Google’s latest Environmental Report, published June 27, 2025, highlights a crucial challenge for the technology giant: sustaining the exponential growth in energy demand driven by artificial intelligence while maintaining a genuine commitment to environmental sustainability.
In 2024, the Mountain View company recorded a 27% increase in electricity consumption across its data centers, a direct consequence of growing demand for computational power to run advanced AI models, such as those underlying large neural networks. However, through massive investments in renewable energy and infrastructure efficiency improvements, Google paradoxically managed to reduce associated carbon emissions by 12%. Specifically, more than 25 new renewable energy projects came online, adding 2.5 GW of clean capacity, while a further 8 GW were contracted for the future. Source
Energy efficiency was significantly improved through innovative technologies such as TPU v5 “Ironwood” chips, which deliver 30 times better energy efficiency compared to Google’s first-generation chips from 2018. Advanced techniques, including machine learning model quantization, accelerated the training of large language models by 39%, reducing the need for computational resources.
But energy consumption is not the only environmental cost of the AI revolution: water usage for cooling data centers has reached concerning levels. In 2023, Google’s data centers used 6.1 billion gallons of freshwater, equivalent to the annual irrigation needs of approximately 40 golf courses. Some facilities, such as the one in Council Bluffs, Iowa, consumed 980 million gallons individually — the equivalent of 6.5 golf courses. In regions such as Northern Virginia, a global data center hub, water consumption increased 65% between 2019 and 2023, raising urgent questions of sustainability and environmental responsibility.
In response, Google has intensified water recycling and replenishment projects, achieving a 64% offset of freshwater used in 2024 (compared to 18% the previous year). These efforts underscore Google’s commitment to mitigating environmental impact, while confirming that the growth of artificial intelligence carries significant and unavoidable environmental costs.
In parallel, Google’s approach to sustainability is profoundly influencing traffic dynamics.
Digital Sustainability and Editorial Selection: How Google Is Changing and What Publishers Must Do
Google has been on a radical path toward environmental sustainability and energy efficiency for several years now, profoundly influencing the way it indexes and organizes the web. In particular, Google’s spider, Googlebot, has become significantly more selective and conservative, reducing re-crawl requests for pages that have remained unchanged for more than a month by 14% compared to 2021. This shift, documented in a white paper presented at the Hot/Storage 2024 conference, has delivered energy savings of approximately 3.2 TWh per year — equivalent to the annual electricity consumption of 280,000 European homes.
At the heart of this new operational philosophy is the Change Rate Estimator, an algorithm that came into effect in Q4 2023. It decides whether a URL genuinely needs to be revisited, deferring by up to 90 days any page with less than a 0.2% probability of change, unless it receives a ping from the sitemap or fresh inbound links. This approach aims to avoid energy and computational waste, which is further reduced by improving page load speed. Internal data presented by Martin Splitt at Chrome Dev Summit 2023 confirms that pages with a Largest Contentful Paint (LCP) below 2.5 seconds receive nearly twice as many crawler visits because they consume less CPU, further reducing energy expenditure.
In parallel, Google has shifted its indexing strategy, moving from an omnivorous approach to an “exclusive catalog.” A study by analyst Vincent Schmalbach, conducted across 1,200 domains, makes this trend clear: in 2025, only 42% of submitted URLs were indexed, versus 90% in 2019. Google today favors content characterized by high semantic originality — evaluated through the new MUVERA* system — domain authority, and complete structured data.
The tightening was most visible in what became known as the “Great De-indexing” between May 27 and 29, 2025, when Search Console recorded 275 million more pages in the Crawled — currently not indexed state. Analysis by Marie Haynes showed that 71% of these pages contained low-originality content, such as paraphrased text, product listings without reviews, or obsolete and poorly read articles. In the same days, Google published version 9.3 of the Quality Rater Guidelines, introducing the “Paraphrased / Low-Originality Content” flag: the connection appears anything but coincidental.
Gary Illyes, a senior analyst at Google, further clarified the concept, describing crawling as “low-hanging fruit” for reducing environmental impact. With fewer useless pages indexed, energy consumption and CO₂ output decrease — aligning environmental and operational goals.
“Crawling is very low-hanging fruit when you’re trying to reduce your environmental impact: fewer useless pages = less CO₂” — Gary Illyes, Search Off The Record, episode “Making crawling more sustainable,” November 11, 2022.
For digital publishers, this transformation represents a genuine revolution that demands a complete revision of editorial and technical strategies. Producing large volumes of generic content is no longer viable: today it is essential to offer unique, high-quality, high-value content. At the same time, it is imperative to optimize the technical performance of websites — above all by reducing load times and improving Core Web Vitals — and to implement regular analysis processes to anticipate potential de-indexing events and adapt quickly to Google’s evolving criteria. Optimizing Core Web Vitals on the server side reduces by 40% the CPU Google spends on rendering, as Abhijit Saha from the Sustainability team has explained.
In a landscape where artificial intelligence and sustainability have become central concerns, digital publishers must act with responsibility and foresight, transforming these challenges into concrete opportunities to stand out and consolidate their position in an increasingly selective digital ecosystem.
Investing in original content that is technically efficient and strictly adheres to E-E-A-T guidelines (Experience, Expertise, Authoritativeness, Trustworthiness) is essential not only for maintaining online visibility, but also for operating responsibly in a digital ecosystem that must contend with increasingly tight physical and energy constraints.
“Indexing everything is no longer sustainable or useful; we aim for an index with a net positive quality” — John Mueller, Office Hours, December 2024
John Mueller reiterated this during the December 2024 Office Hours. The result is a visible compression: only pages that carry new information, hosted on fast and lightweight sites, make it into the index. Uniqueness, performance, and authority signals are no longer recommended — they are non-negotiable.
The future will require publishers to be increasingly aware of the challenges posed by AI and sustainability, transforming them into opportunities to distinguish themselves in a competitive and saturated environment, while keeping the quality and authenticity of their journalistic work at the center.
*MUVERA is a new algorithm developed by Google that revolutionizes search across complex data, such as documents represented by multiple vectors (one per word), simplifying the process without sacrificing precision. In practice, MUVERA condenses these complex representations into a single synthetic vector — called an FDE — that preserves the essence of the original content. This allows the use of much faster search methods, drastically reducing time and computational costs while achieving accuracy comparable to more sophisticated systems. A major breakthrough for anyone working with large quantities of data and content.
Frequently asked questions
