Good, Better, BESS

TLDR: BESS is not one thing. It is a five-layer stack, chemistry at the bottom, money at the top, and risk handed one layer up at every interface.

Everyone says they are "in BESS." Very few people say exactly where. Let's get into it.

Layer 1: Cells and chemistry (the physics layer)

Degradation, cycle life, and thermal behavior are set here, and they are application-specific, not datasheet-specific. A rated cycle life reflects a test protocol. Your project's depth of discharge (how much of the battery you actually use each cycle), C-rate (how hard you charge and discharge relative to capacity), and temperature range may look nothing like that test. The cell doesn't care what the spec sheet says.

The question this layer answers: how does the chemistry age under the actual duty cycle (specific profile of power usage and rest over time)?

Layer 2: Modules and packs (the mechanical layer)

Cooling architecture, cell layout, busbar design (the metal highways that carry current between cells), serviceability: all locked in long before commissioning. Thermal gradients become uneven aging. Design shortcuts become field issues, on a delay.

Layer 3: System integrators (the "We Sell BESS" layer)

Containers, inverters, BMS (battery management system), thermal, installation, often from different vendors. Coordination is the actual product.

The question: when something breaks on site with the interconnection date looming, who actually fixes it? Not who is named in the contract. Who shows up.

Layer 4: Power electronics and controls (the brains)

The inverter moves the energy. The EMS (energy management system) decides what the battery should do. Dispatch logic decides what it actually does. Inverter efficiency shifts with load, dispatch chases revenue, and the clean modeled profile becomes a series of daily judgment calls.

Layer 5: Owners and operators (the reality layer)

Whatever the four layers below didn't resolve lands here, for twenty years: degradation, downtime, augmentation (buying fresh capacity later to backfill what degraded), warranty enforcement.

The question, and it is the best diligence question in storage: what does year 7 look like? A confident recitation of datasheet numbers is a signal. People who have actually operated storage answer differently.

How storage projects actually fail

They drift. A slight degradation mismatch, a slight availability shortfall, slightly optimistic assumptions in the energy model. Each one sits inside the margin of uncertainty on its own. Together they compound, month by month, until the gap between the pro forma and reality is impossible to ignore. By then, the integration team is three projects away.

This is also why a chemically superior cell in a badly executed project will lose to a boring cell in a well-run one. The specs matter. They are not the story.

Good, Better, BESS

Good: it got built. The ribbon was cut. Everyone celebrated.

Better: it performs close to model for the first few years.

BESS: the chemistry fits the application, the design ages evenly, the controls match the actual dispatch environment, and someone still answers the phone in year 7.