AI’s Impact on Big Tech’s Sustainability Commitments

Google and Microsoft maintain net zero pledges despite AI’s growing emissions

Two of the world’s largest technology companies have published sustainability reports that reveal a growing tension between climate ambition and operational reality. Google plans to run on 100% carbon-free energy by 2030. Microsoft aims to be carbon negative by the same date. However, both companies now face rising emissions directly linked to artificial intelligence.

The expansion of AI creates substantial energy demands. Training a single large language model can consume over 1,200 megawatt-hours of electricity. That’s enough to power approximately 120 homes in the United States for an entire year. The same training process generates roughly 500 metric tons of carbon dioxide, equivalent to the emissions from driving between New York and San Francisco 438 times.

For UK businesses, this matters beyond the headlines. As AI tools become embedded in procurement systems, customer service platforms, and supply chain management, the carbon footprint of digital services will increasingly factor into sustainability reporting requirements. Companies that rely on cloud computing and AI-enabled services need to understand how their technology choices affect their own emissions data.

Data centers currently account for between 1% and 2% of global electricity consumption. Artificial intelligence represents approximately 15% of that demand. The International Energy Agency projects that global data center energy consumption will double by 2030 as AI deployment accelerates across industries.

Microsoft reports 29% emissions increase since 2020

Microsoft’s May 2025 sustainability report shows emissions have climbed 29% above the company’s 2020 baseline. Compared to 2019 levels, the increase reaches 48%. The company attributes this primarily to data center expansion required to support AI infrastructure.

Despite these rising figures, Microsoft continues to describe its carbon negative goal as achievable. The company now frames the commitment as a marathon rather than a sprint, acknowledging that the path will be longer and more complex than initially anticipated. This language shift reflects the scale of the challenge posed by AI’s energy requirements.

Google reported 12 million metric tons of emissions in 2024. The company has improved its freshwater replenishment rate to 64%, up from 18% in 2023. Nevertheless, total water consumption increased by 28% in a single year, from 24 billion liters to 30 billion liters. This surge correlates directly with expanded data center operations.

Both companies maintain their original target dates. However, the reports indicate that reaching these goals will require technological breakthroughs and infrastructure changes that are not yet fully deployed. Google now refers to its 2030 target as a moonshot, suggesting the difficulty level has increased substantially.

Cornell study quantifies AI’s projected environmental impact

Research published by Cornell University in November 2025 provides specific projections for AI’s carbon footprint. If current growth trends continue, the industry will emit between 24 and 44 million metric tons of carbon dioxide annually by 2030. This range is equivalent to adding between 5 and 10 million cars to American roads each year.

The same study forecasts water consumption will increase by 731 to 1,125 million cubic meters annually. That volume matches the household water usage of between 6 and 10 million Americans. These figures represent the combined effect of AI training, deployment, and the infrastructure required to support both activities.

Consequently, businesses that use AI-enabled services will face questions about embedded emissions in their supply chains. Scope 3 emissions reporting already requires companies to account for the carbon footprint of purchased goods and services. As regulatory scrutiny increases, the energy profile of cloud computing and AI tools will become a material factor in compliance documentation.

The tech sector’s total carbon footprint already represents approximately 4% of global greenhouse gas emissions. This exceeds the aviation industry’s contribution. Without significant improvements in efficiency and grid decarbonization, AI growth threatens to push this percentage higher.

Energy demands create compliance pressure for UK businesses

For small and medium enterprises in the United Kingdom, the environmental cost of AI matters in practical terms. Many businesses now rely on cloud-based accounting systems, customer relationship management platforms, and automated procurement tools. All of these services run on data center infrastructure that generates emissions.

Public sector suppliers face particular scrutiny. PPN 06/21 requires organizations bidding for government contracts above certain thresholds to publish carbon reduction plans. These plans must include Scope 3 emissions, which covers the services businesses purchase. Therefore, the carbon intensity of your technology providers becomes part of your own compliance obligations.

Additionally, businesses seeking certification under standards like ISO 14001 or preparing for mandatory climate reporting under emerging UK regulations need accurate emissions data. The carbon footprint of digital services represents a growing portion of many companies’ total emissions profiles, particularly for office-based operations with limited physical manufacturing.

Furthermore, customer expectations are shifting. Corporate buyers increasingly ask suppliers about their environmental performance. Tender documents now routinely include questions about carbon reduction strategies and emissions reporting. Your choice of technology partners affects your ability to answer these questions credibly.

Meanwhile, the cost of energy itself continues to rise. Data centers require constant cooling and round-the-clock power. These operational expenses ultimately flow through to service pricing. Businesses should anticipate that AI-enabled services may become more expensive as providers factor in both energy costs and the price of carbon offsets.

Industry responses focus on efficiency and renewable energy

Technology companies are pursuing several strategies to reconcile AI expansion with climate commitments. Microsoft and Google have both invested billions in carbon capture startups. These companies also purchase large volumes of carbon credits to offset emissions that cannot yet be eliminated through operational changes.

Renewable energy certificates represent another major investment area. These certificates allow companies to claim renewable power even when their data centers physically draw electricity from grids that include fossil fuel generation. Critics argue this approach amounts to accounting rather than genuine decarbonization, particularly when certificates are purchased from distant renewable projects.

Cooling technology improvements offer more direct emissions reductions. Liquid cooling systems can reduce water consumption by up to 86% compared to traditional air cooling, according to research published alongside the Cornell study. Server efficiency gains also help, though these improvements have not kept pace with the growth in AI workloads.

Data center location decisions now factor in regional water availability and grid carbon intensity. Siting facilities in areas with low water stress and high renewable energy penetration can reduce environmental impact substantially. Combined with grid decarbonization efforts, researchers estimate these strategies could cut carbon emissions by approximately 73% and water use by 86%.

However, these solutions require time and capital investment. For the next several years, the technology sector’s emissions will likely continue rising even as companies implement mitigation measures. This creates a gap between current trajectories and the targets both Microsoft and Google have set for 2030.

What UK businesses need to know about AI and carbon reporting

• Training a single large AI model generates approximately 500 metric tons of carbon dioxide and consumes over 1,200 megawatt-hours of electricity.
• Microsoft’s emissions have increased 29% since 2020 despite active sustainability programs, primarily due to data center expansion for AI infrastructure.
• Data centers currently use between 1% and 2% of global electricity, with AI representing roughly 15% of that consumption.
• By 2030, AI is projected to emit between 24 and 44 million metric tons of carbon dioxide annually if current growth continues.
• Water consumption from AI infrastructure may increase by 731 to 1,125 million cubic meters per year by 2030.
• The technology sector already accounts for approximately 4% of global greenhouse gas emissions, surpassing the aviation industry.
• Businesses using cloud computing and AI tools should expect these services to factor into Scope 3 emissions reporting requirements.

Commercial implications for sustainability reporting and procurement

The carbon intensity of AI affects carbon reporting compliance for UK businesses in several ways. First, companies must account for emissions from purchased services when calculating Scope 3 figures. Cloud computing, software as a service, and AI-enabled platforms all generate upstream emissions that belong in your carbon footprint.

Second, supplier due diligence becomes more complex. When evaluating technology vendors, businesses should request information about data center locations, renewable energy usage, and carbon reduction commitments. This information directly affects the accuracy of your own emissions calculations and the credibility of published carbon reduction plans.

Third, costs will likely increase. Energy represents a significant portion of data center operating expenses. As electricity prices rise and carbon pricing mechanisms expand, technology providers will face pressure to pass these costs through to customers. Budget planning should account for potential price increases in cloud services and AI tools.

Public sector suppliers need to consider these factors when preparing tender responses. Procurement processes increasingly evaluate environmental credentials alongside price and capability. Your ability to demonstrate that you understand and manage the carbon footprint of your digital infrastructure can influence contract awards.

Moreover, businesses working toward net zero targets must include technology emissions in their reduction strategies. Simply switching to a cloud provider does not eliminate emissions; it transfers them to a different part of your value chain. Effective carbon management requires visibility into these emissions and active engagement with suppliers about their reduction plans.

Questions businesses should ask technology providers

When selecting or reviewing relationships with technology vendors, particularly those offering AI-enabled services or cloud infrastructure, several questions can help you understand the environmental implications. Where are your data centers located? Regional differences in grid carbon intensity mean that identical services can have vastly different carbon footprints depending on physical location.

What percentage of your electricity comes from renewable sources? Ask for specifics about whether this represents physical renewable energy or renewable energy certificates. The distinction matters for understanding actual emissions reductions versus accounting mechanisms.

How do you measure and report Scope 3 emissions? This question matters because your organization’s Scope 3 emissions include the Scope 1 and 2 emissions of your suppliers. Vendors who cannot provide this information create gaps in your own reporting.

What efficiency improvements are you implementing? Cooling technologies, server utilization rates, and infrastructure design all affect energy consumption. Providers actively working on efficiency should be able to describe specific initiatives and their expected impact.

Do you offer tools to help customers understand their usage-related emissions? Some providers now offer carbon calculators or emissions reporting specific to customer accounts. These tools support your own reporting obligations and help identify opportunities to reduce emissions through changed usage patterns.

The gap between ambition and implementation

Microsoft and Google’s reaffirmation of climate targets demonstrates continued commitment to sustainability goals. However, the language in recent reports shows increased caution about achievability. Microsoft’s shift from sprint to marathon and Google’s moonshot framing both acknowledge that original timelines may have underestimated the challenge.

This matters for businesses that have based their own sustainability strategies partly on cloud migration and digital transformation. The assumption that moving to major cloud providers automatically improves your carbon profile may not hold as strongly in an AI-driven computing environment. Due diligence and ongoing monitoring become essential.

Industry-wide emissions trends suggest the problem extends beyond individual companies. The technology sector’s carbon footprint continues growing even as leading firms invest heavily in mitigation. This indicates structural challenges that will require policy interventions, technological breakthroughs, or fundamental changes in how AI systems are designed and deployed.

For procurement professionals and sustainability managers, this creates uncertainty. Contracts signed today may span periods when energy costs, carbon prices, and regulatory requirements all shift substantially. Building flexibility into vendor agreements and maintaining dialogue about emissions management can help businesses adapt as the situation evolves.

Official guidance and industry standards

The UK government provides resources on carbon reporting through the Environmental Reporting Guidelines published by the Department for Energy Security and Net Zero. These guidelines cover requirements for calculating and disclosing Scope 1, 2, and 3 emissions.

For businesses working on public sector contracts, Procurement Policy Note 06/21 details the carbon reduction plan requirements. This includes expectations for Scope 3 reporting and guidance on what constitutes adequate disclosure.

The International Energy Agency publishes data and analysis on data center energy consumption that helps contextualize the sector’s environmental impact. Their projections for energy demand growth provide useful benchmarks for understanding future trends.

Professional bodies like the Institute of Environmental Management and Assessment offer training and guidance on environmental reporting standards. Their resources can help businesses interpret complex requirements and implement effective measurement systems.

For businesses seeking structured support with carbon reporting and compliance requirements, resources like practical training on emissions measurement can help teams build internal capability and confidence in managing these obligations.