By Laura Ross
In February 2021, parts of the U.S. were engulfed by a series of severe winter storms. This resulted in a major electricity generation failure in Texas, where around 4.5 million homes lost power, 57 people died, and the cost of property damage surpassed $195 billion.
The electricity failures, largely attributed to the freezing of the state’s natural gas pipelines, highlighted the problems with the U.S. power grid. During the winter of 2021, production at Texas’ coal and nuclear plants also dropped, and similar events occurred in other states, including Kansas.
In the years since, various energy experts, politicians, and pundits have highlighted the need for major spending on grid-related infrastructure upgrades. What role could artificial intelligence (AI) play in supporting this transformation?
The U.S. Embraces Renewables
In the past six years, power outages have more than doubled and regulators are warning that these rolling outages could become far more widespread. Last summer, for example, the regional grid in the Midwest was short of the amount of energy needed to power 3.7 million homes, while wholesale electricity prices skyrocketed to $5,000 per megawatt-hour.
Plans to rapidly retire fossil fuel plants, in a bid to move toward low-carbon electricity production, coupled with the predicted rise of electric vehicles (EVs), will further destabilize the nation’s existing power systems.
In the United States, the portion of electricity obtained from solar and wind almost quadrupled between 2011 and 2020. McKinsey estimates that by 2026, global renewable-electricity capacity will be up 80% from 2020 levels. Meanwhile, the Biden Administration has set goals for EVs to comprise 50% of all new passenger vehicle sales by 2030, a zero-carbon electricity grid by 2035, and a net-zero carbon economy by 2050.
The decentralized nature of renewable sources, including microgrids, private solar panels, and wind farms, adds a level of complexity to the grid that existing infrastructure cannot withstand. Industry experts believe that a complete overhaul of the U.S.’s antiquated power grid requires an investment of more than $2 trillion, and steps have been taken in this direction. The U.S. Department of Energy, for example, has launched the Building a Better Grid Initiative to upgrade the national grid network as part of President Biden’s infrastructure law.
The use of AI could support a more cost-effective transformation and successfully move the nation toward a cleaner and greener future.
How Can AI Transform the U.S. Power Grid?
As the World Economic Forum highlights, the shift to an increasingly electric world means decentralized, renewable sources will produce more energy.
AI-powered software could seamlessly accommodate the transmission of electricity from these disparate sources to the national grid. In addition, AI algorithms could power energy storage systems (ESS) that offload renewable energy from storage as and when it is needed. Large EV fleets, for example, could be charged without overstressing the grid and software could forecast when and where to offload and store power.
Several companies are already experimenting with these AI-powered technologies. Nvidia, a leading chipmaker, has partnered with energy software company Utilidata to develop a smart grid chip. These chips would enable meters to collect and process data on power needs in real time so utility companies can direct their resources more efficiently.
As well as enabling more accurate forecasting, AI-powered systems can also monitor equipment for potential outages or failures and respond to disruption in seconds, rather than days. This will help prevent the ongoing, and highly destructive, power outages many parts of the U.S. have endured in recent years.
What’s Next for the U.S. Power Grid?
Rather than investing in infrastructure-based solutions, which come at an enormous cost and require many years of planning and construction, policymakers, utility companies, and governments can put more faith in the power of AI.
For lawmakers, that could mean incentivizing communities to generate their own electricity, efficiently managed via AI software. Utility companies, meanwhile, must decide whether to pivot their service offering and establish themselves as software companies, or to partner with existing businesses that have the means and the know-how to develop AI solutions.
