AI’s Insatiable Energy Demands Jeopardize Big Tech’s Climate Goals

[Social Media]

 

The artificial intelligence (AI) boom has propelled Big Tech share prices to unprecedented heights, yet it comes at the expense of the sector’s environmental aspirations. Companies like Google and Microsoft are grappling with the challenge of aligning their soaring energy needs with their sustainability commitments.

 

Google admitted on Tuesday that the rapid advancement of AI technology is jeopardizing its environmental targets. The company disclosed that data centers, which are crucial for AI infrastructure, have contributed to a 48 percent increase in its greenhouse gas emissions since 2019. Google noted "significant uncertainty" around achieving its goal of net zero emissions by 2030, citing the unpredictable future environmental impact of AI. This goal aims to balance CO2 emissions produced by the company with an equivalent amount of CO2 removal or elimination.

 

Microsoft, a major financial supporter of ChatGPT developer OpenAI, similarly acknowledged that its ambitious net zero goal for 2030 might be unattainable due to its AI strategy. This raises the critical question: Can the tech industry reduce AI’s environmental footprint, or will it continue to prioritize AI supremacy despite the environmental costs?

 

AI's threat to tech’s green goals primarily lies in the energy-intensive nature of data centers. These facilities are essential for training and operating AI models, such as Google’s Gemini or OpenAI’s GPT-4. They house sophisticated servers that process vast amounts of data, consuming significant amounts of electricity, which leads to substantial CO2 emissions depending on the energy source. Additionally, the manufacturing and transportation of this equipment generate embedded CO2.

 

According to the International Energy Agency, the total electricity consumption of data centers could double from 2022 levels to 1,000 terawatt-hours (TWh) by 2026, matching Japan's energy demand. Research firm SemiAnalysis estimates that AI could drive data centers to consume 4.5 percent of global energy generation by 2030. The water usage is also notable, with one study projecting that AI could account for up to 6.6 billion cubic meters of water use by 2027—nearly two-thirds of England’s annual consumption.

 

A recent UK government-backed report on AI safety highlighted the carbon intensity of the energy source used by tech firms as a critical factor in assessing the environmental cost of AI. The report also mentioned that a significant portion of AI model training still relies on fossil fuel-powered energy. In their bid to meet environmental goals, tech firms are aggressively securing renewable energy contracts. For instance, Amazon is the world’s largest corporate purchaser of renewable energy. However, some experts argue that this trend forces other energy consumers towards fossil fuels due to insufficient clean energy availability.

 

"Energy consumption is not just growing, but Google is also struggling to meet this increased demand from sustainable energy sources," says Alex de Vries, the founder of Digiconomist, a website monitoring the environmental impact of new technologies. The global push to triple renewable energy resources by the end of the decade aims to reduce fossil fuel consumption in line with climate targets. However, the International Energy Agency warns that even though global renewable energy capacity grew at the fastest pace in the past 20 years in 2023, the world might only double its renewable energy capacity by 2030 under current plans. The burgeoning energy demand from AI data centers could further undermine these efforts.

 

Tech companies may need to invest heavily in building new renewable energy projects to meet their escalating power needs. Onshore renewable energy projects like wind and solar farms can be developed relatively quickly, often in less than six months. However, slow planning processes and a global bottleneck in connecting new projects to the power grid could add years to the timeline. Offshore wind farms and hydro power projects face similar hurdles, with construction times ranging from two to five years.

 

Concerns are mounting over whether renewable energy can keep pace with AI’s expansion. Major tech firms have already secured a third of U.S. nuclear power plants to supply low-carbon electricity to their data centers. Without investing in new power sources, these deals could divert low-carbon electricity from other users, leading to increased fossil fuel consumption to meet overall demand. The energy demand for AI is expected to continue growing. Conventional supply and demand dynamics suggest that as AI’s electricity use increases, energy costs will rise, potentially forcing the industry to become more efficient. However, given the unique nature of the AI sector, the largest companies might choose to absorb these costs, investing billions of dollars to maintain their competitive edge.

 

The most expensive data centers in the AI sector are those used to train "frontier" AI systems like GPT-4 and Claude 3.5, which are at the cutting edge of capability. OpenAI, Google’s Gemini, and Anthropic, the maker of Claude, are in a fierce competition to lead this field. The "winner takes all" nature of this competition means that companies are driven to spend vast sums on developing the most advanced AI systems, regardless of the energy costs.

 

The pursuit of Artificial General Intelligence (AGI)—AI systems capable of performing any human task—further amplifies this trend. The potential to monopolize such transformative technology could justify spending hundreds of billions of dollars on a single training run.

 

While there are ongoing breakthroughs in AI technology that allow for more efficient use of electricity, this efficiency often leads to even more powerful AI systems being developed, rather than reducing overall energy consumption. This phenomenon, known as "Jevons' paradox," was first observed by economist William Stanley Jevons in the context of steam engines: improvements in efficiency led to increased overall use of coal as the cost of steam power fell and new applications emerged.

 

As AI technology continues to advance, the tech industry faces a critical challenge. Balancing the insatiable energy demands of AI with the imperative to meet climate goals will require significant innovation and investment in renewable energy. Without such efforts, the sector’s environmental aspirations may remain out of reach, overshadowed by the relentless drive for AI supremacy.

 

Credit: Mother Jones 2024-07-09

 

Get our Daily Newsletter - Click HERE to subscribe

[Cryingdick]

Nano nuclear reactors can be used. I am looking at a few companies that do this to invest in. It's basically a small nuclear generator that can be dropped in at any data center or factory to make its own grid. As far as waste goes these days it seems we should be able to launch the waste into space at some point. 

[Peterphuket]

8 hours ago, Cryingdick said:

Nano nuclear reactors can be used. I am looking at a few companies that do this to invest in. It's basically a small nuclear generator that can be dropped in at any data center or factory to make its own grid. As far as waste goes these days it seems we should be able to launch the waste into space at some point. 

Have mentioned this before but do you know how much water these data centres and possibly those small modular nuclear generators use.