Op-ed: Cutting the environmental cost of cutting-edge technologies like AI and Bitcoin
As both artificial intelligence and cryptocurrency gain traction in several aspects of our lives, their escalating energy demands have become a critical issue. Like all technologies that improve our quality of life at the cost of using resources, generative AI and cryptocurrencies—Bitcoin (BTC) in particular—have already produced tremendous value while consuming tremendous amounts of energy.
This does not mean, however, that we halt them altogether. It simply means that it’s vital to continuously asses their energy consumption to ensure that the utilization of these innovative tools does not lead to unsustainable power usage. In other words, we can enjoy the benefits of revolutionary tech, but only as long as we don’t wreck our homeworld in the process.
AI’s Energy Consumption, Especially ChatGPT
AI’s energy consumption, particularly for cutting-edge models like OpenAI‘s ChatGPT, has seen a dramatic surge. This increase is not confined to the energy-intensive training phase but extends to the operational phase, where machine learning algorithms are deployed. For example, Hugging Face’s multilingual text-generation AI consumed about 433 megawatt-hours (MWh) just in its training phase, which is sufficient to power 40 average U.S. homes for a year. The ongoing energy demand by AI tools, encompassing both the training and operational phases, already rivals that of small countries, signaling a trend that is just beginning.
These figures underscore the substantial energy footprint of AI technologies, particularly as they become increasingly central in various applications, from healthcare to finance. The growing capabilities and applications of AI have sparked debates about its potential drawbacks, including the risk of increased energy use contributing to climate change. The significant energy requirements for running AI algorithms and machine learning processes pose a challenge in maintaining environmental sustainability. However, it seems very unlikely that further development of the technology will slow down anytime soon.
Bitcoin’s Energy Consumption
Bitcoin’s energy consumption presents a different but equally critical challenge. The Bitcoin Energy Consumption Index reveals a striking annual carbon footprint of 76.79 million tonnes of carbon dioxide equivalent (MTCO2Eq), comparable to the carbon emissions of countries like Oman. Its electrical energy consumption stands at 137.68 TWh, almost matching the power consumption of New Zealand.
Bitcoin’s energy consumption, particularly its mining operations’ significant reliance on fossil fuels, has been a longstanding object of criticism. However, data from recent years suggests a distinct trend toward more sustainable energy sources. According to an August report from the Cambridge Centre for Alternative Finance, technological advancements in mining hardware, such as the development of more efficient application-specific integrated circuits (ASICs), have played a crucial role in this shift. These improvements have not only enhanced the energy efficiency of Bitcoin mining operations but also contributed to a reduction in their overall carbon footprint. While Bitcoin mining still consumes a considerable amount of energy, these trends indicate a move towards a more environmentally conscious approach in the industry, aligning with global sustainability goals.
Solutions
At this point, it is plain that both of these technologies have found voracious market demand, meaning companies will continue to service that demand. As such, the matching demand for greener energy solutions to power them is not just preferable but imperative.
In the field of AI, there is a growing emphasis on optimizing algorithms for greater efficiency and using more energy-efficient hardware. The development of AI models that are less energy-intensive without compromising their capabilities is a key focus, but at a certain point, the burden of reducing environmental degradation must fall on the efficiency of the energy source itself. As such, integrating AI with renewable energy sources is a vital step towards managing its overall energy footprint.
For Bitcoin, a significant shift is taking place toward more sustainable energy sources, driven by the inherent energy-intensive nature of its Proof of Work consensus mechanism. Unlike some other cryptocurrencies that may benefit from a transition to -Proof of Stake—most notably Ethereum, which successfully completed such a transition in 2022, reducing its energy consumption by 99%—Bitcoin’s value and security are closely tied to its PoW architecture. Therefore, a shift in the consensus mechanism is unlikely.
Instead, the focus is on harnessing more sustainable energy sources to power the network. This shift is crucial as the demand for Bitcoin and its underlying technology continues to grow. The technology itself is driving improvements in energy sustainability because, firstly, the widespread adoption and preference for Bitcoin means it will continue to be used extensively, and secondly, there’s a pressing need to ensure that its operation doesn’t exacerbate environmental issues. Thus, the future of Bitcoin mining is increasingly leaning towards leveraging renewable energy sources and innovative technologies to enhance efficiency and reduce the environmental impact, ensuring that the network remains both secure and sustainable.
The allure and utility of AI and Bitcoin make them essential technologies in our modern world. However, their energy-intensive nature necessitates a thoughtful approach to their usage. As we continue to integrate these technologies into our daily lives, it is crucial to develop and adopt practices that ensure their sustainability. Striking the right balance between harnessing their benefits and maintaining environmental responsibilities is key to safeguarding their future and ours. The journey towards sustainable usage of AI and Bitcoin is not only about technological advancements but also about making conscious choices that align with our environmental commitments.