A rack of tall cylinders along one wall of a battery container, plumbed into the ceiling.
They have never been fired at this site. They are being unbolted anyway.
Both plants on this end of Long Island went dark on the evening of Friday, October 2, 2026.
Not after an incident. There was no incident that week.
What is coming out of them is the thing installed to make them safe.
Which raises the question underneath it. Why does a battery hall need a special chemical at all?
Why water does not put out 1 burning cell
A lithium ion cell in thermal runaway is not an ordinary fire, and the difference decides everything that follows.
Most fires need oxygen from the room. A cell going into runaway breaks down its own metal oxide cathode and releases oxygen inside itself, so smothering it does nothing. It will keep reacting in a sealed box.
That leaves temperature as the only handle. Water does not extinguish the cell that is already gone. It cools the neighbors, which is the entire fight, because runaway spreads cell to cell by heat.
So these sites carry two systems at once. Sprinklers to hold the stack below the propagation threshold, which means hours of flow rather than minutes, and a clean agent that floods the space and strips heat out of the flame without soaking the electronics.
The agent of choice is a fluorinated ketone, a liquid that flashes to gas and leaves no residue on a switchgear cabinet.
Both halves of that answer end as liquid on the floor.
2 sites, a quarter mile from the wells
One plant sits on Cove Hollow Road in East Hampton, the other on North Shore Road in Montauk between Fort Pond and Fort Pond Bay.
This is sand. The South Fork sits on a shallow aquifer, and what reaches the ground here does not run off to a river, it goes down.
Public supply wells stand about a quarter mile from the East Hampton site.
The fluorinated ketone matters again at this point, for a reason that has nothing to do with fire. Mixed with water and heat it can break down into an ultra short chain member of the PFAS family.
Short chain is the hard case. The shorter the molecule, the faster it moves with groundwater and the worse it holds onto the carbon filters built to catch the long ones.
Which makes the aquifer the wrong neighbor.
30 hours of sprinklers and 2 wells pulled
The documented sequence starts in May of 2023, when the East Hampton plant caught fire.
Internal sprinklers ran for roughly 30 hours. How much water left the building is disputed, with one account putting it above a million gallons and a court filing cited at more than 2.2 million.
Two public supply wells about a quarter mile away later tested above state guidance for that short chain compound. Both were taken out of service and two further wells at the same site were switched to emergency use.
No concentration figure has been published. The reporting says above guidance and stops there.
State environmental staff put out field findings on the tenth of September, 2026, and a final report is still outstanding.
Their wording was careful. The compound may be linked to the suppression system.
The plant that never burned and 1 word from the state
Here is the part that makes this more than a cleanup story.
Montauk has never had a fire. It went offline alongside East Hampton because the issue is not an incident, it is the inventory sitting in the pipework, in the way that a coated pan or a lost well turns into a closed case.
The town board voted to authorize litigation on the first of October, the operator shut both plants the following evening, and the regional report sets out that week.
A week later the county extended the suspension until the agent is removed and a state approved remediation plan exists, and the local account carries that order.
None of it is a finding of fault. No cause has been established, the final report is pending, and the accounts disagree on the water volume and on the date of the vote.
What is settled is only the removal.
Nobody’s lights flickered
The county fire commission chairman offered the line that sticks.
When both plants went offline, he said, his lights did not even flicker. Two grid batteries left the system and the system did not notice.
That cuts both ways. It means the shutdown cost little, and it means the case for accepting the risk was never about keeping anyone’s lights on this week.
The useful question is narrower than the argument around it. Not whether storage belongs on a sand aquifer, but whether a hall full of cells can be protected by something that is not a fluorinated liquid.
Because the protection is the part that leaked.
A battery fire is a heat problem, and the industry answered it with a chemistry problem.
Read the whole thing?
Get the week's signal, not the noise
Our sharpest reporting on energy, climate and nature — free, once a week.
