Research Finder
Find by Keyword
Amazon Sets an Emissions Target. How Will AWS Manage Rising Power Demand?
Amazon’s new target gives customers and communities a way to track progress as AWS’s growing AI capacity puts more demand on local grids.
11/19/2026
Key Highlights
- Amazon aims to cut emissions per unit of electricity across its global operations by 65% by 2030, compared with 2019.
- Amazon reported matching 100% of its global electricity use with renewables in 2025, while its reported emissions from purchased electricity rose 34%.
- AWS can reduce the electricity needed to run workloads through changes to cooling, hardware, and server utilization, even as total demand grows.
- Regional reporting would help customers and communities see how AWS expansion affects local power demand and whether cleaner generation is coming online in time.
The News
On November 17, Amazon announced an interim goal to reduce carbon emissions per megawatt-hour of electricity used across its global operations by 65% by 2030, compared with 2019. Amazon says it based the target on the International Energy Agency’s Net Zero by 2050 pathway and will report progress annually. The new target sits within Amazon’s broader commitment to reach net-zero carbon emissions by 2040. Amazon announcement
Analyst Take
Amazon deserves credit for setting a measurable goal to cut emissions from the electricity it uses. This is the kind of commitment we’d like to see from more cloud providers, especially as they expand their data center capacity for AI. As communities push back on data center expansion, cloud providers should be able to explain what they’re doing to use less power for the work they run, where the additional electricity will come from, and who is responsible for following through. Amazon’s target gives us one way to judge progress.
Two figures from Amazon’s 2025 sustainability report explain why this matters. Amazon says it matched 100% of its global electricity use with renewables that year. It also reported that emissions from purchased electricity rose 34%, to 3.74 million metric tons of carbon dioxide equivalent. Amazon attributed the increase to electricity use in data centers and electrification across its delivery network and buildings. That last point matters. Moving delivery vans from fuel to electricity can raise electricity demand while reducing fuel use. A data center expansion adds a different kind of load. The companywide total includes both, so it’s important for AWS to explain its own contribution.
Cleaner electricity and lower power consumption require different work. Amazon can procure cleaner power. AWS can improve how much computing it gets from each megawatt-hour through better cooling, higher server utilization, and more efficient hardware. Enterprise customers can decide which workloads to run and how much capacity to provision. Amazon owns the corporate emissions commitment. AWS is in the best position to show how much electricity its efficiency measures save and how those savings compare with demand from new capacity.
There’s also a difference in what people need to know. An enterprise buyer may want to compare two AWS regions that meet its latency and data residency requirements. A community considering a new data center wants to know how much power the site will draw, what upgrades the utility expects, and who will pay for them. AWS can explain its projected demand and efficiency commitments. The utility can explain how it plans to serve that demand, and the utility and regulators can explain how costs are allocated. Amazon’s commitment to report progress every year gives customers and communities something concrete to track. Regional reporting would make that progress more relevant to people who want to understand what a new data center means for their local grid.
What Was Announced
Amazon will track emissions per megawatt-hour of electricity consumed across its global operations and publish progress in its annual sustainability report. If it meets the target, the electricity it uses in 2030 will be associated with 65% lower emissions per megawatt-hour on average than in 2019. Total electricity emissions could still rise if consumption grows quickly enough. Reporting intensity alongside total electricity use and absolute electricity emissions would make that outcome visible.
Amazon says its carbon-free energy portfolio spans more than 700 projects representing over 40 gigawatts of capacity once completed. Its plans include wind, solar, on-site generation, geothermal, and nuclear power. The timing is important. Amazon and X-energy are targeting more than 5 gigawatts of new U.S. advanced nuclear capacity by 2039. For the 2030 goal, I’d pay closer attention to which projects are expected to be generating electricity by then and which grids they serve.
Amazon’s report describes work already underway to improve data center efficiency, including liquid-to-chip cooling and software changes to improve server utilization. Amazon reported a global power usage effectiveness (PUE) of 1.14 for its data centers in 2025. PUE divides total data center energy use by the energy used by IT equipment. A PUE of 1.14 means that for every 100 units of energy used by IT equipment, another 14 went to cooling and other facility overhead. Understanding the impact of AI growth also requires knowing how much electricity the equipment itself uses and how that changes as capacity comes online.
Looking Ahead
The first annual update should show the 2019 baseline, the latest intensity figure, total electricity use, and absolute electricity emissions on a consistent basis. We’d also like to see AWS report electricity use and efficiency improvements in the markets where it is adding significant capacity. If electricity demand rises despite efficiency gains, AWS should explain how much demand the improvements avoided and what it plans to do about the remaining growth.
AWS already provides customers with monthly emissions estimates by region and service through its Sustainability Console. That gives enterprise teams somewhere to start when comparing deployments. For communities, a more useful disclosure would pair projected data center demand with the expected timing of grid upgrades and new generation. Where AWS makes a local commitment, the public should be able to see who owns it, when it is due, and whether it was met.
By 2030, we’d judge Amazon’s progress by whether cleaner electricity and more efficient data centers are bringing absolute electricity emissions down as AWS grows. The credibility of that result will depend on what Amazon and AWS can show along the way, especially in the communities where the new demand is landing. We’d like to see other cloud providers set measurable electricity emissions goals and report their progress alongside total electricity use. That would make it easier to see where cleaner power and efficiency improvements are making a difference as AI demand grows.
Stephanie Walter | Practice Leader - AI Stack
Stephanie Walter is a results-driven technology executive and analyst in residence with over 20 years leading innovation in Cloud, SaaS, Middleware, Data, and AI. She has guided product life cycles from concept to go-to-market in both senior roles at IBM and fractional executive capacities, blending engineering expertise with business strategy and market insights. From software engineering and architecture to executive product management, Stephanie has driven large-scale transformations, developed technical talent, and solved complex challenges across startup, growth-stage, and enterprise environments.



















