The asset you couldn't control

When Vistra converted an old gas-plant site into one of the world's largest lithium-ion battery facilities, the plan was hailed as a clean-energy milestone — the kind of asset a strategy is built around. Phase 1 alone held enough storage to power hundreds of thousands of homes. Then, on a January afternoon, it caught fire, entered thermal runaway, and burned to a total loss. For the board, the meaningful question is not which cell failed or why — the cause is still being established. It is the prior decision: to deploy a novel, high-energy hazard at world-leading scale before the means to contain its worst case were in place.

The clearest image of the event is operational, but its meaning is governance. Firefighters who had toured and inspected the building could do nothing but monitor it as it burned, because a battery in full thermal runaway cannot be extinguished. That helplessness was not theirs to fix on the day. It was the visible expression of a decision made long before: the organisation had deployed this technology faster than it had built the ability to control its failure. When that is true, the worst day is not managed — it is merely survived, if you are fortunate.

With an established hazard, the means of control are mature and known. With a genuinely novel technology, they are not: the operator, the regulator, the insurer and the emergency services are all climbing the same learning curve at the same time. Meanwhile the commercial case — first-mover advantage, policy targets, capital already committed — pushes hard to deploy now and at scale. That is the structural tension, and only one body sits above the commercial urgency with a mandate to hold the line: the board. Deployment pace is therefore not an operational detail. It is a governance decision about how far ahead of demonstrated control the organisation is willing to run.

Moss Landing also shows how the decision usually gets made: on assurance rather than analysis. The batteries were housed in a repurposed 1950s building described in the project application as robust and non-combustible, with large numbers of high-energy racks in one open space, without the compartmentalisation newer modular designs use. Reporting describes the approving body being given assurances, not a worst-case analysis. One commissioner who had voted to approve later said, publicly, that she had been wrong. "Non-combustible" was a description, not a hazard study — and the gap between the two is exactly where a board must refuse to be comforted.

The asset built to anchor the strategy became its opposite. Phase 1 was a complete loss of building and grid assets. Insured losses were estimated by Marsh McLennan at around 180 million dollars — among the largest battery-related property claims in U.S. history. More than 1,200 residents were evacuated and a major highway and local businesses closed. The smoke plume carried metals onto surrounding wetlands, with environmental testing and litigation following. And the regulatory response moved against the technology itself: the county advanced toward a moratorium on new battery projects, and a new state law now requires stronger containment, monitoring and engagement before such facilities open. A board that booked this asset at its build value was carrying a liability it had never priced.

The board's task with any novel high-consequence technology is to keep deployment pace inside the boundary of demonstrated control: a chemistry chosen for inherent safety, a firebreak through separation and containment, early detection protected, suppression matched to the failure mode, and a worst case engineered to be survivable. The standards now most associated with battery storage — NFPA 855 and UL 9540A — and the modular designs newer sites use, matured after Moss Landing was conceived. Deploying at this scale ahead of them was not bad luck. It was a choice about pace, and pace is the board's to set.

There is also a duty to the people who will manage the worst day. At Moss Landing the responders knew the building, but the standard playbook — water — did not apply to the hazard, and the surrounding community met the consequences without warning. The new state law's requirement for engagement with local fire officials before facilities open is, in governance terms, an admission that this should have been a precondition all along. A board should treat one question as part of any sanction: have the people who will manage the worst day already met this hazard — before the worst day?

"An organisation can be world-leading at deploying a technology and a novice at containing its failure. The board's task is to close that distance before scaling, not after."

You will not understand every technology your organisation adopts. You can refuse to let its scale outrun your control of its failure. Three questions hold that line.

1. Worst-case analysis, not vendor assurance, before sanction — Require a documented worst-case failure analysis — magnitude, containability, plume, financial and community exposure — before sanctioning a novel high-consequence deployment. Red flag: the case rests on assurances that something is "safe" or "non-combustible" rather than an analysis of what happens when it fails.
2. Deployment pace inside demonstrated control — Confirm the means of containment — firebreak through separation, protected early detection, suppression matched to the failure mode — exist and are proven at the scale being deployed. Red flag: scale is set by the commercial case and capital committed, while the control measures are still maturing or unproven at that size.
3. Responders and community briefed, worst case survivable, before commissioning — Confirm that local emergency services know the technology and that the worst case has been engineered to be survivable for people and neighbours — engagement now required by law where Moss Landing sits. Red flag: the first time responders and the community meet the hazard is on the day it fails.

None of these requires the board to understand battery chemistry, ventilation rates or fire codes. They require one refusal: not to let commercial pace carry the organisation past the point where it can contain what it has built. That refusal is the whole of the discipline this series has traced — across an assurance system, a counted control, a clean permit, a downgraded space, and now a scaled technology. The form changes; the boundary does not.

Moss Landing is this series' broadest warning. The latent condition was not a faulty part; it was a pace — deploying faster than the discipline to contain. As every business races to adopt new high-energy and high-consequence technologies, that is the board's enduring boundary: never sanction scale you cannot yet contain, and never accept assurance where the worst case demands analysis. A technology you have deployed faster than you can control is not yet an asset. It is an unpriced liability with a commissioning date — and, on the wrong day, a fire your people can only watch.

"Before you sanction scale, ask what the worst case costs and who has already rehearsed it. If the answer is "we are not sure," you have your answer."