Google’s carbon emissions rose 25% and Amazon’s 16% year-over-year, driven by AI infrastructure expansion, according to TechCrunch. Scope 3 emissions — particularly from data center construction, GPU procurement, and semiconductor manufacturing — now dominate the growth, doubling since Google’s 2019 baseline. Amazon added over 1.2 GW of data center capacity in Q4 2025 alone. Both companies face steep technical and financial hurdles to meet net-zero goals, including scaling low-carbon steel/cement and purchasing millions of tons of carbon removal credits.
According to techcrunch.com, Google’s total carbon emissions rose 25% year-over-year, while Amazon’s increased by 16%, driven largely by surging AI infrastructure demands — a development that jeopardizes both companies’ net-zero pledges.
AI-driven emissions surge overwhelms renewable energy gains
Both Google and Amazon released their 2026 sustainability reports this week, revealing stark setbacks in climate progress. Though years of renewable energy procurement had previously stabilized Scope 1 and 2 emissions — those tied directly to owned operations and purchased electricity — AI’s exponential growth has disrupted that trajectory. Google now operates data centers powered partly by newly commissioned natural gas plants, a strategic pivot away from renewables to meet real-time AI compute loads. Amazon reported adding more data center capacity globally in 2025 than any other company, including over 1.2 gigawatt (GW) in Q4 alone — a scale-up directly linked to AI service demand.
The shift reflects a broader industry tension: while tech firms can offset office and legacy data center emissions with wind and solar purchases, AI workloads require continuous, high-density power that batteries and intermittent renewables struggle to supply reliably at current scale.
Scope 3 emissions dominate the upward trend
The bulk of the emissions increase falls under Scope 3 — indirect emissions from value chain activities outside direct operational control. Google’s Scope 3 emissions grew by 2.1 million metric tons last year, doubling since its 2019 baseline. Amazon’s Scope 3 rise stems primarily from capital goods (including data centers and warehouses) and fuel and energy inputs — categories where construction materials like steel and cement carry heavy embedded carbon footprints. Semiconductor manufacturing also contributes significantly: advanced GPU and memory chip fabrication consumes vast energy, especially in Asia, where grids remain fossil-fuel-dominant. Further compounding impact are potent fluorinated greenhouse gases used in chip fabs — some up to thousands of times more warming-potent than CO₂ per unit mass.
Neither company explicitly names AI as the sole driver in public statements, but both acknowledge sharply rising energy use correlated precisely with AI adoption timelines. Both also emphasize AI’s theoretical environmental benefits — a rhetorical stance analysts describe as “protesting too much.”
Net-zero targets face mounting technical and financial hurdles
Achieving net-zero commitments now requires far more than green energy procurement. To compensate for AI’s footprint, Google and Amazon must accelerate investment in low-carbon steel and cement production — technologies still unproven at industrial scale — and purchase millions of tons of verified carbon removal credits. Tim De Chant, Senior Climate Reporter at TechCrunch and author of the source article, notes that while solutions exist, “their embrace of AI hasn’t made it any easier.” De Chant holds a PhD in environmental science, policy, and management from the University of California, Berkeley, and was awarded a Knight Science Journalism Fellowship at MIT in 2018.
Industry observers stress that these challenges are shared across Big Tech: Microsoft, Meta, and others report similar Scope 3 spikes tied to AI hardware procurement and user-device energy consumption. Unlike past digital transitions, AI’s physical infrastructure intensity makes decarbonization fundamentally more complex — requiring coordination across semiconductor supply chains, construction materials producers, and grid operators.
Source: TechCrunch
Compiled from international media by the SCI.AI editorial team.
