The race for artificial intelligence supremacy is underway among the world’s biggest companies, and so far investors are on board. NVIDIA SPDR Dow Jones Industrial Average ETF Trust, a leading artificial intelligence (AI) chip maker, is currently the world’s second-largest company by market capitalization. Microsoft-backed AI model leader OpenAI raised $6.6 billion in an investment round last month, one of the largest private investments in history. Amazon and Google are among the cloud service providers that are making major investments in this technology, along with Microsoft.
This investment frenzy is based on the hypothesis that AI will soon increase productivity across the economy. With its vast data processing power, AI will inevitably transform life and work in an information-driven economy. We are only just beginning to understand its benefits and risks, so a cautious approach with significant industry oversight is warranted.
Meta’s Facebook data center in Eagle Mountain, Utah. (Photo credit: GEORGE FREY / AFP) (Photo credit: GEORGE … (+) FREY/AFP via Getty Images)
AFP (via Getty Images)
AI proponents highlight the potential for new grid management and other decarbonization tools. But for now, the AI industry’s most significant climate impact is through new fossil fuel electricity generation and the associated increase in greenhouse gas emissions. This will exacerbate the risk of economic collapse caused by climate change and dash the AI industry’s hopes of building a strong customer base and achieving profitability.
Ensuring that AI is built on clean energy sources requires action from technology and utility industry leaders, investors, regulators, and policymakers. This not only limits climate-related damage to the global economy, but also helps the AI industry help itself.
Data center construction boom fueled by AI
After analyzing U.S. grid planners’ forecasts for 2023, energy analyst Grid Strategies has come to a surprising conclusion. The forecast for five-year growth in electricity demand is 4.7%, up from 2.6% in the previous year.
Grid strategy analysis based on Federal Energy Regulatory Commission filings
grid strategy
Much of this projected growth is related to new data centers. McKinsey estimates that data centers will account for 30-40% of new net demand by 2030. Additionally, data center growth is strongly linked to AI. Goldman Sachs predicts that by 2028, AI will account for approximately 19% of all data center power demand.
Other important drivers of increased U.S. electricity demand are industrial reshoring and the electrification of transportation, buildings, and industry, both of which are central plans for the Biden-Harris administration’s worker-friendly energy transition plan. Yes, and encouraged by the landmark Inflation Reduction Act.
Among these energy demand drivers, AI is unique in that competitive advantage tends to be given to companies that use more energy (AI companies tend to be more energy efficient by processing more data). , allowing you to train larger and potentially more performant models). This suggests that the rapid acceleration of AI’s energy consumption is unlikely to end soon.
Climate risks that AI cannot manage
Power companies are delaying planned closures of coal-fired power plants as demand for AI-powered energy is predicted to increase rapidly. S&P Global recently predicted that the U.S. power sector will retire just 105,000 megawatts of coal-fired power by 2035, a 21% decrease in retirements predicted the previous year.
Gas-fired power plants, once threatened by the rise of cheap renewable electricity and battery storage, have also been given a new lease of life. More new gas generation capacity was announced in the U.S. in the first half of 2024 than in all of 2020.
New gas-fired power plants announced as of July 1, 2024
sierra club
Unless a new path is chosen, this return to fossil fuel power generation could represent a major setback for the energy transition and efforts to prevent climate collapse.
Avoid new fossil fuel infrastructure this decade
Technology industry leaders are pledging to address these climate risks, but many of them will take place beyond the 2030s, outside the period when scientists say aggressive emissions cuts are most essential. We are to make investments that will bear fruit. For example, Microsoft last month pledged to bring power online from the restarted Three Mile Island nuclear facility owned by Constellation Energy, and this week Google pledged to bring power online from a small modular nuclear reactor being developed by startup Kairos Power. announced plans to purchase electricity.
The Inflation Control Act (IRA) production tax credit is key to making these projects financially viable. Assuming the nuclear waste challenge is resolved, they could be a significant success for the IRA in driving the decarbonization of the U.S. power system. But it will do nothing to prevent climate change from new fossil fuel infrastructure built over the past decade. Fires, floods, and other disasters related to greenhouse gas emissions are already escalating rapidly, and production from existing fossil fuel projects has already exceeded Paris Agreement consumption levels.
David Hester of Horseshoe Beach, Florida, inspects the damage to his home after Hurricane Helen made landfall in September 2024. The storm killed at least 228 people in five U.S. states. (Photo by CHANDAN KHANNA/AFP) (Photo by CHANDAN KHANNA/AFP, Getty Images)
AFP (via Getty Images)
The high-tech and utility industries need to work together to aggressively reduce gas plant expansion, accelerate the planned retirement of coal-fired power plants, and accelerate efforts to install renewable energy and energy storage. This would significantly reduce systemic climate risks and help the United States meet its Paris Agreement commitments.
Investors need to address climate science to fulfill their fiduciary duties
To stop AI-driven emissions from skyrocketing, we need to break the myth that climate change is solely the responsibility of governments. Institutional investors have a fiduciary responsibility to ensure that companies address climate-related systemic financial risks. This is especially true for many investors with diversified portfolios that are highly exposed to economy-wide climate shocks.
Climate change particularly jeopardizes the profitability of AI. Unlike other new technologies, for AI to be profitable, it must generate enough revenue to offset significant annual capital expenditures. Leading observers of the AI industry, such as Jim Covello, head of equity research at Goldman Sachs, question whether AI can secure these returns, as reliability challenges limit its usefulness. I have doubts.
The case for AI investing in this is that companies in nearly every industry sector want to understand how AI can give them a competitive edge. However, if the economy deteriorates significantly, this widespread adoption becomes much less likely. Many potential customers may not be willing or able to add significant expense to their budgets for unproven technology with reliability issues.
Addressing the risk of economic collapse that climate change could pose to AI’s customer base is now a business imperative for the nation’s largest companies and their investors.
How concerned should we be about the risk of a climate-related crash? Investors should at least be sensitive to emerging climate science and its implications. Global stock valuations could fall by up to 40% if emissions reductions do not accelerate, according to an October 2024 analysis by the EDHEC Risk Climate Impact Institute. The projections do not take into account tipping points that, if reached, would further destabilize the economy, such as the collapse of West Antarctic glaciers and the melting of methane-rich Arctic permafrost.
Proposing a carbon price that emphasizes AI
Technology industry analyst Robert Wright says OpenAI CEO Sam Altman is helping us move forward by promoting “the idea that when it comes to technological change, and especially advances in AI, the faster the better.” He claims that he led the group in a dangerous direction. He called for withholding subsidies for AI expansion until governments address climate and safety risks, and instead imposing special taxes on the electricity consumed by AI data centers to reduce these risks. I am proposing.
Sam Altman, OpenAI CEO, May 16, 2023 (Bill Clark/CQ-Roll Call, Inc, via Getty Images)
CQ-Roll Call, Inc (via Getty Images)
Similarly, Brian Deese, former director of the National Economic Council in the Biden-Harris administration, has announced that he is working with utility companies and technology companies to bring clean energy online to power AI’s large-scale computing needs. We propose new policy incentives for A key recommendation is that the federal government impose a fee on the high-tech industry to be paid into the Department of Energy’s Clean Energy Deployment Fund.
Expecting Congress to act on these ideas will require leadership from the technology industry on carbon pricing, which has been lacking so far.
Short-term efforts to decarbonize the power system
Fortunately, meaningful action to decarbonize the power system outside of Congress is possible. Technology and utility industry leadership opportunities include expanding 24-hour carbon-free energy procurement, increasing transparency around forecasted demand, establishing green energy pricing, and modernizing interregional and other power grids. This includes support for development.
Procuring carbon-free energy around the clock is perhaps the most controversial of these efforts, as it is more expensive than the traditional approach of purchasing renewable energy wherever it is cheapest. . However, as a 2024 analysis by Princeton University’s Jesse Jenkins and colleagues showed, traditional procurement approaches have not been effective in reducing emissions. Only by procuring clean energy at the same time as they consume energy will large customers drive investments beyond wind and solar to include energy storage and robust clean energy technologies such as geothermal and hydropower.
In Nevada, Fervo Energy’s first enhanced geothermal project will send carbon-free electricity to the … (+) grid to power Google data centers. (AP Photo/Ellen Schmidt)
Copyright 2023 Associated Press. Unauthorized reproduction is prohibited.
Google has provided significant leadership in the movement toward 24-hour, carbon-free energy procurement. Advocating for carbon-free energy generation and transmission in the Santee Cooper Integrated Resources Plan is an example of how high-tech companies can aggressively advocate for short-term emissions reductions, not just in the coming decades.
Fairness for ratepayers, taxpayers and communities
Data centers are driving up the market price of energy, and this trend will continue until interconnection queues and other energy supply constraints are resolved. Tech companies, utilities, and state regulators should be able to reap profits from their data centers, rather than other companies, to cover the costs of new power plants, power lines, and other system upgrades required by the AI-powered data center boom. The onus is on the tech companies that are gaining to make sure they pay the bills. rate payer.
Efforts are already underway to force ratepayers to pay for data center system upgrades. For example, utilities AEP and Exelon recently challenged Amazon’s nuclear power proposal, saying it could shift up to $140 million in transmission costs to daily ratepayers. Similarly, the Maryland Office of the People’s Counsel recently announced a regional transmission project that would charge Maryland ratepayers a significant portion of $5 billion in capital expenditures to build transmission equipment for a new data center in Virginia. opposed PJM’s proposal.
Regulators and policymakers also need to ensure that technology companies avoid and minimize the impact of the AI-powered data center boom on local environments. Fossil fuel companies have a long track record of avoiding responsibility for their environmental impacts, especially in low-income communities and communities of color that have long served as energy production hubs. At least 120,000 oil and gas wells across the United States have been abandoned by their original owners, leaving a lack of financially viable operators. Local residents will suffer the health effects of this harmful legacy, and the cleanup costs will ultimately be borne by taxpayers. Regulators and policymakers must learn from this history and ensure that the technology and utility industries take full responsibility for the environmental impacts of AI energy infrastructure.
