The three largest US tech companies emitted 119 million metric tonnes of carbon dioxide equivalent in the past year, a nearly 20% increase from the previous year, driven almost entirely by the infrastructure buildout required to power artificial intelligence systems. This reversal marks a dramatic departure from years of climate progress, as Microsoft, Amazon, and Google race to construct the physical backbone needed to train and operate AI models like ChatGPT and other large language models (LLMs) that have become central to their business strategies.

How Did Tech Giants' Emissions Spike So Dramatically?

The numbers tell a striking story. Microsoft's carbon emissions rose 25% in the financial year ending March 2026, reaching 20 million metric tonnes of carbon dioxide equivalent, with the company explicitly attributing the increase to "expansion of our datacenter infrastructure." Google reported an 18% increase, while Amazon saw a 16% overall increase and a 20% jump in supply chain emissions tied to datacenter construction.

What makes this particularly notable is the timing. Prior to this year, Microsoft's emissions had remained relatively flat at 16 million metric tonnes in both 2023 and 2024. The sudden acceleration reflects the scale of investment pouring into AI infrastructure. The world's biggest tech companies are on track to spend $765 billion this year alone, mostly on building AI datacenters in locations ranging from Norway to North Tyneside in the United Kingdom.

The physical demands of AI are staggering. The Uptime Institute, which rates and inspects datacenters, estimates that big datacenter projects announced last year would consume 1.3% of the world's electricity usage, representing a near-doubling of current datacenter demand. The majority of that new power demand will come from US projects.

Why Are Datacenters So Energy-Intensive?

Training and running AI models requires enormous amounts of computing power. Each query processed by an AI system, each model trained from scratch, and each update to existing systems demands electricity. Datacenters house thousands of specialized processors working around the clock, consuming power continuously whether the facility is operating at full capacity or not. The infrastructure itself, from cooling systems to power distribution, adds additional energy demands on top of the raw computing needs.

The scale of planned expansion is breathtaking. JLL, a US property consultancy, expects approximately 1,200 datacenters to be built globally between now and 2030, with demand overwhelmingly driven by AI.

What Do These Emissions Mean in Real Terms?

To put the numbers in perspective, the combined annual emissions of Microsoft, Amazon, and Google now equal roughly one-third of France's total carbon footprint. The previous year's emissions were roughly equivalent to the 2024 emissions of Czechia, an entire nation. This comparison underscores how concentrated the environmental impact of AI infrastructure has become in just a handful of companies.

Despite these increases, all three companies maintain ambitious net-zero targets. Google and Microsoft have committed to achieving net-zero emissions by 2030, while Amazon has set a 2040 deadline. However, experts question whether these goals remain achievable given current trajectories.

What Are the Hidden Complications Behind These Emissions?

One often-overlooked aspect of the emissions problem involves carbon credits. Shaolei Ren, a professor of electrical engineering at the University of California, Riverside, noted a troubling trend in the companies' sustainability reports.

"While companies are actively investing in or purchasing carbon credits, the figure suggests a possible lack of credit supply in the carbon market to meet the technology companies' needs. Everyone is talking about the lack of physical goods and infrastructure like power, but there may also be a lack of virtual goods, carbon credits," Ren explained.

Shaolei Ren, Professor of Electrical Engineering at University of California, Riverside

This creates a paradox: even if companies want to offset their emissions through carbon credits, the market may not have enough credits available to match the scale of their emissions growth.

There's also a broader accountability problem. When companies migrate their operations to cloud services provided by Microsoft, Amazon, and Google, they outsource their digital carbon footprint to these cloud giants. This means the emissions appear on the tech companies' balance sheets, but the actual responsibility for those emissions is diffused across hundreds of client organizations.

How Are Experts Responding to These Claims?

Cecilia Rikap, an economics professor at University College London, expressed skepticism about the tech giants' sustainability messaging.

"Claims by Microsoft, Amazon and Google about their clouds being ecologically friendly and sustainable are a marketing strategy. Governments should remember these expanding carbon footprints when the very same companies offer addressing the ecological crisis with AI solutions," Rikap stated.

Cecilia Rikap, Economics Professor at University College London

This critique highlights a fundamental tension: the same companies promoting AI as a solution to climate change are simultaneously driving up their own emissions at unprecedented rates.

Steps to Understand the AI-Climate Tradeoff

• Track Emissions Transparency: Monitor the annual sustainability reports released by major tech companies to understand how their carbon footprints are changing year-over-year and what specific activities are driving increases.
• Evaluate Offset Claims: When companies claim to be carbon-neutral or net-zero, examine whether they're relying on carbon credits and whether those credits represent genuine emissions reductions or simply financial transactions.
• Consider Outsourced Emissions: When your organization uses cloud services, recognize that your digital carbon footprint is being counted in the cloud provider's emissions, not your own, which can obscure the true environmental cost of digital operations.
• Assess Infrastructure Expansion Plans: Look at announcements about new datacenter construction in your region or globally, as these projects represent the physical manifestation of AI's growing energy demands.

The fundamental challenge facing the tech industry is that the infrastructure required to power AI systems is expanding far faster than the renewable energy sources needed to power them sustainably. Until that balance shifts, the emissions from AI datacenters will likely continue climbing, even as companies maintain their net-zero commitments.