The Utility Briefing

Nearly three months after a fire at a Mission District substation knocked out power to 137,000 San Francisco residents in December 2025, city officials are still pressing for answers. The outage — one of the most disruptive in recent San Francisco history — has triggered a formal push for investigation, with municipal leaders demanding accountability from Pacific Gas & Electric over what caused the failure and, more pointedly, why it happened at a location with a documented history of similar incidents.

That last part is the real story here.

A single substation fire is a serious operational event. A recurring fire at the same substation is an institutional failure. The distinction matters enormously, not just for PG&E's regulators and ratepayers, but for the broader utility industry as it grapples with aging infrastructure, escalating climate risk, and the accelerating demands of electrification.

The Problem Isn't Just the Fire

When a critical piece of infrastructure fails repeatedly at the same location, it signals one of a few uncomfortable truths: the root cause was never properly identified the first time, corrective actions were insufficient or poorly implemented, or the asset was deprioritized in capital planning despite known risk indicators. In any of these scenarios, the failure is not simply technical — it is procedural.

A substation that has caught fire before is not an unknown risk. It is a documented risk. The question regulators should be asking is what data PG&E had, when they had it, and what decisions were made — or deferred — as a result.

This is precisely where process intelligence and operational visibility become non-negotiable. Utilities generate enormous volumes of maintenance records, inspection reports, work order histories, and equipment performance data. The challenge is rarely a lack of data — it is whether that data is being translated into action before an asset fails catastrophically, not after.

What Better Operational Intelligence Looks Like

Forward-thinking utilities are increasingly deploying continuous monitoring frameworks that flag deteriorating assets based on historical incident patterns, maintenance cycle deviations, and real-time sensor inputs. When a substation has a prior fire event in its record, that history should automatically elevate its risk classification and trigger more aggressive inspection and remediation timelines. It should not take a second blackout affecting six figures of customers to prompt a review.

The San Francisco case also raises a service quality dimension that utilities can no longer afford to treat as secondary. Prolonged outages disproportionately affect vulnerable populations — those who depend on medical equipment, lack resources to relocate, or live in older housing stock without backup capacity. The 137,000 figure is not just a statistic; it represents a service obligation that was not met, and a trust deficit that compounds with every delayed explanation.

A Regulatory Reckoning in Progress

San Francisco's push for investigation may ultimately produce fines, mandated audits, or accelerated infrastructure spending. But the more durable outcome — for PG&E and the industry — would be a genuine shift toward predictive, data-driven asset management that treats prior failure as a forward-looking risk signal rather than a closed case.

The utilities that will lead in the next decade are not those that respond best to crises. They are those that build the operational intelligence to see crises coming — and make the organizational commitment to act before 137,000 people lose power on a December night.

When a single frustrated customer can launch a petition and attract nearly 60,000 signatures within weeks, something has gone wrong well before the first bill arrived in a mailbox. That's the situation Duke Energy now faces in North Carolina, where customers reporting utility bill spikes of $150 or more during the January-to-mid-February cold stretch have turned a seasonal billing cycle into a full-blown public trust crisis.

Duke's explanation — that extreme cold weather drove higher consumption — is almost certainly accurate. The physics are unambiguous. But accuracy alone doesn't rebuild confidence, and that distinction matters enormously for the industry at large.

The Real Problem Isn't the Bill. It's the Surprise.

Customers aren't just angry about paying more. They're angry because they didn't see it coming. In an era when consumers receive real-time alerts for credit card transactions, flight delays, and grocery delivery windows, receiving a bill that's double or triple the norm — with no prior warning — feels like a system failure, not a weather event.

This is fundamentally a process visibility problem. The data to anticipate these spikes existed in real time: smart meter readings, degree-day accumulation models, consumption trend anomalies. The question utilities must ask themselves is not why did bills go up, but why didn't customers know they were going to?

"The data to anticipate these spikes existed in real time. The question is why customers weren't warned before the bill arrived."

Operational Intelligence Has to Flow Both Ways

Modern utility operations are increasingly data-rich. Advanced metering infrastructure gives utilities near-continuous visibility into consumption patterns at the household level. Yet that intelligence too often flows in one direction — inward, toward billing systems — rather than outward, toward the customers whose behavior and budgets it directly affects.

A proactive mid-cycle alert — "Based on your current usage, your bill this month is tracking 40% higher than last year. Here's why, and here's what you can do" — doesn't eliminate the higher bill, but it transforms the customer experience entirely. It converts a shock into an informed expectation. That's not a small thing. That's the difference between 60,000 petition signatures and 60,000 customers who feel respected.

What the Industry Should Take From This

Duke Energy is not uniquely culpable here. Unexpected bill spikes during extreme weather events are an industry-wide vulnerability, and winter 2024 has exposed it across multiple service territories. What makes the Duke situation particularly instructive is the speed and scale of the customer response — a signal that tolerance for opacity in utility billing is eroding fast.

Regulators in North Carolina and beyond will be watching. If utilities don't close the loop between the operational data they already possess and the customer communication infrastructure they deploy, they invite exactly this kind of intervention — audits, legislative scrutiny, and mandatory refund proceedings that are costly, reputationally damaging, and largely avoidable.

The utility sector is undergoing a once-in-a-generation transformation in grid technology and data capability. The industry's ability to retain public trust — and its regulatory compact — increasingly depends on putting that intelligence to work for customers, not just for operations. Extreme weather isn't going away. The standard for customer communication can't stay where it is.