 million tons of CO2 emissions annually by 2030, comparable to the current total emissions of the country of Austria.  

Renewable Gaps and Regional Realities

The U.S., home to a third of the world’s data centers, is rapidly expanding its data center footprint, with the Midwest emerging as a key hotspot. The region’s central location and infrastructure potential make it attractive, but it faces significant renewable energy bottlenecks. In states like Indiana, Illinois, and Wisconsin, power grids remain heavily reliant on fossil fuels and lack the capacity to quickly integrate large-scale renewable energy. 

For instance, Chicago, Illinois, ranks third in data center capacity after Dallas and will require at least 8,500 megawatts (MW) of new capacity between 2030 and 2049 to meet growing energy demand, driven largely by data center expansion. The 25 new data centers proposed in Illinois alone would consume as much energy as the state’s five nuclear plants currently generate. 

This raises a critical question: how can we address immediate grid gaps amidst soaring demand? Coal plants are shutting down, and while natural gas is a step forward, new EPA regulations cast doubt on its cost-effectiveness for new proposals. Meanwhile, renewable energy development lags behind expectations, even with incentives from the Inflation Reduction Act. 

Faced with these challenges, some companies are turning to expensive nuclear power. Microsoft’s recent move to reopen Pennsylvania’s Three Mile Island nuclear plant, a site infamous for the worst nuclear accident in U.S. history, shows a new willingness to revisit controversial solutions. Companies like Amazon have invested close to a billion dollars in small modular nuclear reactors that offer the promise of stable, localized power, but they remain unproven at scale, with the first units unlikely to come online until 2030. 

This leaves us at a sustainability crossroads: accept the realities of today’s grid and find ways to take action now, or delay decarbonization further. 

The Time Value of Carbon

Delaying action risks exacerbating the climate crisis. But what often is less taken into consideration, is the “time value of carbon” and other greenhouse gases (GHGs).  GHG reductions achieved today through infrastructural improvements or high-quality carbon offsets provide immediate climate benefits that often can outweigh the impact of solutions like expensive carbon capture or nuclear energy development that can only be deployed later. When creating Science Based Targets initiative (SBTi) goals or 2030 pathway strategies we should all keep in mind that earlier GHG reductions are more impactful than equivalent reductions achieved later. 

The Unique Value of Place-based Offsets

Regionally focused offsets, especially those matched to local emissions sources, provide unique opportunities to drive equity and credibility in climate action strategies. By investing in local infrastructure improvements: retrofitting buildings for energy efficiency or plugging methane-leaking orphaned oil and gas wells, companies can simultaneously advance climate goals and support the communities they operate in. This localized approach ensures decarbonization efforts address the needs of affected communities, aligning operational expansion with equitable development and producing high-integrity sustainability claims that are both impactful and credible. 

PPAs, VPPAs and the Offset Advantage

Power Purchase Agreements (PPAs) and Virtual Power Purchase Agreements (VPPAs) are widely used by tech companies to achieve Scope 2 emissions neutrality. However, these mechanisms often fall short of providing immediate, localized emissions reductions. The reliance on volatile energy markets and long-distance renewable projects, such as wind farms located far from operational centers, can lead to inefficiencies and higher costs for energy that isn’t directly utilized. 

While PPAs and VPPAs support the development of renewable energy infrastructure, they do little to address existing grid emissions or localized fossil fuel dependencies. Offsets, when done right, could help bridge this gap. Initiatives like Tradewater’s methane abatement projects enable tech companies to achieve near-term reductions while transitioning to renewable energy. Furthermore, “like-for-like” offsetting, where the GHG reduction from offsets aligns with the type of gas being emitted, enhances credibility. For instance, data centers reliant on natural gas can invest in methane abatement as an effective step toward thoughtful GHG neutrality. 

A Balanced Approach to Data Center Sustainability

Achieving sustainability in the data center industry requires a practical and balanced approach. Ambitious net-zero goals must be paired with actionable steps, including investments in energy infrastructure and the integration of high-quality offsets. While carbon removal technologies often dominate the conversation, immediate reduction offsets provide a more accessible and impactful solution for measurable progress without undermining long-term ambitions. 

In regions that are targeting data center expansion, offsets that reduce emissions while delivering local benefits are critical. As data center development increases, these projects can help mitigate strain on local grids and reduce vulnerabilities to outages and climate-related disasters. 

The AI revolution offers the tech sector a unique opportunity to lead by example. By adopting innovative offset solutions and regional decarbonization strategies, companies can align their growth with a sustainable future, for both their industry and the planet. 

Calli O'Neal

Calli is a senior at Pitzer College studying environmental analysis on the policy track with a Spanish minor. She is passionate about equitable global supply chains and helping decarbonize our world. She lives out of her backpack in her free time
and prefers to be on the road.

Back to Blog

Permanence

Emission reductions are considered permanent if they are not reversible. In some projects, such as forestry or soil preservation, carbon offset credits are issued based upon the volume of CO2 that will be sequestered over future decades—but human actions and natural processes such as forest fires, disease, and soil tillage can disrupt those projects. When that happens, the emission reductions claimed by the project are reversed.

The destruction of halocarbon does not carry this risk. All destruction activities in Tradewater’s projects are conducted pursuant to the Montreal Protocol , which requires “a destruction process” that “results in the permanent transformation, or decomposition of all or a significant portion of such substances.” Specifically, the destruction facilities Tradewater uses must meet or exceed the recommendations of the UN Technology & Economic Assessment Panel , which approves certain technologies to destroy halocarbons, including the requirement that the technology achieve a 99.99% or higher “destruction and removal efficiency.” Simply put, this means that Tradewater’s technologies ensure that over 99.99% of the chemicals are permanently destroyed. During the destruction process, a continuous emission monitoring system is used to ensure full destruction of the ODS collected.

Accuracy

Some carbon offset projects necessarily rely on estimations or assumptions when calculating the emission reductions from project activities. Forestry projects, where developers make assumptions about the carbon that will be sequestered over future decades if trees are conserved, are a perfect example. Such projects sometimes result in an overestimation of the environmental benefit of the project.

Tradewater’s halocarbon projects avoid the issue of overestimation by consistently conducting extremely precise testing and measurement of the amount of refrigerant destroyed in each project.

  • Every container of ODS that Tradewater destroys is weighed by a third-party using regularly calibrated scales. The ODS is then sampled by a third-party and analyzed by an accredited refrigerant laboratory to determine its species and purity. These two steps combine to ensure that credits are issued only for the precise volume and type of refrigerant destroyed.
  • The destruction facilities that Tradewater uses continuously monitor the incineration process during destruction events to ensure that over 99.99% of the ODS is destroyed. This monitoring is mandated by regulatory protocols and is part of the verification process to which projects are subjected.
  • Tradewater accounts for the project emissions created during the collection, transport, and destruction of ODS, and the number of offsets issued is reduced by a corresponding amount. The protocols that we use also build in other reductions to account for substitute chemicals that will be used to replace the destroyed refrigerants. Tradewater publishes this information in the documentation for all its ODS destruction projects. These documents outline how the material was obtained, the project emissions calculations, the test results, and the amount and type of ODS chemicals destroyed, among other information.

    • Additionality

    It is a basic requirement of all carbon offset projects that the underlying project activities are additional. “Additional” means that the projects would not happen in the absence of a carbon market. Tradewater’s halocarbon projects simply would not happen – and the gases would be left to escape into the atmosphere – without the sale of the resulting carbon offset credits. This is because there is no mandate to collect and destroy these gases. It is still permissible to buy, sell, and use halocarbons that were produced before the ban. There are other reasons halocarbon destruction projects are additional:

    • There are no incentives or financial mechanisms to encourage halocarbon destruction. According to the International Energy Agency and United Nations Environment Program, “there is rarely funding nor incentive” to recover and destroy ozone depleting substances in storage tanks and discarded equipment. And collecting, transporting, and destroying halocarbons is time-intensive and expensive. The burden to collect and destroy these gases therefore remains prohibitive outside of carbon offset markets—meaning that if organizations like Tradewater do not do this work, nobody else will.
    • Countries are not focused on the need to collect and destroy halocarbons. The Montreal Protocol has been celebrated as a success because of its production ban. This success, however, ignores the legacy gases produced before the ban and is a blind spot for government regulators. In the U.S., for example, the Environmental Protection Agency (EPA) developed a Vintaging Model in the 1990s to estimate the quantify of ozone depleting substances left in circulation. Based on the inputs and assumptions put into the model, the EPA predicted that no CFCs would be available for recovery beyond 2020 in the United States. But this prediction did not prove accurate. Tradewater has collected and destroyed more than 1.5 million pounds of CFCs globally in recent years and continues to identify thousands of pounds per week.
    • International carbon accounting standards do not require corporations to measure or track emissions tied to halocarbons, and refrigerants are specifically excluded from Science Based Targets initiative (SBTi) commitments. These commitments derive from emissions reporting under the GHG Protocol, which requires companies to report on emissions only from new generation refrigerants, such as hydrofluorocarbons (HFCs), but does not establish any obligation to report inventories or emissions of refrigerants still in use, such as CFCs and HCFCs. All these factors combine to make Tradewater’s carbon offset projects highly additional. As Giving Green, an initiative of IDinsight, concluded: “Tradewater would not exist without the offset market, so this element of additionality is clearly achieved.” The case for additionality is not so clear for some other project types, such as forestry and landfill gas carbon projects. For example, some forests are already being conserved for their beauty, or for use as parks, and generate carbon offset credits only because those conservation efforts do not yet have full formal protection in place to avoid deforestation in the future. Similarly, methane from landfills can be used to make electricity or captured as compressed natural gas, thereby creating additional revenue streams to support the activities, beyond the sale of carbon credits.