How Big Tech Is Tackling Climate Change: A Guide for IT Decision-Makers

Big tech companies are investing heavily in renewable energy and carbon removal to offset the massive power demands of artificial intelligence and cloud computing, though their progress is complicated by expanding data center footprints and indirect supply chain emissions. This guide breaks down the specific climate pledges of major providers, the reality of their operational emissions, and what these sustainability efforts mean for businesses selecting AI and cloud infrastructure.

What Are the Core Climate Pledges of Major Tech Providers?

Major technology firms have set voluntary net-zero targets that often precede the deadlines established by the Intergovernmental Panel on Climate Change (IPCC). However, the structure and scope of these pledges vary significantly across the industry.

While these commitments represent substantial financial investments, evaluating them requires understanding the difference between direct operational emissions (Scope 1 and 2) and indirect value chain emissions (Scope 3). Google was the first major company to achieve a net-zero lifetime net carbon footprint, but the industry as a whole is now shifting its focus to the much harder challenge of decarbonizing their broader supply chains and addressing the energy intensity of AI.

How Data Centers and AI Are Driving Energy and Water Consumption

The rapid expansion of artificial intelligence is fundamentally altering the energy profile of the technology sector. Digital technology currently accounts for 1 to 2 percent of global emissions, but some analyses place the broader ICT sector’s footprint as high as 2 to 3.9 percent of worldwide emissions—a figure comparable to global aviation fuel emissions.

The constant push for scale in AI has led companies to develop hugely energy-intensive computational models that optimize for accuracy over more efficient, sustainable alternatives. Training and running these models requires massive data centers, which are notoriously resource-heavy. Furthermore, the physical infrastructure supporting these centers carries a heavy water footprint; research indicates that water usage in semiconductor manufacturing will double by 2035, driven largely by the demand for AI chips.

This growth has led to scrutiny over how companies report their environmental impact. A Guardian analysis suggests that between 2020 and 2022, emissions from company-owned data centers were likely more than seven times greater than official reports indicated, with companies relying on carbon credits to mask the true operational costs of AI systems. The tension between AI expansion and climate goals is evident in infrastructure planning; for example, Microsoft is considering ending its 24/7 clean energy goal in Virginia due to the rapid construction of electricity-hungry data centers in the region.

Why Scope 3 Supply Chain Emissions Remain the Biggest Hurdle

For hardware and cloud providers, Scope 3 emissions—indirect emissions that occur in a company’s value chain, including manufacturing, shipping, and product use—make up the vast majority of their total carbon footprint. These emissions are difficult to measure and control because they rely on the practices of external suppliers.

Apple provides a clear case study in Scope 3 management. The company has cut its total carbon emissions by about 60% since 2015, a reduction driven primarily by reshaping its supply chain rather than just optimizing its retail stores and offices. In 2024, suppliers using renewable electricity helped avoid 21.8 million tonnes of emissions, a 17% increase over the previous year. Today, Apple has 17.8 gigawatts of renewable electricity operational across its global supply chain.

"We’re incredibly proud of the progress we’re making toward Apple 2030… As we get closer to 2030, the work gets even harder – and we’re meeting the challenge with innovation, collaboration and urgency." — Lisa Jackson, Apple’s Vice President of Environment, Policy and Social Initiatives

Microsoft is also addressing this by requiring its suppliers to report on their emissions, forcing transparency deeper into the hardware manufacturing pipeline. However, convincing thousands of global suppliers to transition to renewable energy remains a massive, ongoing logistical and financial challenge.

E-Waste and the Circular Economy in Hardware Manufacturing

The environmental cost of big tech extends beyond energy consumption into the physical materials required to build servers, smartphones, and networking equipment. Computing technologies rely heavily on minerals that are often procured under difficult and exploitative conditions, creating significant ecological and social externalities.

To mitigate this, companies are investing in the circular economy. Apple has increased its use of recycled cobalt in its batteries and utilizes recycled aluminum in its enclosures. However, critics note that increasing recycled material usage fails to fully address the core issues of overproduction and the growing volume of electronic waste.

When data center servers reach the end of their useful life—often accelerated by the need to upgrade hardware for newer AI workloads—they become e-waste. While tech companies are improving their internal hardware recycling programs, the global e-waste recycling rate remains low. For IT decision-makers, the circular economy means evaluating not just how a vendor powers their data centers, but how they manufacture, lifecycle-manage, and ultimately dispose of the physical hardware that powers the cloud.

What This Means for IT Decision-Makers Choosing Cloud and AI Tools

For business owners, IT managers, and procurement leaders, evaluating a vendor’s climate pledge requires looking past annual net-zero marketing claims. The choices your company makes regarding cloud providers and AI tools directly contribute to your own Scope 3 emissions, which are increasingly subject to regulatory scrutiny under frameworks like the European Union’s Corporate Sustainability Reporting Directive (CSRD) and the U.S. SEC’s climate disclosure requirements.

When selecting an AI or cloud provider, IT decision-makers should prioritize the following:

• Hourly Energy Matching: Move beyond vendors that rely on annual Renewable Energy Certificates (RECs). Look for providers utilizing granular tracking, such as Google’s Time-based Energy Attribute Certificates (T-EACs), which verify that clean energy is being consumed on an hour-by-hour basis, ensuring actual grid decarbonization rather than accounting offsets.
• Scope 3 Transparency: Require vendors to disclose their supply chain emissions and their specific targets for hardware manufacturing. Microsoft’s mandate for supplier reporting is a strong baseline for vendor transparency.
• Policy Alignment: Corporate climate responsibility is not just about operations; it is about lobbying. InfluenceMap reports that despite positive internal programs, major tech companies remain members of cross-sector industry groups that lobby against binding climate measures necessary to meet global climate goals. Align your vendor choices with your own corporate governance standards.

Practical Takeaway for Enterprise Buyers

Before signing a multi-year cloud or AI infrastructure contract, request the vendor’s latest environmental progress report. Specifically, verify if they use hourly renewable energy matching rather than annual offsets, and check if their Scope 3 reduction targets include explicit, measurable milestones for semiconductor manufacturing and data center hardware lifecycle management. If a vendor cannot provide this data, their climate commitments may not align with your enterprise sustainability reporting requirements.

Frequently asked questions

Will technology fix climate change?

Digital technology currently accounts for a small fraction of global emissions, but its potential to enable a 20% global emission reduction by 2030 is significant if managed correctly. However, the unchecked growth of AI and data centers threatens to outpace these efficiency gains, meaning technology is an enabler of climate solutions rather than an automatic fix.

Is ChatGPT or Google worse for the environment?

Comparing specific AI models is difficult because companies do not always disclose granular energy use per query. However, training large language models requires massive computational power, and a study noted that “training ChatGPT used at least 700,000 liters of water” for cooling. Both rely on energy-intensive data centers, making the underlying infrastructure’s energy source the primary environmental factor rather than the specific model interface.

What is the #1 polluter on planet Earth?

The technology sector is not the #1 polluter; energy production, agriculture, and heavy industry emit far more greenhouse gases. However, the ICT sector produces an estimated 2 to 3.9 percent of worldwide emissions, and its growth rate is accelerating rapidly due to the computational demands of artificial intelligence and blockchain technologies.