Can Google’s AI Ambitions Match its Sustainability Promises?

As AI reshapes modern technology, its environmental footprint is becoming increasingly difficult to ignore.

AI is now embedded in everything from healthcare to daily productivity tools, promising major leaps in efficiency and capability.

But behind this surge in innovation lies a growing ecological cost—one that casts doubt on the tech industry’s sustainability credentials.

At the centre of this debate is Google, a company both praised for its climate leadership and challenged for the impact of its data-hungry AI infrastructure.

Its report, 'The AI Opportunity for Europe’s Climate Goals – a Policy Roadmap', outlines the promise of AI in advancing sustainability, even as its operations raise questions about whether such technology can ever truly be green.

It offers actions surrounding AI that European leaders can utilise to enhance their sustainability efforts. The report adds to the joint BCG-Google report on Accelerating Climate Action with AI (2023) and Google's EU AI Opportunity Agenda (2024) to split action into three pillars – deploy, enable and guide. 

This report also comes after the release of the EU AI Act in August 2024.

According to the EU Artificial Intelligence Act website, the EU AI Act: “is a European regulation on artificial intelligence (AI) – the first comprehensive regulation on AI by a major regulator anywhere.

“The Act assigns applications of AI to three risk categories. First, applications and systems that create an unacceptable risk, such as government-run social scoring of the type used in China, are banned.

“Second, high-risk applications, such as a CV-scanning tool that ranks job applicants, are subject to specific legal requirements. Lastly, applications not explicitly banned or listed as high-risk are largely left unregulated.”

This act also established new transparency requirements on the environmental impact of AI models, helping to ensure sustainability remains a top priority for organisations. 

How can decision-makers embrace AI to enhance climate action?

Google explains how EU policymakers must ensure they have a secure framework to craft AI solutions that enhance climate action. 

This can be achieved by improving access to technology infrastructure, expanding access to high-quality climate data and developing skills and raising awareness.

The EU can accelerate the digitalisation of climate data by expanding access to funding for data collection technologies in high-emissions industries. Not only do targeted measures such as voucher schemes help small and medium-sized enterprises access high-speed broadband, but improving electricity grid infrastructure will help facilitate the development of energy-intensive facilities, such as data centres.

Expanding access to high-quality climate data will remove a significant barrier companies face in the development of AI tools for climate action – siloed, incomplete or non-interoperable datasets.

This can be solved by promoting interoperability across systems and devices in key industries and ensuring policymakers use the European Data Governance Board to encourage the release of public datasets.

Skilled individuals are vital in ensuring AI technologies, infrastructure and data do not fall short of enhancing climate action. 

The EU can offer specialised training and programmes to enhance public sector expertise and upskilling efforts across the private sector should integrate AI training with sustainability knowledge. 

By combining these methods, the EU can ensure organisations and individuals have the expertise, awareness and ability to craft effective AI solutions in the movement to a greener future.

Energy-intensive innovation

Google has long positioned itself at the forefront of environmental technology.

It achieved carbon neutrality in 2007 and has matched its entire electricity use with renewable energy since 2017.

The goal now is to operate fully on carbon-free energy by 2030. Yet the expansion of AI threatens to compromise this commitment.

The launch of Gemini, Google’s flagship generative AI model, marks a turning point.

This system underpins many core Google services, including Gmail, Google Docs and even Google Earth.

According to Kate Brandt, Google’s Chief Sustainability Officer: “Google Earth has democratised geospatial information for a wide range of users and use cases. It renders a 3D representation of Earth, allowing people to explore our planet from endless vantage points.”

She adds that with Gemini capabilities, businesses can now analyse land data more efficiently when evaluating sites for renewable energy.

Despite these benefits, Gemini is built on vast computing power.

These workloads run across Google’s cloud infrastructure, increasing energy demand in data centres already under strain.

Unlike traditional web services, AI requires extensive data training and constant updating. The environmental price is high.

Between 2019 and 2023, emissions from Google’s data centres rose by 48%.

The company attributes this to the growth of cloud services, but industry experts suggest the surge is tightly linked to the rapid scaling of AI models. This growth puts Google in a difficult position—championing sustainability while contributing to the very emissions it claims to reduce.

As Fieke Jansen from the University of Amsterdam’s DATACTIVE project puts it: “Google’s AI ambitions are incompatible with its climate goals. You cannot continue to grow your emissions and claim to be on a path to sustainability.”

Lack of transparency in impact

Part of the challenge is the lack of clarity around the specific environmental cost of AI.

Google releases annual sustainability reports, but these do not detail the energy or water usage tied to individual services, especially those powered by AI.

This lack of disaggregated reporting means the environmental impact of Gemini and similar models remains obscured.

In February 2024, Google embedded Gemini across Gmail, Google Docs and Android smartphones, broadening the scope of AI but also potentially inflating its hidden emissions.

Fieke adds: “Without disaggregated reporting, there’s no way to verify the impact of AI specifically. That makes accountability nearly impossible.”

This issue is industry-wide.

Other tech firms also fail to separate AI from general operations in environmental disclosures, making it difficult for regulators, researchers and the public to understand AI’s real-world consequences.

Water use is another area of concern.

Google’s data centres consumed 5.6 billion gallons of water in 2021, much of it used in cooling systems to prevent overheating.

As AI workloads expand, so does water demand.

In areas such as The Dalles, Oregon, where Google operates a major facility, local communities have expressed concern about the stress on already scarce water supplies during drought conditions.

This figure is likely to have increased in the years since as AI models and infrastructure are scaled up. 

Sustainable AI

In response to mounting scrutiny, there have been some positive signals from Google and others in the industry. 

In 2023, Google DeepMind launched efforts to improve model efficiency by refining architectures, reducing redundancy in training and using smarter scheduling to minimise energy draw. 

These are steps towards what researchers are calling “green AI” – systems designed to be more environmentally conscious from inception.

However, such initiatives are still in their infancy compared to the rapid pace of AI deployment.

Towards a sustainable AI future

There are some efforts to align AI development with sustainability goals.

Google DeepMind has started optimising AI models for efficiency, reducing redundancy, streamlining training processes and using scheduling strategies to cut energy consumption.

These steps fall under the umbrella of “green AI,” which advocates for more environmentally conscious model design.

Yet progress remains slow compared to the pace at which AI is being deployed.

While the technology promises solutions to global sustainability challenges – from optimising energy grids to enhancing agricultural outputs – it simultaneously creates new environmental burdens.

“There is no doubt that AI has enormous potential to help solve global sustainability challenges,” says Fieke. “But the way it’s currently being scaled raises serious questions about whether the solution is becoming part of the problem.”

This conflict is not limited to Google.

Microsoft, Meta and Amazon are also pushing generative AI across their platforms while publishing high-profile climate targets.

As digital infrastructures expand, the gap between ambition and action grows.

Regulatory pressure may soon force change.

Governments are introducing legislation that demands clearer environmental reporting for digital technologies.

These changes could bring accountability to the sector, compelling firms to demonstrate tangible progress, not just intentions.

Google, like its peers, must now reconcile its dual role as a leader in AI and a self-declared champion of sustainability.

The key lies in greater transparency, smarter design and a commitment to reducing real emissions—not just offsetting them.

As Fieke warns: “Tech companies can’t rely on offsetting or distant targets. They need to prove their models are sustainable now.”

Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE. 

Discover all our upcoming events and secure your tickets today. 

AI Magazine is a BizClik brand