The Federal Energy Commission (FERC) on July 16 ordered artificial intelligence data centers to establish mandatory standards for maintaining the security of the power grid. This order applies to data centers with access to capacities exceeding one gigawatt and must be implemented by December 31, 2023.
New Challenges for the Power Grid
Reports indicate that by 2027, 40 percent of artificial intelligence data centers will face operational limitations due to a lack of electricity. The current power consumption of data centers in the United States is approximately 80 gigawatts, and it is projected to reach 150 gigawatts by 2028. This increase in power consumption places significant pressure on the power grid, prompting FERC to respond to these statistics.
Read more: Nexans Electra on the verge of a major leap in underwater cables
Impact of New Loads on the Grid
Tom Iford, a global operations expert at Oracle, emphasizes that the new loads from data centers could have a similar impact on the power grid as the loss of a large power plant. These heavy loads can be quickly removed from the grid, which is dangerous because operators must instantaneously match electricity production with consumption loads.
Aaron Nimala, global director of operational technologies at Oracle, also believes that data centers should be recognized as active participants in the power grid rather than ordinary customers. This change in load classification, according to FERC's directive, will become a federal scheduling matter.
Challenges of Wildfires and Risk Management
Wildfires have also exacerbated existing problems. In this regard, the Oregon Court of Appeals ruled in April 2023 in favor of Pacificorp, overturning a one billion dollar judgment. On the other hand, South Dakota has passed a law prohibiting strict liability claims in wildfire cases.
In California, Governor Gavin Newsom has proposed accelerating payments to wildfire victims. Iford points out that liability for wildfires places additional pressure on operators, as this can lead to bankruptcy and billion-dollar lawsuits.
Using Artificial Intelligence to Predict Hazards
Companies are striving to utilize artificial intelligence to mitigate risks arising from wildfires and new data center loads. Nimala states that by combining weather forecasts, LiDAR vegetation data, and historical outage patterns, companies can generate risk scores for their equipment. This action could lead to a 35 percent reduction in outage loads under severe weather conditions.
Key components of these systems have already been implemented, but whether companies are adopting them quickly enough remains an unanswered question. Data centers and wildfires are not waiting for answers.
Read more: Development of electric charging infrastructure at the Port of Vancouver in collaboration with KemPower and EVlution Charge · Meta and entry into the subscription world: a revolution in the economy of artificial intelligence
