A technology adoption curve with no safety curve beneath it.
Microsoft has announced plans to deploy up to 1,000 EV charging stations across its Redmond, Washington campus. Genentech operates a large and growing EV fleet. Amazon, Google, Apple, and hundreds of smaller corporate campuses are following the same trajectory โ driven by sustainability commitments, employee demand, and fleet electrification mandates.
None of this is wrong. But nearly every one of these organizations has overlooked a critical question: what happens when one of these vehicles catches fire in your parking structure at 2 a.m.?
Lithium-ion battery fires are not ordinary vehicle fires. They are categorically different in how they start, how they behave, how long they last, and what it takes to suppress them. The gap between what most facilities teams know about EV fire risk and what they need to know is significant โ and it's growing as charging infrastructure scales faster than safety planning.
The safety gaps on most campuses right now:
- No pre-incident plans (PIPs) filed with local fire departments โ responding units arrive with no information about EVSE locations, utility shutoffs, or structural access points
- EVSE installed adjacent to structural columns or egress paths without fire separation review under NFPA 1 or NFPA 88A (parking structures)
- Sprinkler systems designed and installed for gasoline-fueled vehicle fires โ overhead suppression provides limited effectiveness against floor-level lithium-ion battery thermal runaway
- No drainage capacity planning for high-volume water application โ 3,000+ gallons per vehicle produces runoff that can overwhelm standard floor drains and compromise adjacent structures
- No EVSE telemetry monitoring for anomalous charging events โ temperature spikes and abnormal current draw are early warning signals that most platforms expose but no one is watching
- Facility security and EHS teams untrained for initial EV fire response โ the first 5 minutes before fire department arrival are often the most consequential
- No defined utility shutoff procedures or documentation for responding fire units โ finding the right breaker in a large campus structure during an active incident costs critical time
- No fleet-side thermal event history analysis โ some EVs arrive on your property already stressed from prior fast-charging cycles or minor impact damage that increases failure risk
Thermal runaway doesn't behave like a car fire.
When a traditional gasoline vehicle ignites, the fire department's response is well-understood: suppress the fuel source, cool adjacent surfaces, ventilate. The fire has a defined behavior profile and a well-established suppression approach. Fire departments have decades of training for it.
Lithium-ion battery thermal runaway is different in nearly every dimension. Once a battery cell enters thermal runaway โ triggered by overcharging, physical damage, manufacturing defect, or heat exposure โ a chain reaction begins that cannot be stopped. It can only be managed.
The re-ignition problem
An EV battery pack that appears extinguished is not safe. Cells deep within the pack can remain in a degraded but unstable state for 24 to 72 hours after initial suppression. Without continued monitoring and cooling, re-ignition is a documented, predictable risk โ not a theoretical one. Most parking structures have no protocol for this, and most corporate security teams are not equipped to manage a multi-day monitoring operation for a single vehicle.
The suppression math
Standard suppression for a passenger vehicle fire requires approximately 500 gallons of water. A single EV battery fire requires an estimated 3,000 to 8,000 gallons โ depending on battery size, pack configuration, and how early suppression begins. In a multi-story parking structure with standard NFPA 13 sprinkler coverage designed for conventional vehicles, the available water supply may be insufficient. Drainage systems designed for normal fire suppression may not handle the volume or the pH of lithium-ion battery runoff, which introduces environmental hazard considerations on top of the fire risk.
The cascade risk
In high-density EVSE deployments โ rows of Level 2 chargers packed tightly into a parking structure โ the heat output from one vehicle in thermal runaway can propagate to adjacent vehicles. The closer the spacing, the higher the cascade risk. This is especially relevant for companies planning large-scale deployments where space efficiency drives charger layout decisions without fire separation analysis.
"The fire service has not yet caught up to the scale of EV deployment in corporate and residential settings. Most departments are still developing their response protocols. That means the burden of preparation falls on the facilities teams who own these structures โ before an incident, not after one."
What NFPA says โ and what most teams haven't read.
Several NFPA standards directly address EV charging infrastructure and battery fire risk. Most corporate facilities teams are unaware of the specific provisions that apply to their buildings.
Relevant NFPA Standards for EVSE Deployments
- NFPA 1 โ Fire Code: covers EVSE installation requirements, including placement, clearances, and emergency disconnects
- NFPA 13 โ Standard for the Installation of Sprinkler Systems: the basis for most corporate parking structure suppression โ not written with EV thermal runaway in mind
- NFPA 72 โ National Fire Alarm and Signaling Code: governs detection systems that may need updating to address EV-specific detection requirements
- NFPA 88A โ Standard for Parking Structures: covers construction, protection, and special hazard requirements for parking facilities
- NFPA 855 โ Standard for the Installation of Stationary Energy Storage Systems: the primary standard for EV charging infrastructure fire protection โ most recently revised to address large-scale EVSE deployments
Compliance with these standards is not optional โ but more importantly, compliance alone is not sufficient. Standards represent minimum requirements established through a consensus process that inherently lags real-world deployment realities. A campus deploying 1,000 EV chargers in a multi-story structure should be asking not just "are we code-compliant?" but "are we actually prepared?"
AI doesn't extinguish fires. It helps prevent them.
The application of AI to EV fire risk is not about suppression automation โ it's about early warning, pattern recognition, and continuous monitoring that human teams can't sustain at scale.
EVSE telemetry monitoring
Most commercial EVSE platforms โ including ChargePoint, Blink, EVgo, and others โ expose real-time charging data via API: session duration, current draw, voltage, temperature, and charging anomalies. AI-driven monitoring systems can watch these data streams continuously, flagging abnormal charge curves or thermal signatures that indicate a vehicle or charger in a degraded state. A campus with 1,000 chargers cannot have a human watching all of them. AI can.
Risk modeling for EVSE deployment plans
Before a single charger is installed, AI tools can model fire propagation scenarios based on charger placement, structural configuration, suppression system coverage, and drainage capacity. Running these models during the design phase โ rather than during a post-incident review โ allows facilities teams to identify and address cascade risks, separation gaps, and suppression deficiencies before they're built in.
Fleet-level battery health analytics
For organizations operating large EV fleets, AI-powered telematics analysis can identify vehicles that show charge cycle degradation patterns associated with elevated thermal event risk โ enabling proactive fleet maintenance decisions before a battery reaches a dangerous state.
The organizations getting this right are doing five things.
Across facilities teams that have taken EV fire risk seriously, a consistent pattern of preparation has emerged. These organizations are not necessarily the largest or most sophisticated โ they're the ones that asked the right questions early.
What prepared organizations are doing
- Filing EV-specific pre-incident plans with their local fire department โ including EVSE locations, breaker maps, access routes, and water supply information
- Conducting a formal suppression system review to assess adequacy for EV thermal runaway scenarios โ not just code compliance
- Reviewing EVSE placement against NFPA 1 and NFPA 88A separation requirements before installation, not after
- Deploying EVSE telemetry monitoring with defined escalation thresholds for anomalous charging events
- Running tabletop exercises with facility security, EHS, and local fire department representatives โ so everyone knows what to do before they need to
We help facilities teams close the gap โ before an incident forces it.
Lometa AI brings a combination that doesn't otherwise exist in this space: active fire service credentials in high voltage vehicle firefighting, real-world EVSE deployment experience, and the operational AI consulting background to translate risk into actionable planning. Our EV & EVSE Risk practice covers everything from initial risk assessment through pre-incident plan development, suppression review, and AI monitoring strategy.
For companies actively expanding EV charging infrastructure โ like Microsoft's 1,000-charger Redmond campus buildout โ the most cost-effective moment to address these risks is before the infrastructure is installed, not after.