Google’s carbon emissions increase by 48% in five years due to energy consumption required by AI

Google has reported that its Carbon emissions have increased by 48 percent in the last five years due to the energy consumption required by the use of Artificial Intelligence (AI) in its products, reaching emissions of 14.3 million tons of carbon dioxide (CO2) in 2023.

With the rise of AI, big tech companies have developed their own models and technologies, as is the case of Google with Gemini, in order to offer tools and functions in their services that provide new ways of working, creating content or communicating.

However, using this technology to carry out different tasks often involves a very high energy consumptionwhich increases carbon emissions in data centers from where AI is operated. As a result, it results in a setback in the fight against climate change.

In this framework, Google has announced that, during the year 2023, its Total carbon dioxide emissions have risen to 14.3 million tonswhich represents an increase of 48 percent compared to emissions recorded in 2019.

This was announced by the technology company in the latest results of its annual Environmental Report, in which it offered an analysis of its efforts to take advantage of technology, especially AI, to drive positive environmental change and operate sustainably, in addition to pointing out its intention to achieve net zero emissions across all its operations by 2030.

However, the US company has clarified that as they continue to integrate AI into their products, the Reducing emissions “may be a challenge”. In fact, their carbon emissions during 2023 have increased by 13 percent compared to the previous year.

This represents a CO2 emissions will grow by almost 50 percent over a 5-year periodsomething Google has attributed to increases in energy consumption at its data centers and emissions from its product chain.

Specifically, as he explained, this increase in consumption is due to the energy demands required by their AI technologies. As he pointed out, given that There is increasing demand for AI tools at a general levelalso increases the computational intensity of AI and therefore the associated emissions.

With all this, as detailed in the report, during the past year the Total electricity consumption of its data centers grew by 17 percent And as Google’s infrastructure “continues to drive the digital transition,” the company expects this trend will continue to increase in the future.

MINIMIZE YOUR ENVIRONMENTAL IMPACT

In this context, Google has also stated that Scaling up AI and using it to accelerate action on climate change is “as crucial as addressing the environmental impact associated with it.” The technology company has therefore indicated that it is carrying out various initiatives to reduce its environmental footprint.

For example, as he stressed, they build more energy-efficient IT infrastructures, supported by responsible water use practices and a commitment to minimizing resources. In doing so, they achieve that a Google-owned data center and operated by Google is approximately 1.8 times more energy efficient than an enterprise data center.

Thus, in 2023 the average effectiveness of the annual energy use of its data centers was 1.10 instead of the industry average of 1.58. That is, its Data centers use approximately 5.8 times less overall energy for each unit of IT equipment power.

Using AI to boost climate action

Google has also stated in its report that the AI has the potential to help mitigate between 5 and 10 percent of global greenhouse gas emissions by 2030To this end, he said they are promoting AI in key areas such as information organization, improved prediction and better optimization.

An example of this initiative is the fuel efficient routes functionorganizes information by analyzing traffic, terrain and the engine of the vehicle in question to suggest the most efficient route. With these types of initiatives, Google has helped reduce more than 2.9 million metric tons of emissions from the end of 2021 to the end of 2023. This is equivalent to taking approximately 650,000 cars off the road for one year.