The largest battery in the world isn't in a phone or an electric vehicle. It's sitting in a field in California, Australia, or Texas, wired into the power grid, and it can pump hundreds of megawatts into the network within milliseconds of being needed. Grid-scale battery storage — banks of lithium-ion cells packaged in shipping containers and connected to the utility grid — has grown from a curiosity to a fundamental piece of power system infrastructure in under a decade. By 2025, global grid-scale battery storage capacity reached roughly 267 gigawatts of power and 610 gigawatt-hours of energy , up from virtually nothing in 2015. The cost of storing a megawatt-hour of electricity dropped from around $1,500 in 2014 to below $117 by 2023. That trajectory has made batteries competitive with gas peakers for short-duration storage and unleashed a wave of project development that shows no sign of slowing. AI-generated image Inside a modern battery energy storage system (BESS) facility, lithium iron phosphate modules are stacked in racks and connected to inverters that interface with the utility grid. What Grid Batteries Actually Do The power grid has an unusual property: electricity must be generated at exactly the same moment it's consumed. Unlike a warehouse that can hold inventory for weeks, the grid has essentially zero inherent storage — or it did before batteries arrived. This balance requirement has historically been managed by spinning reserve: gas turbines and other flexible generators kept running at partial output, ready to ramp up within minutes when demand spikes or a generator trips offline. Batteries can do several things for the grid, and understanding the distinctions matters because different services require different system designs: • Frequency regulation: The grid runs at 60 Hz (in North America) or 50 Hz (most of the world). When supply and demand fall out of balance, frequency drifts. Batteries respond in under 100 milliseconds — far faster than any gas turbine — to inject or absorb power and correct frequency deviations. This is the most valuable service batteries provide, measured in dollars per megawatt per hour. • Peak shaving: Electricity is most expensive during peak demand periods, typically hot summer afternoons when air conditioning loads surge. A battery charged overnight during cheap off-peak hours and discharged during the 4 to 8 PM peak reduces the utility's need for expensive peaking generators and cuts transmission congestion costs. • Renewable integration (time-shifting): Solar panels produce power from roughly 8 AM to 6 PM, peaking around noon. Electricity demand peaks in the early evening. Batteries bridge this gap by storing midday solar and discharging it when people get home from work — a function so common in California that it's now called the "duck curve" problem, named after the shape of the net load graph. • Black start capability: After a major grid outage, conventional power plants need electricity to restart themselves — a chicken-and-egg problem. Battery systems can provide the initial power to bootstrap a grid restart without depending on another generator already being online. • Transmission deferral: A battery sited at a congested point on the grid can store energy when the line is underloaded and dispatch it locally during peak periods, delaying or eliminating the need for expensive transmission upgrades. Each of these services generates different revenue streams for battery operators, and real-world systems often "stack" multiple services simultaneously to maximize return on investment. A battery doing frequency regulation during off-peak hours might shift to peak shaving during the evening and provide black start reserves overnight. The Chemistry: Why LFP Won the Grid Storage Race Not all lithium-ion batteries are the same, and the grid storage market has decisively converged on one specific chemistry: lithium iron phosphate (LFP) . Understanding why requires a brief tour of the chemistry options. The lithium-ion family includes several cathode chemistries, each with different characteristics. NMC (nickel manganese cobalt) batteries, widely used in electric vehicles, offer high energy density — more kilowatt-hours per kilogram — making them ideal for applications where weight matters. NCA (nickel cobalt aluminum) is similar. Both NMC and NCA use cobalt, which is expensive, supply-chain-constrained, and mostly mined in the Democratic Republic of Congo under conditions that have attracted ethical scrutiny. LFP uses iron and phosphate instead of nickel and cobalt. This makes LFP cells significantly cheaper and eliminates the cobalt supply chain concern. LFP also has a flat voltage discharge curve, meaning it delivers consistent power across most of its charge cycle, and it is exceptionally stable chemically — a property that translates directly into safety and longevity. Chemistry Energy Density Cycle Life Safety Relative Cost Grid Use LFP 90-160 Wh/kg 3,000-6,000+ Excellent Lowest Dominant NMC 150-220 Wh/kg 1,000-2,000 Good Medium Some NCA 200-260 Wh/kg 500-1,000 Moderate High Rare Sodium-ion 100-160 Wh/kg 2,000-15,000 Excellent Potentially lower Emerging fast Vanadium flow 15-50 Wh/kg 10,000+ Very good High upfront Long-duration niche Iron-air 25-35 Wh/kg Designed for 20 yr Very good Target: <$20/kWh Long-duration, early For a grid storage project, weight is largely irrelevant — a battery farm isn't going anywhere. What matters is cost per kilowatt-hour of storage, cycle life (how many times the battery can charge and discharge before its capacity degrades significantly), and safety. On all three counts, LFP wins. An LFP pack designed for grid use can typically complete 4,000 to 6,000 full charge-discharge cycles before reaching 80% of original capacity, translating to 15 to 20 years of operation at one cycle per day. AI-generated image Modern battery storage projects are often co-located with solar farms, storing excess daytime generation for dispatch during peak evening demand. How a Grid Battery System Is Built A utility-scale battery storage project is not one giant battery. It's thousands of individual cells organized into a hierarchy of increasing scale, each level managed by control systems that balance load, temperature, and state of charge. At the base level, individual battery cells — typically cylindrical, prismatic, or pouch-format — are the fundamental electrochemical units. Each cell stores a small amount of energy, typically 50 to 300 watt-hours depending on format and chemistry. Cells are wired in series and parallel combinations inside a battery module , which typically delivers 48 to 100 volts and several kilowatt-hours. Modules are assembled into battery racks inside purpose-built enclosures, and multiple racks fill a battery container — typically a modified 20-foot or 40-foot shipping container format that holds 500 kilowatt-hours to 4 megawatt-hours per container. Each container includes its own battery management system (BMS), which monitors cell voltage, temperature, and state of charge in real time, balances charge distribution across cells, and triggers safety shutdowns if anomalies are detected. The BMS is the brain of the pack — without it, variations in individual cell characteristics would cause some cells to overcharge while others undercharge, accelerating degradation and creating safety risks. 267 GW Global grid battery power capacity (2025) 610 GWh Global grid battery energy capacity (2025) $117/MWh Levelized cost of storage (2023) 57.6 GWh US BESS deployed in 2025 (record) 70 GWh US BESS forecast for 2026 <1 sec Battery response time for frequency regulation Containers connect to power conversion systems (PCS) , which are essentially large inverters that convert the direct current from the batteries into grid-compatible alternating current. For utility-scale projects, these inverters work at medium voltage (typically 480V to 35kV) before stepping up to transmissio