Batteries Are Getting Built. Pumped Storage Takes Years. Here’s Why That Matters.

Every investor sees the battery buildout. Fewer investors are honest about what 4 hour storage can’t do when the grid is under stress for days, not minutes. Pumped storage hydro is not “the old tech.” It’s a different asset class, with different failure modes, and in the right place it’s still the cleanest way to buy duration at scale.

Start with the system need: demand up, dispatchable under pressure

NERC’s latest long-term assessment shows demand forecasts climbing sharply, driven in part by large loads like data centers, alongside significant announced retirements.
That combination is why long-duration flexibility keeps coming back into the conversation.

At the same time, batteries are scaling fast:

• EIA reports cumulative U.S. utility-scale battery storage capacity exceeded 26 GW in 2024, after adding 10.4 GW that year.
  
• EIA expected 18.2 GW of utility-scale battery storage additions in 2025 (planned), and noted developers planned 4.4 GW of new natural gas-fired capacity in 2025.
  

So the question is not “batteries or pumped storage.” Batteries are already happening. The real question is:

Where do batteries stop being enough, and where does pumped storage become the lowest-risk way to buy reliability?

What pumped storage is (and why markets value it differently)

FERC’s plain English definition is the simplest: pumped storage moves water between two reservoirs at different elevations, pumping when demand is low and generating when demand is high.

Two investor relevant implications:

1. It can provide ancillary services that help with grid reliability and renewable integration.
   
2. It is an infrastructure asset with a long technical life, where the core “storage medium” is water and gravity, not a battery cell warranty.
   

The U.S. installed base is not massive, but it is real:

• FERC reports 24 pumped storage projects constructed and operating, totaling over 16,500 MW of installed capacity.
  

That installed base exists because pumped storage solves specific problems well: large power, long duration, system services, and stability.

The cost reality: pumped storage is expensive per kW, but it’s not built for 2 hour economics

If you underwrite pumped storage like a short duration battery, you’ll reject every good project.

NREL’s 2024 Annual Technology Baseline (ATB) gives a useful view of capital cost ranges for pumped storage resource classes:

• Closed loop pumped storage (two new reservoirs): roughly $3,029/kW to $4,501/kW across classes shown (2021 dollars per kW in ATB table).
  
• Projects using one existing reservoir can show lower costs in the table (classes shown around $1,730/kW to $3,250/kW, with wider variability).
  
• Round trip efficiency is typically framed in a range, with 80% used as a central estimate and sources reporting 70%–87%.
  

So yes, pumped storage is capital-intensive. But the underwriting case is about:

• Duration (8-12 hours is common framing in planning)
  
• Ability to cycle hard without “cell degradation economics”
  
• Monetizing system services and capacity value over long periods
  

Comparison table: pumped storage vs 4-hour batteries vs gas peakers

This is a blunt table on purpose. You need a quick screen before you go deep.

Attribute	Pumped Storage Hydro	Utility-Scale Li-ion BESS (typical 4-hour build)	Gas peaker / fast-start engines
Best at	Long duration + grid services	Short duration shifting, fast deployment	Firm capacity, weather resilience (if fuel is firm)
Duration	Hours to many hours	Usually 1-4 hours (some longer)	As long as fuel is available
Asset life	Multi-decade infrastructure	Shorter life with augmentation/repower cycles	Multi-decade with overhauls
Siting	Hard (water, head, land)	Easier (but needs interconnection)	Easier than hydro (but needs gas + interconnection)
Grid value	Capacity + ancillary + stability	Energy shifting + some ancillary	Capacity + reliability products
Primary investor risk	Permitting + schedule + capex	Revenue saturation, degradation, replacement cost	Gas deliverability, emissions policy risk, siting

I’m not arguing pumped storage replaces batteries. I’m arguing that if you need “keep the system standing” duration and services, pumped storage is one of the few proven tools.

Underwriting pumped storage: 10 questions that decide the deal

If you only remember one section, make it this one. These are the questions that determine whether pumped storage is a real asset or a perpetual development story.

1. Is the site closed-loop or does it touch natural waterways?
   Closed-loop generally de-risks environmental opposition compared to projects heavily tied to river operations (case-by-case, but directionally true).
   
2. What is the head, and what is the water balance?
   You’re underwriting physics. Not PowerPoint.
   
3. Where is the interconnection, and what upgrades are implied?
   Do not ignore this because “it’s hydro.” The grid is still the grid.
   
4. What is the real constructability story?
   Geotech, access roads, dam design, spoil, and laydown can swing capex more than most people admit.
   
5. What is the permitting path and stakeholder map?
   Local opposition can kill timeline. Timeline kills IRR.
   
6. What is the commercial revenue stack, in order of confidence?
   Capacity and ancillary services are often more important than pure arbitrage for pumped storage. But you need to show it with market rules and historical pricing, not vibes.
   
7. Who is the offtaker, if any?
   Pumped storage can be merchant, but merchant risk is not free. Long-duration storage often needs some form of contracted support to be financeable.
   
8. What are your long-term O&M and refurbishment assumptions?
   Hydro O&M is real, but it’s not battery augmentation.
   
9. What is the exit strategy?
   Is this a long-hold infrastructure asset, or are you building for flip at notice-to-proceed?
   
10. What kills the project, specifically?
    Make a kill list. Fund the work that retires those kill risks first.
    

A simple chart: “best tool by duration”

This is how I explain it to investors without turning it into a chemistry lecture.

Best-fit resource by needed duration (illustrative)

0–2 hours    | Batteries ██████████ | Pumped Storage ██ | Gas ██████

2–6 hours    | Batteries █████████  | Pumped Storage ████ | Gas ███████

6–12 hours   | Batteries █████      | Pumped Storage ██████████ | Gas ████████

12+ hours    | Batteries ██         | Pumped Storage ██████████ | Gas ██████████ (fuel-dependent)

Batteries win on speed and modularity. Pumped storage wins when the problem is sustained and system-level.

What to do today: an investor checklist

If you want exposure to pumped storage (or want to know when to pass), here’s a concrete plan.

1. Pick 2–3 target regions, not 20
   
   • Look for constrained transmission, high volatility, or policy support for long-duration reliability.
     
2. Screen sites like a miner
   
   • Head, geology, access, land control
     
   • Then grid proximity
     
   • Then permitting complexity
     
3. Treat interconnection as a gating item early
   
   • If you can’t see a believable POI path, do not burn years in licensing first.
     
4. Decide how you’ll monetize before you finalize the design
   
   • Pure arbitrage is rarely enough
     
   • Model capacity and ancillary value explicitly
     
5. Stage capital against kill risks
   
   • Spend first on work that answers “can this be permitted and built at the cost we think?”
     

Bottom line

Pumped storage is expensive and slow compared to batteries, and that is exactly why it can be investable. It’s hard to build, but once built it behaves like infrastructure. FERC’s operating fleet is only 24 projects totaling over 16,500 MW, which is a reminder that this is a scarce and specialized asset class, not a commodity product.
If you underwrite it with the right yardstick (duration, grid services, long life, and real siting), pumped storage can be the cleanest way to buy long-duration reliability at scale.

What to watch next (5 bullets)

• How quickly battery additions continue to scale relative to long-duration needs (EIA’s recent build and planned additions are significant).
  
• The pipeline of pumped storage projects moving from permits to licenses and actual construction starts.
  
• Market rule changes that better pay for long-duration capacity and essential reliability services
  
• Regional transmission buildouts and whether they reduce (or shift) the value of long-duration storage
  
• Water and land constraints becoming the real limiter for hydro-style solutions (especially in high-growth regions)
  

Sources

• FERC, Pumped Storage Projects (definition + installed base)
  
• NREL, 2024 ATB: Pumped Storage Hydropower (capital cost ranges, assumptions, efficiency range)
  
• EIA, U.S. battery capacity increased 66% in 2024 (26 GW cumulative; 10.4 GW added in 2024)
  
• EIA, Planned 2025 additions (18.2 GW batteries; 4.4 GW natural gas planned)
  
• NERC, 2024 Long-Term Reliability Assessment (load growth and retirement context)
  

DOE, Hydropower Market Report materials (project pipeline context)