On The Wire: Today's Top Stories In Finance & Tech
Your daily briefing of some of the most important stories from the crypto, finance, and tech space.
AI's energy use will balloon in the coming years, but the narratives around it are misguided.
Bitcoin is frequently criticized for the huge amount of energy used by proof-of-work miners, who use dedicated hardware to crunch through trillions of hashes every second to help secure the network.
Last year, estimates suggested that Bitcoin uses around 150 terawatt-hours of electricity per year, enough to power a country like Argentina. While it's hard to come up with precise figures, Bitcoin is likely responsible for a little more than 0.5% of the world's electricity usage, though more than half of that comes from renewable sources.
Now, there's another villain in town: AI.
The artificial intelligence (AI) industry could consume as much energy as a country the size of the Netherlands by 2027, warns a new study. Big tech firms have scrambled to add AI-powered services since ChatGPT burst onto the scene last year. They use far more power than… pic.twitter.com/XP9gKnBFyn
— Net Zero Watch (@NetZeroWatch) October 11, 2023
A new study warns that the rapid expansion of artificial intelligence (AI) means that the technology could end up consuming as much power as the Netherlands by 2027: Potentially over 130 TW/h. That's almost as much as Bitcoin.
AI applications demand a high level of computational resources, just like PoW crypto mining. The study, conducted by Alex De Vries, a PhD candidate at the VU Amsterdam School of Business and Economics, looked at the rate AI is currently expanding, and especially the role of Nvidia, the chip designer that supplies around 95% of all AI processing hardware. De Vries used Nvidia's expected supply volume by 2027 to estimate that AI could consume anywhere from 85 to 134 terrawatt-hours (TWh) of electricity annually.
De Vries suggests that AI should be used only when absolutely necessary, and highlights the need for more transparency within the AI sector regarding energy consumption and environmental impact.
Doomberg, a pseudonymous collective of journalists and industry analysts that has become one of the most successful financial Substack publishers, is well-known for its informed views on the importance of energy. The amount of energy the world consumes correlates closely with the growth of the global economy.
We like to say that energy is life, and your standard of living is defined by how much energy you get to harness.
The quest to access more energy, and use what energy we possess more efficiently, is closely interwoven with economic prosperity—and with technological progress.
Perhaps unsurprisingly, the history of the world—especially the last 200 years—is one of exponentially increasing energy use.
In a free market, what "should" take place is irrelevant. So long as something is profitable, it will almost certainly happen.
Industries that are intensely harmful (fast food, for example) therefore thrive, because there is demand. It takes a significant enforcement effort to successfully ban industries that are deemed illegal, should the profits be considered by criminals to outweigh the risks—drug and human trafficking being two obvious examples.
The point is that the free market isn't concerned with morality, but with profit. If something is profitable, someone, somewhere will do it. In the case of Bitcoin mining, or AI, a ban or regulation designed to curtail it would simply mean that activity is pushed to a more favorable jurisdiction.
In the case of China banning Bitcoin mining, a lot of activity ended up in neighboring countries, as well as the US. If a major economy decided to restrict AI on energy grounds, it's a fair assumption that others would seek to capitalize on that.
The challenge therefore becomes now how to limit AI, but how to maximize clean energy production.
Despite these challenges, there is optimism that AI could also contribute to solving environmental problems, as well as other pressing issues of our time, such as finding new treatments for cancer and other medical conditions.
Recent experiments suggest AI could help reduce contrails from aircraft, which contribute to global warming. Additionally, AI may accelerate research into nuclear fusion, offering a green, limitless power supply.
Any environmental benefits, though, are incidental to the adoption of AI, which will happen regardless. One way or another, humanity's addiction to energy isn't going to end.
Subscribe to our newsletter and follow us on Twitter.
Everything you need to know about Blockchain, Artificial Intelligence, Web3 and Finance.