Can video intelligence help prevent battery fires? 

On the night of January 11, 2026, a pallet of electric scooter batteries in a logistics warehouse in Tarnowo Podgórne, Poland, entered thermal runaway. A grey-white vapour cloud appeared. Within minutes, the entire 11,000-square-metre facility was engulfed – over 80 fire departments deployed, losses exceeding €100 million, the warehouse partially demolished. 

By the time the fire alarm activated, it was already too late. 

It is a story becoming familiar across the world – from people’s homes, to warehouses, to dedicated Battery Energy Storage Systems holding gigawatt-hours of renewable power for national grids. Lithium-ion batteries are everywhere. The technology monitoring them has not kept pace. 

Battery fires don’t start with fire 

This is the critical point that most fire safety systems are not built around. 

Lithium-ion thermal runaway – the chain reaction that causes battery cells to overheat and ignite, produces a distinct sequence of visible signals before it becomes a fire. Swelling. Off-gassing. A vapour or smoke cloud that precedes ignition by minutes. In the Tarnowo Podgórne incident, these signs appeared on CCTV footage well before the blaze took hold. 

Traditional fire detection systems are not designed to detect any of this. They are designed to detect fire. By the time they activate, the window for early intervention has already closed. 

This is not a flaw in the alarm system. It is a fundamental limitation of any technology that waits for fire to begin before responding to it. 

What video intelligence would have changed 

The surveillance footage from Tarnowo Podgórne tells a clear story. The pallet sat in frame. The off-gassing began. The vapour cloud grew. Then ignition. Then, within minutes, total loss. 

Every frame of what was about to happen was recorded. Nothing acted on it. 

This is the gap that video intelligence exists to close – not better cameras, but the ability to understand in real time what the footage means, and alert the right person before the situation becomes irreversible. 

After the incident, the footage was run through our smoke and fire detection system. Sentry identified the incident at the moment the battery began to off-gas and the first smoke appeared – well before any flames became visible. 

Proactive fire detection 

Traditional fire alarms are reactive by design – they respond to fire already present, not the conditions that precede it. Battery fires demand something different: the ability to detect the visual precursors to ignition that appear in the minutes before thermal runaway becomes a blaze. 

Sentry monitors video feeds continuously, identifying early indicators of thermal events such as vapour clouds, smoke wisps, off-gassing – without relying on predefined rules. Because it learns what normal looks like at each site, it detects what does not belong and alerts operators immediately, with the context needed to act. 

In Tarnowo Podgórne, that would have meant an alert at the first sign of off-gassing – minutes before the fire alarm triggered, and before a salvageable situation became a catastrophic one. 

A growing risk across logistics, manufacturing and beyond 

The Tarnowo Podgórne fire is not an isolated incident. As the electrification of logistics, retail, and industrial operations accelerates – more e-scooters, e-bikes, power tools, and industrial battery systems moving through warehouses, distribution centres, and manufacturing facilities every day –  the volume of lithium-ion batteries in operational environments is growing rapidly. 

So is the risk. Lithium-ion thermal runaway can be triggered by physical damage, manufacturing defects, improper storage, or temperature extremes – many of which are invisible until the off-gassing begins. Standard fire safety infrastructure was not designed with this risk in mind. 

The risk extends well beyond logistics. Battery Energy Storage Systems – large-scale lithium-ion installations used to store renewable energy for grids, commercial buildings, and industrial facilities – represent one of the fastest-growing and highest-risk environments for thermal runaway.  As global BESS capacity expands rapidly alongside renewable energy deployment, the gap between what cameras can see and what traditional detection systems act on becomes an increasingly costly one to leave open. 

The question every operations leader should be asking 

Your cameras are already recording what happens in your facility around the clock. The question is whether your infrastructure has the intelligence to recognise the early signs of a battery fire – and act on them before they become something far worse. 

In Tarnowo Podgórne, the footage existed. The warning signs were there. The intelligence to interpret them in real time was not. 

That is the problem we exist to solve. 

Video intelligence for fire, smoke and radiance detection 

Video intelligence adds a real-time safety layer on top of your existing cameras, turning every frame into a chance to spot risk before it becomes an emergency. 

IntelexVision’s Sentry platform delivers real-time video intelligence that enables early detection of threats and proactive action across energy, utilities, and emergency services by reducing false alarms from wildlife, weather, and lighting, supporting limited operator teams over large zones, and accelerating the detection of safety breaches and operational anomalies.  

With optical smoke, flame, and heat signature analysis, event classification (for example, open flame versus environmental haze), and integration into site-wide alarm and evacuation protocols, Sentry enables faster emergency response, reduces downtime and equipment loss, and provides early visual detection when it matters most. 

Key takeaways 

  • Lithium-ion battery fires produce visible precursor signals: vapour clouds, off-gassing, smoke – before ignition occurs 
  • Traditional fire alarms are reactive: they detect fire, not the conditions that precede it 
  • The latest video intelligence – like Sentry – can provide earlier warning for battery-based fires as shown in the Tarnowo Podgórne  
  • Early detection of battery fires requires continuous, intelligent monitoring of video – not sensor thresholds 
  • As lithium-ion battery volumes in operational environments grow, the window between first visible sign and uncontrollable fire will remain as narrow as ever 

The Sentry video intelligence platform provides AI-powered smoke, fire, and hazard detection for warehouses, logistics hubs, manufacturing facilities, and critical infrastructure.

To see how video intelligence can protect your operation, book a demo today. 

Related articles