The burgeoning impact of generative AI, which is reshaping industries as diverse as medicine, education, music, and computing, across the globe, has ushered in a fresh concern. While the world is immersed in the consumption of everything AI has to offer, which goes beyond chatbots and image generation tools, the challenge posed by the technology – significant energy consumption – has been noted.
A recent report noted that the demand for electricity to fuel AI’s rapid growth is staggering, raising environmental and logistical concerns that are yet to be fully addressed.
The power consumption of AI has become a critical issue as the technology matures. According to recent reports, major AI developers like Microsoft and Google consumed a combined 48 TWh (terawatt-hours) of electricity in 2023 alone, with each company using 24 TWh. This consumption level surpasses that of over 100 nations, including countries like Ghana and Tunisia, according to a detailed analysis by Michael Thomas.
Tekedia Mini-MBA edition 15 (Sept 9 – Dec 7, 2024) has started registrations; register today for early bird discounts.
Tekedia AI in Business Masterclass opens registrations here.
Join Tekedia Capital Syndicate and invest in Africa’s finest startups here.
To put this into perspective, the energy needs of Microsoft’s and Google’s AI operations could meet the entire electricity consumption of Azerbaijan, a nation with a population of 10.14 million and an estimated GDP of $78.7 billion.
Electricity consumption is not the only concern. AI systems also require substantial amounts of water for cooling. It has been noted that a single AI query can use as much water as a standard water bottle.
The environmental degradation associated with AI advancements mirrors concerns previously raised about cryptocurrency mining. Cryptocurrency, particularly Bitcoin, has been criticized for its massive energy consumption. The energy-intensive process of mining Bitcoin led to significant environmental concerns, contributing to China’s decision to ban cryptocurrency mining in 2021.
The country cited the substantial electricity consumption and its environmental impact as primary reasons for the ban. Bitcoin mining’s power demands were so high that, at its peak, it consumed more electricity annually than countries like Argentina and Ukraine combined.
Despite the clear parallels between AI and cryptocurrency in terms of energy consumption, no country, including the United States, is taking substantial steps to address the electricity usage challenges posed by AI.
This lack of action is concerning, given the exponential growth in AI applications and the associated energy demands. AI’s extensive use of resources, which underlines its impact on the environment is emerging amid efforts by companies like Google and Microsoft to champion renewable energy and seek alternative power sources.
Elon Musk, cofounder of OpenAI and founder of xAI, Grok, has been vocal about the potential dangers of unchecked AI growth. Musk has warned that we are on the verge of a significant technological breakthrough with AI, but there may not be enough power to sustain its advancements by 2025.
However, in response to these challenges, some AI leaders are exploring alternative power sources. Sam Altman, CEO of OpenAI, is investigating nuclear fusion as a potential solution to power AI systems. Microsoft has also taken steps in this direction, partnering with Helion to develop nuclear fusion technology. Helion aims to generate nuclear energy by 2028, with current efforts focused on training large language models (LLMs) to expedite the regulatory process.
While nuclear fusion promises a non-existent environmental impact, many scientists and researchers argue it may be too late to address the climate crisis solely with this technology. They suggest that fission and renewable energy sources might offer more immediate and practical solutions.
The high AI energy consumption poses a challenge to the financial success of AI-driven companies like Microsoft. The economic incentives for continued AI development are notable in Microsoft’s market valuation, which recently surpassed $3 trillion, a milestone attributed to its early adoption and investment in AI.
CEO Satya Nadella credited the company’s financial performance, including increases in revenue, operating income, and net income, to the transformative potential of AI. However, this economic success further complicates the issue, as the drive for profit may overshadow the environmental and logistical challenges associated with AI’s energy demands.
The US has been reluctant to roll out regulations to address other AI concerns. Though the Senate has introduced a bill dubbed COPIED Act to address some of the issues, it largely focused on the deepfake concerns. Also, the European Union’s AI Act, even though it covers a wide range of issues, did not touch energy consumption. This means that AI’s high energy consumption is a fresh concern that policymakers must include in their cart of things to do about AI.