News stocks
- The artificial intelligence boom is straining America's power grid.
- AI queries can use 30 times more energy than traditional searches, massively escalating power demand.
- "The cost of that Gen AI architecture is freaking out of control," Baird strategist Ted Mortonson said.
Sign up to get the inside scoop on today’s biggest stories in markets, tech, and business — delivered daily. Read preview
Thanks for signing up!
Access your favorite topics in a personalized feed while you're on the go.
Advertisement
First it was electric vehicles. Then it was bitcoin. Now its AI.
All three trends have sparked ongoing concerns about the power-hungry nature of new technologies as they push America's shaky power grid to the limit.
It appears that the AI boom, which is still in its early days, might be the biggest stressor on the country's electric grid.
That's because mega-cap tech companies are spending hundreds of billions of dollars on power-hungry AI-enabled GPU chips, which are housed in massive data centers that require state-of-the-art cooling technologies to dissipate the heat generated from the computers.
Advertisement
AI research company Hugging Face has estimated that generative AI search queries can use 30 times as much energy as a traditional Google search.
And with hundreds of millions of users already interacting with AI tools like ChatGPT, the power demand for AI technologies is only set to rise.
Bank of America put into perspective the challenges faced by the power grid as it grapples with surging demand from AI data centers.
"Manufacturing, data centers, artificial intelligence, and the push for electrification are expected to add massive demand to an already-tight electrical grid. Intermittent wind and solar cannot provide the needed power and tight supplies could lead to higher prices, bottlenecks, and outages," Bank of America said in a recent note.
Advertisement
Some eye-opening stats about the US power grid cited by Bank of America include:
- "The US grid produces 1,250 gigawatts (GW) of electricity from 9,200 generating units. Sometimes called 'the world's largest machine,' the American power grid has 600,000 miles of transmission lines, enough to wrap around the Earth 24 times. The average age of transformers, transmission lines, and other grid equipment is 40-50 years old."
- "Demand is rising for the first time in a decade. Over the past ten years, power demand rose just 0.4% per year. Over the next decade, the growth rate is expected to be 2.1% to 2.8%. Expected future demand of 70 GW by 2030 is like adding another state of Michigan to the grid every year."
- "Supply is tight and hard to add. No major utility projects are expected before 2026, and 160GW of fossil fuel supply has been shut in the past decade. Regulatory, permitting, and political obstacles often thwart new energy and mining efforts. Our colleagues expect only 55-60GW of capacity to be added in the near future."
- "Wind and solar struggle to make up the difference. They run only 24-40% of the time, producing much less than 'nameplate' capacity figures would suggest. Adding batteries brings extra strain: battery storage is 141 times more expensive than liquefied natural gas, and every KWh of battery storage required 50KWh of energy to create it."
Baird managing director and tech strategist Ted Mortonson told Business Insider last month just how big of a problem the power demand of AI is.
"The cost of that Gen AI architecture is freaking out of control. Oracle on their conference call basically said they are now constructing 70 megawatt data centers, going to 200 megawatts. That's the size of a city. So, they're so power hungry," Mortonson said. Oracle announced in its earnings call in March that it would invest $10 billion to expand data center capacity in order to service the huge demand for generative AI.
Advertisement
Amazon woke up to this realization earlier this year, evidenced by its decision to buy a nuclear power plant in Pennsylvania for $650 million.
A recent report from The Wall Street Journal said that Amazon's cloud unit is nearing a deal with Constellation Energy for electricity that would be directly supplied from a nuclear power plant on the East Coast.
This demand boom has led to a renaissance in utility stocks, with the sector surging 8% so far this year, and Goldman Sachs believes the gains can continue.
"While investor interest in the AI revolution theme is not new, we believe downstream investment opportunities in utilities, renewable generation and industrials whose investment and products will be needed to support this growth are underappreciated," Goldman Sachs said in a note earlier this year.
Advertisement
The bank highlighted four top utility stocks to buy, including Xcel Energy, NextEra Energy, Southern Co., and Sempra.
"US power demand likely to experience growth not seen in a generation. Not since the start of the century has US electricity demand grown 2.4% over an eight-year period, with US annual power generation over the last 20 years averaging less than 0.5% growth," Goldman Sachs said.