Originally published: May 16, 2024
For years tech giants have been helping climate catastrophists shut down reliable fossil fuel electricity, falsely claiming they can be replaced by solar/wind.
Now the grid they've helped gut can't supply their growing AI needs.1
-
For the last decade, tech giants such as Apple, Microsoft, Meta, and Google have, through dedicated anti-fossil-fuel propaganda and political efforts, promoted the shutdown of reliable fossil fuel power plants in favor of unreliable solar and wind.
-
Tech giants have propagandized against reliable fossil fuel power plants by falsely claiming to be "100% renewable" and implying everyone could do it. In fact, they have just paid utilities to credit them for others' solar and wind use and blame others for their coal and gas use.2
-
In addition to their “100% renewable” propaganda, tech giants directly endorsed people and policies who shut down reliable fossil fuel power plants.
E.g., The RE100 coalition, including Google, Apple, Meta, and Microsoft, advocates for policies to “accelerate change towards zero carbon grids at scale by 2040.”3
-
Companies' propaganda that solar/wind could rapidly replace fossil fuels has proven false.
Statewide blackouts in California (2020) and Texas (2021) were caused by the failure of solar/wind—which can go near zero at any time—to make up for lack of reliable fossil fuel capacity.
-
Thanks in significant part to tech giants' advocacy, we have now shut down enough reliable power plants to be in a nationwide electricity crisis.
For example, most of North America is at elevated/high risk of electricity shortfalls between 2024-2028.4
-
The anti-fossil-fuel, pro-unreliable solar and wind political climate that tech giants have fostered is getting much worse, as the Administration has pledged to further reduce reliable electricity supply via power plant shutdowns and add artificial demand through EV mandates.
Biden's EV mandate - a dictatorial attack -
While for years tech giants didn't seem to have any concern about the electricity supply disaster their propaganda and policies were bringing about, they are now very interested because of the accelerating power requirements of computing, above all the hyper-competitive AI space.
-
To function at its potential, AI requires massive amounts of power. E.g., state-of-the-art data centers can require as much electricity as a large nuclear reactor.5
-
Electricity demand from US data centers already doubled between 2014 and 2023. Now with the fast growth of energy-hungry AI, demand from data centers could triple from 2.5% to 7.5% of our electricity use by 2030, according to Boston Consulting Group.6
-
In large part due to AI, nationwide electricity demand is projected to skyrocket. Official 10-year projections for the US have summer and winter peak demand rising by over 79 gigawatt and over 90 gigawatt. 90 gigawatt is equivalent to adding the entire power generating capacity of California (!)7
-
Given the woeful underpowered grid that AI giants have helped bring about, dramatically rising demand from AI will not only contribute to massive electricity shortages, but it will also destroy a lot of potential for AI to occur in the United States.
-
Limited and expensive electricity will force data centers to operate with higher cost or lower capacity within the US—or take a performance hit in the form of increased latency (which can drastically reduce the value of the product) by moving offshore.
-
Not only is offshoring data centers destructive from an economic standpoint, it also poses a substantial security risk. E.g., Building a data center in China—which we already depend on dangerously for critical minerals—gives the CCP physical power over more parts of our economy.
-
Economically, data centers are a gold mine of opportunities.
Globally, data centers employed 2M people full-time in 2019, many in high-skill/high-pay jobs—and this number is forecast to increase nearly 300K by 2025.
Our gutted grid will cost many Americans these opportunities.8
-
In the face of woefully inadequate electricity supply for their AI goals, tech giant CEOs are finally speaking up about the lack of power.
E.g., Meta CEO Mark Zuckerberg said in an interview that energy will be the #1 bottleneck to AI progress.
-
It is not enough for tech giants to warn us about the lack of reliable power. They need to take responsibility for their anti-fossil-fuel advocacy that helped cause it. And they need to support energy freedom policies that allow all fuels to compete to provide reliable power.
End preferences for unreliable electricity -
An example of a tech giant influencer not taking any responsibility for causing the electricity crisis is BlackRock CEO Larry Fink, who pushed companies and governments to adopt "net-zero" policies using mostly solar/wind, but now admits they can't power AI data centers!
-
A better attitude toward electricity was expressed by OpenAI CEO Sam Altman: "There will always be people who wait and sit around and say ‘we shouldn't do AI because we may burn a little more carbon’... the anti-progress streak" and this "is something that we can all fight against.”9
-
America faces a choice. We can either continue our current trajectory, descend into a Third World grid, and become totally inhospitable for AI, or we can adopt energy freedom policies and become a world leader in both AI and electricity.
-
Share this article with tech giant CEOs and tell them to publicly apologize for damaging our grid and to commit to energy freedom policies.
Google: @sundarpichai (sundar@google.com) Apple: @tim_cook (tcook@apple.com) Meta: @finkd (zuck@fb.com) Microsoft: @satyanadella (satyan@microsoft.com)
Michelle Hung contributed to this piece.
References
-
WSJ - The Coming Electricity Crisis
Washington Post - Amid explosive demand, America is running out of power↩
-
Bloomberg - Data Centers Now Need a Reactor’s Worth of Power, Dominion Says↩
-
Lawrence Berkeley National Laboratory - United States Data Center Energy Usage Report (2016)↩
-
NERC - 2023 Long-Term Reliability Assessments
CA Energy Commission - Electric Generation Capacity and Energy↩
-
Uptime Institute - The people challenge: Global data center staffing forecast 2021-2025↩