Even if AI turns out that it is not as substantial energy as people as people, it can still mean problems with energy networks in the USA.
Technology companies are already burning the growing amounts of electricity to train and conduct up-to-date AI models. And they ask for a lot more electricity when they try to compensate for each other. This growing demand is already beginning to transform the energy system, and the tools try to build up-to-date gas plants and pipelines.
But all these plans to transform the American energy system can be based on AI bubbles. Because excessively excited investors pump money to technology companies who are afraid to lose fashion, but still threatened with the development of AI tools, which eventually floped, tools are also dealing with a wave of speculation over the energy needs of data centers.
“Uncertainty is disturbing”
Uncertainty is uncertain, taking into account the costs that Americans can end up paying Higher media bills and more pollutionlast report warns. The transition to cleaner and more affordable energy sources is slowly progressing in the USA. This is in danger, unless technology companies and media require greater transparency and decide on more renewable sources of solar energy and wind energy.
“Although AI Boom provides exciting possibilities, there are many risk that it is not approaching energy needs with intentional and conscious reaction, which takes into account long -term effects,” said Kelly Poole, the main author of the report published in this month by a group of shareholders as a group of spokesperson as SOW and organization of the SOW environment and organization of the SOW environment and organization of the environment and environmental organization and environmental organization organization and environmental organization
The fleet of national gas -fired power plants would escalate by almost a third if all up-to-date gas projects proposed between January 2023 and January 2025, as the AI generative industry was sizzling. The amount of up-to-date gas capacity, which, proposed by the media and independent developers, increased by 70 percent at that time, driven largely by the growing demand for electricity in the data center.
Before the generative AI, the demand for electricity was almost flat For over a decade for energy efficiency benefits. But up-to-date data centers, counterfeit for artificial intelligence, are much more energy -saving than in the past. The computer stand in a conventional data center can utilize 6-8 kilowatt power-approximately the equivalent of the power used by three houses in the USA, Dan Thompson, the main research analyst from S&P Global, explained in briefing. AI, however, requires stronger computer systems to perform more elaborate tasks. According to Thompson, the power required to start one of these high -density racks is a power worth about 80 to 100 houses or up to 100 kilowatt.
“You look basically at the distribution of the power of a small town,” he said.
Why does it matter? Power forces actually act as an uncertain act of balancing. If the power supply cannot meet the escalate in demand, this can lead to higher media bills and potential failures. On the other hand, excessively building up-to-date risks related to the creation of orphaned assets, for which tools and their clients end, regardless of whether they actually need them in the long term. That is why it is so significant to get an right forecast of future demand.
“You look basically at the distribution of the power of a small town.”
And although artificial intelligence consumes a lot of energy, projections for the future become shadowy. “Speculators flood the market,” says the report, trying to build and reverse data centers. Trying to overtake long waiting times to connect to the energy network, some of these speculators ask for power, even before they get capital or customers to make sure they can introduce a project to the finish. There may also be some double or triple counting (or more) when it comes to forecasting AI energy demand because of the upcoming developers more than one tool to get several quotes.
In the southeast, the main center of data centers, tools provide for up to four times a larger escalate in demand compared to independent analyzes of industry trends, According to the report at the beginning of this year at the Institute for Energy Economics and Financial Analysis (IEEFA). At the national level, he prepares up to 50 percent of a greater escalate in demand than expects the technology industry, separate report From December 2024, the United States.
The media themselves recognized this risk of recent earnings. The proposed projects trying to connect to the network “They can be overstated anywhere from three to five times, which can materialize,” said Jim Burke, general director of Vistra Energy based in Texas in Q1 call This year.
Despite the uncertainty, they still build up-to-date gas power plants and pipelines to satisfy this demand. After all, building a up-to-date infrastructure is one of The most lucrative ways to increase profits. And now the Trump administration – whose campaign was Excited with oil and gas insert – encourages relying on fossil fuels. For example, in Louisian Local Entergy Proposed building of three new gas plants To feed the giant up-to-date meta data. The data center is estimated that it consumes the same number of electricity up to 1.5 million houses and leads to 100 million tons of carbon dioxide emissions over 15 years.
This is a clear contrast to Biden administration goal consisting of starting the power network at 100 % energy without coal pollution by 2035. The only way to stop climate change in your paths is to get rid of pollution of heating planet from fossil fuels. Building a rush of up-to-date gas infrastructure, of course, moves the nation in the opposite direction.
There are solutions that minimize all these risk, as you indicate and Sierra Club in their report. Tools may require developers to disclose the number of other media to which they introduced a data center proposal and how far they are in the finalization of the project. During recessing contracts, they may also require long -term service contracts, escalate deposits insurmountable and escalate the fees for cancellation of the project.
Technology companies also clearly play a substantial role, improving the energy efficiency of their technologies and investing in renewable energy sources. There have been technological giants for years, including Amazon, Meta and Google The best corporate buyers renewable energy. Entering this type of long -term contracts for the construction of up-to-date solar and wind farms may now have an even greater impact, counteracting the withdrawal of financial incentives by the Trump administration for renewable energy sources, if companies are ready to prioritize their sustainable development goals, such as their AI ambitions.
0 Comments