A battery safety standard certifies a design — once, on a representative sample, before anything ships. It is essential work. But the battery that catches fire two years later is a specific, aged, possibly damaged unit that no test ever watched. Closing that gap means turning written standards into checks that run on real batteries, every day.
What a safety standard actually does
A safety standard is, at heart, a shared definition of "tested enough." It specifies a set of procedures — overcharge, short circuit, crush, thermal shock, vibration — and a set of pass/fail criteria, so that a battery design can be evaluated consistently by a lab and certified for a given use.
The landscape is broad and mostly well-matched to purpose. UN 38.3 covers the transport safety testing required before lithium batteries can be shipped. IEC 62619 addresses the safety of industrial secondary cells and batteries, while UL 1973 covers batteries for stationary and similar applications. UL 9540A is a test method specifically for characterising fire propagation in energy-storage systems. For electric vehicles, UL 2580 and the IEC 62660 series set testing expectations, and IEC 62133 covers portable sealed cells. Each one exists so that a buyer, a regulator or an insurer can trust a design without re-running the experiments themselves.
The gap between a certificate and a battery
Almost all of these are type tests: they are performed on a small number of representative samples, at a single point in time, before the product is deployed. That design produces a certificate. The certificate says the design, built and operated as intended, met the criteria on the day it was tested.
What a type test cannot do is watch the individual battery that later ships to a customer. It cannot see that unit age through hundreds of cycles, sit through a heatwave, take a knock during installation, or carry a subtle manufacturing outlier that the sampled units did not. The standard certifies that a design can be safe. It does not — and was never meant to — confirm that this battery, in this state, still is.
A certificate proves a design passed a test. It does not prove the battery in front of you would pass that test today.
Failures happen to units that passed
This is why so many real-world incidents involve products that were properly certified. The conditions that actually drive failure — gradual capacity fade and impedance growth, mechanical abuse, contamination, an unusually hot or cold environment, a cell that drifts out of balance with its neighbours — overwhelmingly emerge after the test, in service. Certification is a snapshot; safety is a moving target.
None of this is a criticism of standards. They set a floor, and that floor matters. The point is narrower: certification and field safety are two different questions, and a certificate alone answers only the first.
Living compliance: standards as runtime checks
The constructive response is to treat a standard not only as a hurdle cleared at launch, but as a source of intent that can be enforced continuously. A standard cares about a recognisable set of physical limits: temperature bounds, voltage windows, isolation and insulation resistance, cell-to-cell balance, resistance to thermal propagation. Those are not abstractions — they are measurable quantities.
That makes them monitorable. With a genomic baseline for each cell — its own structure, composition and electrochemical signature captured early in life — those same limits can be checked live, against the unit's individual reference rather than a generic one. We call this living compliance: the written requirement becomes a runtime rule, evaluated against the real battery as it operates.
What it gives operators
Translating standards into continuous checks changes what an operator can show, and when they can act:
- Continuous evidence — an ongoing record that each battery is still within limits, rather than a one-time certificate that ages the moment it is issued.
- Per-unit limits — thresholds anchored to each cell's own genomic baseline, instead of a single fleet-wide number that is too loose for some units and too strict for others.
- An audit trail — a defensible, queryable history for regulators, insurers and incident investigators, replacing reconstruction after the fact.
- Early intervention — a drift detected and acted on before a limit is breached, rather than a post-mortem after one is.
Standards set the bar; data proves you clear it
Safety standards do exactly what they were designed to do: they establish a credible bar and a common language for meeting it. What they cannot do alone is follow every battery through the years of service where its real risk is decided.
A genomic data layer is how that follow-through becomes practical. It does not replace certification — it extends its intent into daily operation, turning the requirements of a standard into checks that run against each cell continuously. Standards set the bar. A genomic data layer is how you prove that every battery is still clearing it, every day.