The Pacific Gas and Electric Company (PG&E) is an American investor-owned utility (IOU) with publicly-traded stock. The company is headquartered in the Pacific Gas & Electric Building, in San Francisco, California, United States. PG&E provides natural gas and electricity to most of the northern two-thirds of California, from Bakersfield and to the north side of the County of Santa Barbara, to near the Oregon State Line and Nevada and Arizona State Line, which represents 5.2 million households.Overseen by the California Public Utilities Commission; PG&E is the leading subsidiary of the holding company PG&E Corporation, which has a market capitalization of $3.242 billion as of January 16, 2019. It was founded by George H. Roe after California's Gold Rush and by 1984 was the United States' "largest electric utility business". PG&E is one of six regulated, investor-owned utilities (IOUs) in California; the other five are PacifiCorp, Southern California Edison, San Diego Gas & Electric, Bear Valley Electric, and Liberty Utilities.In 2018 and 2019, the company received widespread media attention when critics assigned the company primary blame for two separate devastating wildfires in California, leading to a loss in federal court and a formal finding of liability on their part.. On January 14, 2019, PG&E announced that it was filing for Chapter 11 bankruptcy in response to the financial challenges associated with the catastrophic wildfires that it was liable for in Northern California that occurred in 2017 and 2018. The company hoped to come out of bankruptcy by June 30, 2020, and was successful on Saturday, June 20, 2020, when U.S. Bankruptcy Judge Dennis Montali issued the final approval of the plan for PG&E to exit bankruptcy.
The accelerated growth of AI data centers has introduced a new class of large and highly volatile loads to the power grid. Unlike traditional data centers, which are dominated by relatively steady and predictable IT loads, AI data centers exhibit rapid power transients, leading to increased harmonic injection and a higher likelihood of power system instability. This work examines the characteristics of AI data centers’ workload and evaluates methods to mitigate them. Specifically, artificial load modulation, GPU ramp-rate tuning, power capping, and the integration of multi-level energy storage to provide supplementary power during dynamic load fluctuations are reviewed. Several energy storage methods are discussed to reduce transients due to power oscillations at AI data center interconnections and to improve the electric grid’s power quality. New power delivery architecture and system-level coordination are also identified as promising approaches to alleviating the power quality challenges.
Electricity price signals in modern power systems exhibit complex dependence structures that render forecasting inherently challenging. Our analysis of real-world pricing signals from the California Independent System Operator (CAISO) reveals complex temporal group effects, whereby the influence of explanatory variables on electricity prices persists across consecutive blocks of time due to underlying economic and operational drivers. In response, we propose a multivariate statistical method based on a Group Lasso formulation to forecast the vector of day-ahead electricity prices, by leveraging multi-feature temporal group effects. Our approach is evaluated on two full years of electricity prices from CAISO, demonstrating considerable improvements in point and probabilistic forecast metrics compared to a wide array of statistical and deep learning methods. Theoretical and empirical analyses confirm the effectiveness of the proposed approach in modeling realistic group effects, maintaining both interpretability and low computational complexity. When retrospectively evaluated on test data from a recent international electricity price forecasting challenge, the proposed method ranked in second place, despite having access to significantly less information than competing approaches. Finally, the proposed method is independently validated against two operational electricity price forecasting systems in CAISO, demonstrating competitive predictive performance and practical relevance.
Driving range and charging have been widely studied to understand battery electric vehicle (BEV) adoption patterns, yet evidence remains limited on how they jointly relate to BEV utilization. This study examines the relationship between driving range, charging infrastructure, and BEV mileage, using data from the 2019 and 2024 waves of the California Vehicle Survey. We estimate multivariate linear regression models to assess how self-reported availability of home and workplace charging options—by charger level (Level 1, Level 2, and DCDC chargers)—affects annual BEV mileage, considering vehicle driving range. Our key results indicate that range and charging options near the work location act as substitutes influencing the demand for driving. Availability of charging opportunities at workplace is associated with higher mileage primarily among shorter- and mid-range BEVs, and this association diminishes as driving range increases. In contrast, we find no significant interaction between range and charging options at home. These results highlight the importance of jointly considering vehicle range and charging infrastructure in planning discussions to optimize BEV adoption and usage patterns.
