It's becoming a consensus that "AI ultimately relies on electricity and energy." $Microsoft(MSFT)$ , a beneficiary of this AI wave, recently signed the largest-ever single corporate power purchase agreement in history.
Microsoft inked a five-year deal with $BROOKFIELD ASSET MANAGEMENT LTD(BAM)$ and its affiliate, Brookfield Renewable, worth over $10 billion. This is to accelerate the expansion of renewable energy capacity and develop over 10.5 gigawatts (GW) of new renewable energy.
$10 billion? That's the same amount Microsoft invested in OpenAI earlier, showing the strategic significance of energy for the company's future development.
Microsoft isn't the only one investing heavily in both AI and energy. OpenAI's founder, Altman, personally invested $375 million in Helion Energy, a private fusion energy company in the US.
Altman once said: "Future AI needs an energy breakthrough because AI will consume far more electricity than people expect." He predicts that nuclear fusion or cheaper solar energy and storage are the future directions for AI.
Undoubtedly, the rapid development and application of AI are driving the power demand of data centers to grow exponentially. Data centers and supercomputing centers are like "electricity-swallowing monsters."
Thomas Thornton, a stock strategist at Bank of America Merrill Lynch, predicted:
The power consumption of AI workloads will grow at a compound annual growth rate of 25% to 33% in the next few years.
According to forecasts, by 2030, global data centers' power demand could reach 126-152 GW, adding approximately 250 terawatt-hours (TWh) of power demand during this period, equivalent to 8% of the total US power demand in 2030.
Jensen Huang, the "GPU guru," suggested in a speech that $NVIDIA Corp(NVDA)$ helps improve computational efficiency and reduce energy consumption. If computational speed doesn't accelerate, "we might need 14 different planets, 3 different galaxies, and 4 suns to power all this."
When Musk talked about the future development of the AI industry at the Bosch ConnectedWorld 2024 conference, he was even more explicit. More than a year ago, it was a shortage of chips, and after that, it was a shortage of electricity. "Next year, electricity won't be able to meet all the chip demands."
On the other hand, energy companies are increasingly using AI to enhance their operational efficiency and sustainability. AI is already transforming key business areas: from reducing carbon emissions, preventing cyber attacks, to predicting mechanical failures.
Energy + AI are increasingly intertwined in a multifaceted relationship. The next beneficiaries of AI won't just be Nvidia, the providers of computing power, but also the users of AI, including traditional energy industries.
Comments