(Oil & Gas 360) – As Artificial Intelligence (AI) expands to potentially revolutionize every aspect of human existence, the expected electrical demands required by the power devouring technology in its earliest stages present a formidable planning and financing challenge for utilities and other power plant developers.
The so-called electrical boom, the result of everything from growing economies and populations, cloud computing and cryptocurrency, and policies to electrify the economy wherever possible, has placed demands on utilities and the grid unforeseen just ten years ago.
America’s electrical grid, sometimes referred to as “the largest machine in the world” due to its complexity, importance, and sheer power, presents engineers, financiers, policymakers and consumers infrastructure challenges among the most profound in human history.
The closest analogy to the paradigm-shifting power dynamics might be the fast integration of the Internet into everyday life around the turn of the century, but the internet revolution included a cautionary tale no doubt recalled today by some of those preparing to power what is next with AI and electricity.
Back then, exponential growth driven by the internet from places such as Silicon Valley prompted a power plant construction frenzy where it was thought that natural gas-fired power plants would create a more efficient and thus more profitable financial model for electricity generation.
Construction was based on the financial assumption that the marginal cost of constructing and operating new power plants would be covered by long-term power prices. However, over-building capacity led to tough times for some utilities. For instance, California’s Calpine, one of the most aggressive capacity-adding utilities of the time, got overextended and went bankrupt, as did another market leader at the time, NRG energy.
Today, global terra wattage usage for data centers is projected to reach one thousand terawatt hours by the end of next year. Total terawatt consumption in the United States has ranged roughly between 4,000-4,200 terawatts annually since 2010.
The world-changing potential of AI means that utilities have to confront upgrading outdated infrastructure not designed for modern power needs, particularly as it relates to some of the fundamental components of the grid such as transformers and transmission constraints. The average age of a transformer in the United States is over thirty years old with a long lead time for new ones.
Strategic planning and capital allocation involves having to calculate the risks and benefits of AI for utilities where the technology can help operations be more efficient and resilient in the face of such calamities such as wildfires, cold snaps and hurricanes. At the same time, utilities also confront some renewable power mandates in various jurisdictions, making the integration of intermittent renewables into a system that increasingly depends on reliable 24/7 energy delivery a challenging prospect.
Other risks from integration of the relatively new technology include hacking risk, and growing debates over who should bear the cost of data center integration to the grid as consumers face the likelihood of higher costs for electricity as AI grows more powerful, and thus more power hungry.
By Jim Felton for oilandgas360.com
