Google’s 2026 environmental report reveals that the company’s electricity use rose 37% in 2025 compared to the previous year — the fastest growth rate in the company’s history. The main driver is the ongoing expansion of data centers and infrastructure to support its continuously growing AI services.
Even as Google continues to invest heavily in clean energy, this report shows that AI’s rapid growth is creating new challenges for the tech industry’s sustainability goals.
AI Needs More “Electricity” Than Most People Realize
Every time a user sends a request to AI — whether generating text, analyzing data, or creating an image — behind the scenes tens of thousands of high-performance servers inside data centers are processing it.
The more people use AI,
- the more data centers need to be built
- the more GPUs and servers need to be installed at massive scale
- the more high-efficiency cooling systems are required
- the more electricity and water are needed to keep systems stable around the clock
“100% Renewable Energy” Doesn’t Mean Zero Carbon Emissions
Many people may have seen Google announce it runs on 100% renewable energy. In reality, that figure is based on Renewable Energy Certificates (RECs) — an accounting mechanism that confirms a company has supported clean energy production equal to the amount of electricity it uses.
However, the actual electricity flowing into data centers may still come from grids that rely on fossil fuels, and these certificates cannot offset the carbon emissions generated across the supply chain — such as chip manufacturing plants, server production, or data center construction materials.
The Real Challenge Lies in Scope 3
The report states that roughly 80% of Google’s total carbon footprint comes from Scope 3 emissions — indirect greenhouse gas emissions from the supply chain.
Examples include:
- AI chip manufacturing
- Server manufacturing
- Data center construction
- Equipment transportation
- Steel and cement production
AI Is Getting More Efficient, Yet Energy Use Keeps Rising
Google states that its newer AI models are significantly more efficient — for example, Gemini uses less energy per query compared to earlier versions.
However, improved efficiency hasn’t reduced overall electricity use. As AI becomes cheaper to run, more people use it, and in more varied ways, driving overall infrastructure demand up in turn.
This phenomenon aligns with an economic concept known as the Jevons Paradox, which explains that as a technology becomes more efficient and cheaper, demand for it tends to rise enough that total resource use ends up higher than before.
Google Keeps Investing in Clean Energy, But Can’t Keep Pace With AI’s Growth
Even as electricity demand rises sharply, Google continues to invest in clean energy.
In 2025, the company signed new clean energy supply agreements totaling more than 12 gigawatts (GW) — a company record — spanning solar, hydro, and nuclear power, along with future collaboration on fusion energy projects.
However, Google acknowledges in the report that AI infrastructure expansion is still outpacing the rate at which the power grid is decarbonizing, making its 2030 Net Zero target an increasingly difficult goal to reach.
What Organizations Should Learn From This
This story isn’t just about the environment — it also points to another truth:
AI isn’t just software. It’s massive infrastructure that depends on energy, hardware, and global cloud systems.
For organizations building an AI strategy, looking only at a model’s capabilities may not be enough. Other factors are worth considering as well, such as:
- Infrastructure efficiency
- Appropriate use of cloud resources
- Energy costs
- Supply chain sustainability
- ESG and Net Zero targets
AI is one of the most powerful technologies of the digital age, but using AI sustainably doesn’t just mean choosing the smartest model.
Just as important is designing the right system architecture, using resources efficiently, and applying AI where it genuinely creates business value.
At Dragons Move, we believe sustainable digital transformation requires a balance between innovation, efficiency, and environmental responsibility — so organizations can grow while using technology in a way that’s both high-quality and sustainable.
Key Takeaways
- Google’s electricity use rose 37% in 2025 due to AI infrastructure expansion — the largest increase in the company’s history
- Google’s data centers used approximately 42 million MWh of electricity per year, close to New Zealand’s entire national electricity consumption
- Even though the company matches 100% of its electricity use with Renewable Energy Certificates, supply chain (Scope 3) carbon emissions continue to rise and account for roughly 80% of its total carbon footprint
- Making AI more efficient hasn’t reduced overall energy use, because demand is growing even faster
- Organizations should build AI strategies that weigh efficiency, cost, and sustainability together — not just a model’s raw capabilities
Source
This article was written and analyzed based on a TechTimes report, “Google AI Electricity Up 37%: Renewable Certificates Cannot Cover the Supply Chain Carbon.”