Battery Powered cities? Ain't happening, no matter what Claude says

/

Reader CT Tempest writes:

BSD often provides AI's take on a subject so here's one more from Claude AI:

>>>>
Query:
What about the argument that "levelized cost" of solar energy isn't comparable to fossil fuel cost because of intermitancy and the requirement of fossil fuel plants to provide backup?

Claude:
Where the critics have a point
This is a real issue. The concept that captures it better is sometimes called "LCOS" (levelized cost of storage) or "system LCOE" or "firm power cost" — basically, what does it cost to deliver dispatchable, on-demand electricity from solar? Once you add 4–8 hours of battery storage, the cost goes up meaningfully. It's still trending down fast, but it's not the $0.03 headline number anymore.

Where the critics overstate it
That said, this argument gets stretched too far in a few ways. First, battery costs have dropped roughly 90% since 2010, and solar-plus-storage is already winning contracts against new gas plants in places like the U.S. Southwest and Australia. The "backup" cost is real but shrinking fast. Second, fossil fuel LCOE has its own hidden costs that don't show up in the number — carbon externalities, air pollution health costs, fuel price volatility, and geopolitical risk. If you're going to add system costs to solar, intellectual honesty requires adding externalities to fossil fuels too.

The honest bottom line
The pure LCOE comparison oversells solar's advantage — that's fair. But even when you account for integration and storage costs, solar-plus-batteries is increasingly competitive with new fossil fuel generation in most markets, and the cost trajectory is clearly in solar's favor. The "solar is cheap but you need gas backup" argument was very strong 10 years ago. It's getting weaker every year as storage costs plummet, but it hasn't fully gone away yet — especially for the last 10-20% of grid reliability (the "long duration storage" problem for multi-day cloudy stretches or seasonal variation).

So both sides are partially right, basically. The headline LCOE numbers are a bit misleading, but the broader trend of solar driving costs down is real even after you adjust for intermittency.
>>>>

I'm still betting on solar. Stay tuned.

And here’s what GROK has to say on the subject:

Q: In the debate over the true cost of solar and wind) emergy, I often hear that batteries can supply the necessary energy storage for times when the wind doesn't blow and the sun doesn't shine (nightime isn't good for solar generation, for instance). How practical is battery storeage for a large city, for instance, and how much would it cost. How many hours of electricity provide for that city using present day technology? Hurs, days, weeks? What would it cost?

https://x.com/i/grok/share/d4927204e5f24da1aa7dd85d2560c7e4

Battery storage using current (2025) lithium-ion technology is practical for short durations like 4–8 hours of a large city’s electricity needs, but becomes extremely expensive and logistically challenging for days, and effectively impractical for weeks.

ember-energy.org

For illustration, consider New York City (a typical large U.S. city example, with Con Edison serving NYC and parts of Westchester). NYC’s annual electricity consumption is roughly 50–55 TWh (about 50,000–55,700 GWh).

nyc.gov

This equates to:

  • Daily energy use: ~137–153 GWh per day.

  • Average power demand: ~5.7–6.4 GW.

  • Peak power demand: ~12–13 GW (summer records around 12.6–13.3 GW).

    coned.com

Battery storage must be sized for both power (GW, to meet peak demand instantly) and energy (GWh, to sustain output over time). Present-day grid-scale batteries are overwhelmingly lithium-ion (mostly LFP chemistry for cost and safety). Typical utility-scale systems today are designed for 4–8 hours of duration at full power.

eia.gov

How Many Hours/Days/Weeks of Storage?Storage capacity needed (in GWh) ≈ city’s daily energy use × number of days (or hours/24). This assumes the batteries supply the city’s full demand during low/no renewable generation (e.g., calm nights or multi-day lulls). Real-world systems add ~10–15% buffer for ~85–90% round-trip efficiency.

  • 4–8 hours (common today): ~23–51 GWh (roughly 1/6 to 1/3 of one day’s use). This covers daily solar “duck curve” shifting or short wind lulls. Feasible and increasingly deployed.

  • 24 hours (1 full day): ~137–153 GWh. This would fully bridge one full day/night cycle without renewables.

  • 2–3 days: ~274–459 GWh.

  • 1 week (7 days): ~960–1,070 GWh.

  • 2+ weeks: 2,000+ GWh (multiple terawatt-hours).

Current largest individual battery projects are in the 10–12 GWh range (with some planned larger), and total U.S. utility-scale battery capacity is only ~137 GWh as of end-2025.

seia.org

One day of NYC storage would require roughly the entire current U.S. fleet; a week would require 7–8× that.What Would It Cost?2025 all-in installed costs for utility-scale lithium-ion battery energy storage systems (BESS, including batteries, inverters, installation, and grid connection) have fallen sharply to roughly $120–150/kWh of energy capacity in many global markets (pack prices alone ~$108/kWh). U.S. figures can be higher depending on project specifics, but the trend is strongly downward.

energy-storage.news +1

Approximate capital costs for NYC-scale storage (using ~$150/kWh mid-range; does not include land, ongoing O&M, degradation, or replacement every 10–15 years/3,000–5,000 cycles):

  • 4 hours: ~$3.5–7.5 billion.

  • 24 hours (1 day): ~$20–23 billion.

  • 1 week: ~$144–160 billion.

These are upfront capital costs only. Batteries also incur replacement costs over time (they degrade), and you still need excess renewable generation (or other charging sources) to fill them. Levelized cost of storage is dropping (recent estimates ~$65/MWh in some analyses), but long-duration use makes it far more expensive per kWh delivered.

energy-storage.news

Practicality with Present-Day Technology

  • Hours (4–8 h): Highly practical and already happening. Batteries excel at short-term arbitrage, frequency regulation, and smoothing daily solar/wind variability. Costs have plummeted ~80–90% in the last decade, and deployments are surging (U.S. added a record ~58 GWh in 2025 alone).

    seia.org

  • Days: Marginally practical but very expensive. A few large projects approach multi-day scale in aggregate, but no single city has built anything close. Material demands (lithium, etc.), land use, and fire-safety/permitting become real constraints at this scale.

  • Weeks: Not practical today with lithium-ion. Costs balloon into the hundreds of billions, the physical footprint would be enormous (even at high energy density), supply-chain limits would be strained, and batteries are not optimized for infrequent, long-duration discharge (they’re better for daily cycling). For seasonal or multi-week lulls common in wind/solar-heavy grids, alternatives like pumped hydro, compressed air, or hydrogen are more suitable (though also limited geographically or by efficiency).

In the broader solar/wind cost debate: batteries handle intra-day and short-term intermittency well and are getting cheaper fast, but they are not a complete, low-cost solution for prolonged low-generation periods. High renewable penetration typically requires a mix of overbuilding renewables + transmission, short-duration batteries, longer-duration storage, and/or firm backup (e.g., gas with CCS, nuclear, or demand response). Pure battery reliance for multi-day/week coverage would dramatically raise the “true cost” of a renewables-heavy grid for a large city. Costs continue to fall, and emerging chemistries (e.g., sodium-ion, flow batteries) may help for longer durations, but as of 2025–2026, lithium-ion remains dominant for grid-scale use under ~12 hours.