Google and Amazon emissions surge 25% and 16% as AI demands grow

The environmental toll of artificial intelligence is coming into sharp focus. Google and Amazon both released sustainability reports this week showing significant increases in carbon emissions — 25% for Google and 16% for Amazon compared to the previous year — as both companies race to expand AI capabilities.

Neither company explicitly attributes the increases to AI in their reports, but the evidence is clear when examining their energy consumption patterns and infrastructure investments. Both acknowledge substantial growth in energy use coinciding with rising AI deployment, and both devote considerable space to discussing how AI might eventually benefit the environment.

Why it matters

These figures represent a critical inflection point for Big Tech's climate commitments. Companies that once appeared on track to meet net-zero targets now face a fundamental tension between AI ambitions and environmental pledges. The problem extends beyond energy consumption to include the physical infrastructure and manufacturing processes required to support AI at scale — challenges that lack ready solutions.

The Scope 3 problem

The bulk of the emissions growth comes from Scope 3 categories — indirect emissions from purchased goods, services, and capital investments that companies don't directly control. For tech giants, this includes GPU purchases, semiconductor manufacturing, and data center construction.

Google's Scope 3 emissions increased by 2.1 million metric tons last year and have doubled since 2019, the company's baseline year. The primary driver appears to be capital goods, particularly data center infrastructure, since Google's consumer hardware products are relatively small and low-power.

Amazon's situation is similar but more pronounced. The company added more data center capacity globally in 2025 than any other company, including over 1.2 gigawatts in the fourth quarter alone. This aggressive expansion in capital goods and the associated fuel and energy requirements pushed Scope 3 emissions higher than Google's increase.

The renewable energy ceiling

For years, tech companies could offset their carbon footprints by purchasing renewable energy to power offices and modest data centers. AI has rendered that approach insufficient. While renewable energy plus battery storage remains technically viable for powering AI data centers, companies including Google have begun investing in natural gas plants to meet immediate power demands.

The construction materials present an even thornier challenge. Steel and cement production for data centers generates substantial emissions, and while startups are developing low-carbon alternatives, these technologies aren't yet ready to scale to meet Big Tech's needs.

Semiconductor manufacturing adds another layer of complexity. GPU and memory chip production consumes enormous amounts of energy, and many advanced fabrication facilities are located in Asian countries where electrical grids remain heavily dependent on fossil fuels. The manufacturing process also uses chemicals that are potent greenhouse gases, capable of warming the atmosphere thousands of times more effectively than equivalent amounts of CO2.

A difficult path forward

The challenges aren't insurmountable, but meeting net-zero commitments will require substantial investment and operational changes. Companies will need to dramatically increase renewable energy purchases, fund development of advanced low-carbon steel and cement manufacturing, and purchase millions of tons of carbon removal credits.

The sustainability reports make clear that AI's rapid growth has fundamentally altered the trajectory toward carbon neutrality, transforming what appeared to be achievable goals into significantly more difficult targets.

These details were first reported by TechCrunch.