AI growth drags down big tech's climate goals.
The race to build artificial intelligence infrastructure is making it harder for the world's biggest technology companies to keep their climate commitments, with new sustainability reports showing greenhouse gas emissions climbed sharply in 2025 as data center expansion accelerated.
Amazon and Google both reported higher emissions over the past year, largely tied to the rapid construction of AI infrastructure and the growing energy demand of advanced AI systems.
Amazon said its greenhouse gas emissions increased 16% year over year to about 80.8 million metric tons of carbon dioxide equivalent. According to the company, much of the increase came from expanding data centers and emissions across its supply chain, as well as fuel used for deliveries.
Google reported an 18% rise in what it describes as its "ambition-based" emissions, while emissions directly tied to its operations, excluding purchased electricity, climbed 20%. The company attributed part of the increase to its growing portfolio of AI data centers.
AI is making climate goals harder to reach
The reports highlight the growing challenge of balancing explosive AI growth with long-term carbon reduction targets across the technology industry.
Amazon is aiming to reach net-zero emissions by 2040, while Google is targeting 2030.
Google's electricity consumption jumped 37% in 2025, although the company said increased use of clean electricity helped slightly reduce emissions associated with purchased power. Amazon, meanwhile, reported a 34% increase in emissions linked to purchased electricity.
Google acknowledged the pace of AI deployment is creating new obstacles, stating in its sustainability report: "Our AI infrastructure buildout is currently accelerating faster than the grid is decarbonizing."
Amazon also cautioned that progress will not follow a straight path. "While the speed and scale of AI adoption is unique — and the change is happening faster and more broadly than anything else we’ve encountered in our lifetimes — the need to stay stubborn on our vision and flexible on the details is familiar territory," Chief Sustainability Officer Kara Hurst said in the company's report.
A wider industry challenge
The trend extends beyond Amazon and Google.
Previous sustainability reports showed Microsoft's emissions increased by 23%, while Meta reported a 64% jump, as both companies continued to invest heavily in AI infrastructure, Bloomberg reported.
Building modern AI facilities requires vast amounts of steel, concrete, computer hardware, and electricity, resulting in emissions well beyond those from operating the centers themselves. Google reported a 25% increase in supply-chain emissions, largely driven by hardware manufacturing and data center construction.
Environmental groups argue the industry's emissions trajectory is moving in the wrong direction. "We're essentially in a climate crisis and we should not be having emissions growth at all, arguably, and yet the data centers are going in the opposite direction," said Sasha Luccioni, co-founder and chief scientific officer of Sustainable AI Group, according to Bloomberg.
Why it matters
The latest reports underscore that AI's environmental costs are becoming as much a business issue as a climate one.
Technology companies are investing billions in new AI infrastructure while simultaneously spending heavily on renewable energy and other decarbonization efforts. However, much of the electricity powering new data centers still comes from fossil fuels, particularly natural gas, meaning emissions may continue rising before cleaner energy supplies can catch up.
For investors, customers, and regulators, sustainability targets are increasingly being tested against the realities of AI expansion rather than traditional cloud computing growth.
AI's next bottleneck could be energy
The AI race has centered on securing enough processors and building larger data centers for years. The latest sustainability reports suggest another constraint is emerging: access to clean electricity. Companies can build computing capacity faster than power grids can transition away from fossil fuels, creating a gap between AI ambitions and climate commitments.
The challenge is unlikely to disappear quickly. Even as major tech companies continue investing in renewable energy, the construction of AI infrastructure itself carries a substantial carbon footprint. That means future progress will depend not only on better AI hardware, but also on how quickly energy systems and supply chains become cleaner.
Also read: Meta’s Hyperion data center is using 72-ton construction robots to help build solar infrastructure as AI campuses demand more electricity and speed.