Google’s greenhouse gas emissions have risen 48 percent over the past five years, the company disclosed in its most recent environmental reporting, as the energy demands of artificial intelligence infrastructure continue to outpace efficiency gains. The increase, measured against a 2019 baseline, marks one of the starkest admissions yet from a major tech firm about the environmental cost of the AI boom.
The disclosure, contained in Google’s annual environmental report, attributes the surge primarily to electricity consumption from data centers built to train and run AI models, along with associated emissions from its supply chain. Google has pledged to reach net-zero emissions across its operations and value chain by 2030, but the company itself has acknowledged that the target has become significantly harder to hit as AI workloads scale.
What’s Driving the Increase
Data centers that power AI systems like Google’s Gemini models require enormous amounts of electricity, not only to run computer chips but also to cool the facilities housing them. As demand for AI tools has grown among consumers and businesses, Google has expanded its data center footprint globally, adding new campuses in the United States, Europe, and Asia.
Google’s report notes that total electricity consumption across its data centers has climbed sharply in recent years, and that a meaningful share of that power still comes from grids reliant on fossil fuels, particularly in regions where renewable infrastructure has not kept pace with new construction. The company has said it is working to sign long-term clean energy contracts and invest in nuclear and geothermal power to offset the growth, but those projects take years to come online.
Scope 3 Emissions Remain the Biggest Challenge
A large portion of the emissions increase falls under what’s known as Scope 3 emissions — indirect emissions from a company’s supply chain, including the manufacturing of servers, chips, and other hardware. Google has said these emissions are especially difficult to control because they depend on suppliers’ own energy sources and manufacturing processes, many of which are based in countries with carbon-intensive electricity grids.
Environmental researchers have pointed out that the semiconductor industry, which produces the specialized chips used to train large AI models, is itself energy- and water-intensive. As AI companies race to build ever-larger models, demand for these chips has surged, indirectly driving up emissions tied to their production.
Industry-Wide Pattern
Google is not alone in facing this tension. Other major technology companies operating large AI infrastructure have reported similar increases in their own carbon footprints in recent years, even as they maintain public commitments to sustainability goals. Analysts say the pattern reflects a broader industry dilemma: the same AI systems being marketed as tools to improve efficiency and solve complex problems, including climate-related ones, are themselves significant contributors to energy demand and emissions growth.
The tension has also drawn scrutiny from within the AI industry itself. Concerns about the pace and oversight of AI development have surfaced in other contexts recently, including a researcher’s resignation from Anthropic over AI safety warnings, underscoring broader unease about how quickly the technology is scaling relative to safeguards, whether around safety or sustainability.
Company Response
In its report, Google reiterated its long-term commitment to sustainability, pointing to investments in renewable energy purchase agreements, advances in chip efficiency, and research into cooling technologies that reduce water and power use at data centers. The company has also touted AI’s potential to help address climate challenges, citing internal tools designed to optimize energy grids and reduce waste in other industries.
Google has said it remains “committed” to its 2030 net-zero goal, while acknowledging that “the path to get there has become less clear” given the scale of AI-related infrastructure growth.
Still, the company has not detailed a specific plan to reverse the current emissions trajectory, and some sustainability experts argue that voluntary corporate targets may need to be paired with stronger regulatory frameworks to meaningfully curb emissions growth tied to AI expansion.
What Comes Next
The disclosure is likely to intensify scrutiny of tech companies’ climate commitments as AI adoption continues to expand across consumer and enterprise markets. Regulators in the European Union and several U.S. states have signaled interest in requiring more detailed and standardized emissions reporting from large data center operators, which could increase pressure on companies like Google to show measurable progress rather than aspirational targets.
For now, the gap between Google’s stated climate goals and its rising emissions serves as a case study in the broader challenge facing the tech industry: balancing the rapid commercial rollout of AI technology with the environmental costs of the infrastructure required to sustain it.
