The AI boom sure seems to be anti-ESG in a few ways. The more we learn about AI, the more difficult it will be for companies to increase their use of the technology while concurrently touting ESG bona-fides. Here are a few things to know about AI’s impact on land use, energy consumption, air emissions and employment.
Land use: Bloomberg reported that Blackstone is currently building new AI data centers in the Phoenix area. One is currently under construction on a parcel of land “larger than 60 football fields” while another in the planning stage on 400 acres, “all but erasing the land’s farming roots.” QTS, the AI company Blackstone acquired in 2021, already faced challenges in acquiring land: in Manassas, Virginia “residents and conservationists fought a proposed multibillion-dollar, 900-acre development with tracts next to a state forest and a Civil War battlefield.” Municipal land use and zoning requirements are trying to catch up to the technology, but still lag. There will be controversy around those changes, and I have no doubt that environmental justice matters will play an important role.
Energy consumption: QTS “estimates that its projects, once complete, will tap into some 6 gigawatts of electricity, equal to the needs of roughly 5 million homes.” At another AI data center company, Applied Digital Corporation, CEO Wes Cummins admitted they already need more energy than planned at the new Garden City, Texas facility – opened less than two months ago. Cummins said in management’s January 16, 2024 analyst call that improvements are needed at “a substation that’s in the area to get fully up to the 200 [megawatts of capacity]. There’s two ways to go there. There’s getting approved for wind plus grid is one route, and the other is this, I think it’s a capacitor bank that needs to be installed…”
Scientific American published an article last October predicting that by 2027, AI servers “would consume at least 85.4 terawatt-hours of electricity annually – more than what many small countries use in a year.” I wonder how 85.4 TW of new demand in each of the next 3 years fits into global net zero and renewable energy plans, expectations and goals. This poses a major quandary for Scope 3 emissions determinations for computer hardware manufacturers and companies just using AI processing – which at the moment seems like everyone.
Air emissions: AI and data centers do have air emissions. Their massive use of electricity and their criticality require emergency backups. Industrial-size generators typically fill this need, fueled by propane, natural gas or diesel. Generators run – and therefore emit air pollutants – when the main electricity source is cut-off, curtailed or when doing so is cheaper than paying a power utility (e.g., peak-shaving). They also run – and emit – for short periods (anywhere from 15 minutes to an hour) at least monthly as part of preventive maintenance programs. Finally, there are leaks and releases from fuels and other chemicals.
Employment: Is AI going to take jobs away from humans? Definitely. According to CNN.com, IBM CEO Arvind Krishna said just this week “that the company plans to use AI to replace 7,800 jobs over the next five years.” Yesterday, UPS announced it will let go 12,000 managers which, Bloomberg reports, was “made possible by new technologies including artificial intelligence,” CEO Carol Tome’ said. Reports about other companies vary, but CNBC claimed “37% [of companies surveyed] say the technology replaced workers in 2023. Meanwhile, 44% report that there will be layoffs in 2024 resulting from AI efficiency.” And May 2023 alone saw 4,000 jobs lost to AI, according to data from Challenger, Gray & Christmas.
The AI push won’t slow down anytime soon (although who remembers the metaverse??). Companies looking to expand their use of the technology should be thoughtful in how doing so aligns – or conflicts – with sustainability, climate, employee satisfaction/retention and DEI programs. Corporate ESG positions, metrics and goals may need to be adjusted to reflect AI use. You might also want to begin setting the stage for public reporting and even ESG metrics for CEO compensation.
I’m sure this won’t be the last time I write about AI.
If you aren’t already subscribed to our complimentary ESG blog, sign up here: https://practicalesg.com/subscribe/ for daily updates delivered right to you.
Disclaimer: I own shares in Applied Digital Corporation.