Energy Storage Beyond Lithium-Ion: Sodium, Iron-Air, Flow Batteries and the Race for the Long-Duration Grid

· by · in Critical Technology. ·
Investigative energy-storage comparison showing lithium-ion, sodium-ion, iron-air and flow batteries connected to a modern electric grid with thermal, pumped-hydro and hydrogen storage in the background.

SURVXCOM CRITICAL TECHNOLOGY STACK / ENERGY STORAGE REPORT

Why the future grid will not be powered by one battery chemistry—and how sodium-ion, iron-air, flow batteries, thermal storage and other systems are competing to solve different versions of the same problem: when electricity is generated and when it is needed.

Technology Stack Article 012

CRITICAL TECHNOLOGY HUB: Explore the complete 30-article SURVXCOM Critical Technology reading path. This article belongs to the Energy, Materials & Industrial Systems lane.

EDITOR’S NOTE: This report compares energy-storage technologies by use case, duration, maturity, supply chain and system role. It does not treat every non-lithium technology as a direct substitute for lithium-ion. The evidence hierarchy prioritizes U.S. Energy Information Administration data, Department of Energy and national-laboratory research, current project documentation, manufacturer disclosures and independent reporting. Commercial announcements are distinguished from delivered, commissioned and operating systems.

The energy-storage debate is often framed as if the world is waiting for one miraculous battery. A better battery chemistry appears.

It becomes cheaper than lithium-ion. The grid changes.

Problem solved.

That is probably not how the next phase of energy storage will unfold. The power system has too many different storage problems.

A battery that stabilizes frequency for seconds is solving a different problem from one that shifts solar energy from noon to evening. A four-hour battery serving an afternoon peak is solving a different problem from a system expected to keep electricity available through three days of poor wind and cloudy weather.

A data center requiring rapid ride-through power has different needs from a utility trying to avoid building a peaker plant. A remote microgrid has different economics from a 1-gigawatt transmission system.

This is why the phrase energy storage hides an entire technology stack. Lithium-ion dominates today’s battery-storage market because it combines high round-trip efficiency, mature manufacturing, falling cost, fast response and an enormous supply chain created by consumer electronics and electric vehicles.

It is not disappearing.

In fact, U.S. deployment is accelerating. The Energy Information Administration says developers plan to add roughly 24 gigawatts of utility-scale battery storage in 2026 after a record 15 gigawatts in 2025. Texas alone accounts for more than half of planned additions.

But the market is also beginning to expose lithium-ion’s limits. Most grid batteries today are designed around relatively short discharge durations—often a few hours. Extending duration by adding more lithium cells means adding more of the expensive part of the system. Fire protection, augmentation, thermal management and critical-mineral exposure remain design considerations.

That creates room for technologies optimized around a different proposition: do not try to beat lithium-ion at everything. Beat it where duration, materials, safety, cycle life, siting or supply-chain resilience matter more than compactness.

Sodium-ion replaces lithium with a far more abundant element and is beginning to move toward commercial stationary storage. Iron-air batteries use reversible rusting to target one-hundred-hour storage.

Flow batteries separate power from energy by storing liquid electrolyte in tanks. Thermal systems store heat rather than electrons.

Pumped hydro stores gravitational potential energy. Compressed-air systems store energy in pressurized gas.

Hydrogen can move the time horizon toward days, weeks or seasons, though with lower round-trip efficiency and much more infrastructure. The result is not a battery replacement cycle. It is a storage specialization cycle.

Key Judgments

  • Lithium-ion remains the dominant grid battery. Its scale, efficiency, response speed and manufacturing maturity make it difficult to displace in short-duration applications.
  • The market is beginning to segment by duration. Seconds, four hours, overnight, multi-day and seasonal storage are not the same engineering problem.
  • Sodium-ion is the most credible near-term chemistry challenger for stationary storage. CATL has moved into commercial-scale sodium-ion BESS commitments, while U.S. companies are also developing domestic supply chains.
  • Iron-air targets a different market entirely. Form Energy’s 100-hour architecture trades compactness and round-trip efficiency for low-cost materials and multi-day duration.
  • Flow batteries decouple power from energy. Increasing duration can be achieved largely by increasing electrolyte tank volume rather than duplicating the entire power stack.
  • Thermal storage is underappreciated. When the end use is heat—or when stored heat can be converted back to electricity—thermal systems can avoid some electrochemical constraints.
  • Pumped hydro remains the benchmark for very large storage. It is mature and durable but geographically and permitting constrained.
  • Compressed-air and gravity systems are infrastructure technologies, not battery technologies. Their economics depend heavily on geology, civil works and site conditions.
  • Hydrogen expands duration but sacrifices efficiency. It may make sense for rare, long-duration events or sector coupling rather than daily cycling.
  • AI and data centers are strengthening the business case for long duration. Their power demand is continuous, while grid supply increasingly varies by hour and weather.
  • The correct comparison is system value, not chemistry hype. A lower-efficiency technology can still be valuable if energy capacity is cheap and discharge events are infrequent.
  • The future grid is likely to use a portfolio. Lithium-ion for fast and short-duration work; other technologies for longer, safer, cheaper or more supply-resilient roles.

The First Question Is Duration

Energy storage is usually discussed in megawatts. That is only half the story.

Megawatts measure how quickly a storage system can deliver power. Megawatt-hours measure how much energy it contains.

A 100-megawatt battery containing 400 megawatt-hours can run at full output for roughly four hours. A 100-megawatt system containing 10,000 megawatt-hours can theoretically sustain full output for about one hundred hours.

Those are completely different assets even though their power rating is identical. This is why long-duration energy storage is difficult to define with one number. National-laboratory work notes that “long duration” is often associated with more than ten hours, but the actual duration required for resource adequacy can range from several hours to multiple days depending on the grid.

MILLISECONDS / SECONDS
frequency response • power quality
        │
        ▼
MINUTES
ride-through • contingency response
        │
        ▼
1–4 HOURS
solar shifting • peak shaving • ancillary services
        │
        ▼
4–12 HOURS
evening / overnight shifting
        │
        ▼
12–100 HOURS
weather events • multi-day reliability
        │
        ▼
DAYS / WEEKS / SEASONS
deep resilience • seasonal imbalance

THERE IS NO SINGLE "BEST BATTERY" ACROSS THIS ENTIRE LADDER

That distinction should govern every technology comparison. This point matters because the technology should be evaluated as part of the surrounding system rather than as an isolated claim or capability.

LABORATORY
materials / cell concept
    │
    ▼
PROTOTYPE
integrated hardware
    │
    ▼
PILOT
limited real-world deployment
    │
    ▼
DELIVERED
equipment physically on site
    │
    ▼
COMMISSIONED
connected and tested
    │
    ▼
OPERATING FLEET
repeatable field evidence
    │
    ▼
BANKABLE SCALE

CONTRACTED CAPACITY CAN SIT ANYWHERE ABOVE THIS LADDER.
IT IS NOT THE SAME AS OPERATING CAPACITY.

Why Lithium-Ion Still Wins

Any article titled Beyond Lithium-Ion should begin by acknowledging how difficult lithium-ion is to beat. The chemistry benefits from decades of manufacturing scale.

Cell factories are enormous.

Power electronics are standardized. Developers understand bankability.

Utilities understand operating behavior. Insurers increasingly understand the risks.

Suppliers can deliver containerized systems with predictable interfaces. And round-trip efficiency is high.

EIA’s current market data shows how dominant the technology has become in U.S. utility-scale storage. Battery capacity increased from roughly 20 gigawatts on a time-adjusted basis in 2024 to more than 33 gigawatts in 2025, and developers plan another 24 gigawatts of battery additions in 2026.

California demonstrates what that deployment can do. EIA reports that CAISO battery capacity increased sharply through April 2026 and that battery discharge in the first five months of the year was roughly triple the level of the same period two years earlier, increasingly shifting midday solar into evening demand.

This is a mature grid function.

Where lithium-ion becomes less elegant is duration. If energy duration doubles, developers typically add more cells.

The energy-bearing component remains expensive. That creates an opening for technologies whose energy reservoir is cheap even if their power conversion equipment is not.

Sodium-Ion Crosses Into Commercial Scale

Sodium-ion is the most obvious chemistry to compare with lithium-ion because the architecture is familiar. Both move ions between electrodes through an electrolyte.

Both can be packaged into cells, modules and containerized battery systems. But sodium is vastly more abundant than lithium and can reduce exposure to some constrained battery materials.

The tradeoff is lower energy density. For vehicles, that matters enormously.

For a stationary battery sitting beside a substation, it may matter much less. The important 2026 change is commercial scale.

CATL unveiled its TENER sodium-ion energy-storage system in June and says Chinese deliveries will begin in September, with cumulative shipments expected to reach 1 gigawatt-hour by year-end. The company says international deliveries begin in 2027.

More important than the product launch is the order book. CATL says it signed a three-year, 60-gigawatt-hour sodium-ion storage agreement with HyperStrong and later announced a 5-gigawatt-hour European partnership with Alfen.

These are corporate commitments, not evidence that all of that capacity is installed. But they indicate that sodium-ion is moving from demonstration chemistry toward manufacturing strategy.

The United States is moving as well. American startups are targeting grid and data-center applications with sodium chemistries intended to reduce China-linked battery supply-chain exposure. ESS announced a modular sodium-ion BESS in July after expanding from its iron-flow roots, while other U.S. developers are pursuing sodium-ion and sodium-based stationary systems.

The key question is not whether sodium replaces lithium. It is whether sodium becomes the second major mass-manufactured electrochemical platform.

Characteristic Lithium-ion Sodium-ion
Manufacturing maturity Very high Rapidly emerging
Energy density Higher Lower
Stationary suitability Excellent Potentially excellent
Material abundance More constrained Sodium highly abundant
Supply chain Massive but geographically concentrated New supply chain still forming
Main strategic role Incumbent short-duration storage Possible lower-cost / safer diversification

Iron-Air and the 100-Hour Bet

Iron-air batteries make a different bet. They do not try to compete with lithium-ion on compactness.

They compete on duration and material cost. Form Energy’s system uses iron, air and water in a reversible process often described as reversible rusting.

During discharge, iron reacts with oxygen. During charging, electrical energy reverses the reaction.

The company targets approximately one hundred hours of storage. That is not an evening battery.

It is a multi-day reliability asset. The architecture is attractive because iron is abundant and inexpensive.

The tradeoff is much lower energy density and lower round-trip efficiency than lithium-ion. If a battery cycles every day, efficiency losses matter greatly.

If a multi-day system is called only during rare weather events or prolonged supply gaps, cheap stored-energy capacity may matter more. That is why long-duration storage economics cannot be reduced to efficiency alone.

Form Energy says its contracted commercial portfolio has expanded to more than 80 gigawatt-hours of iron-air systems. Recent projects include utility partnerships and data-center-linked developments, including work involving Xcel Energy, Google and Crusoe.

The scale is notable.

But contracted portfolio is not installed fleet. The technology still has to prove manufacturing ramp, field reliability, maintenance, real project economics and grid integration at scale.

LITHIUM-ION
high efficiency
high power
compact
best for frequent short-duration cycling

        VS.

IRON-AIR
low-cost active materials
large footprint
lower efficiency
designed for ~100-hour duration

QUESTION:
HOW OFTEN WILL THE GRID NEED THE STORED ENERGY?

Flow Batteries: Energy in Tanks

Flow batteries solve the duration problem differently. Instead of storing all active material inside a sealed cell, they store liquid electrolytes in external tanks and pump them through an electrochemical stack.

This creates a useful separation:

the stack determines power;

the tank volume largely determines energy. If a project needs more duration, it may add more electrolyte and tank capacity without proportionally duplicating the entire power-conversion system.

Vanadium redox flow is the most established version. Invinity is currently delivering and commissioning commercial vanadium-flow projects. Its 20.7-megawatt-hour Copwood project in the United Kingdom was delivered in May and is expected to become Europe’s largest operating vanadium flow battery when grid-connected. The company has also been selected to design a gigawatt-hour-scale system in Switzerland, though that project remains in engineering rather than completed deployment.

Iron-flow systems pursue similar architectural advantages with different chemistry. ESS commissioned iron-flow systems at Turlock Irrigation District in California in May and has joined Salt River Project and Google on a 5-megawatt / 50-megawatt-hour pilot in Arizona.

A third-party utility report on an ESS installation at Burbank Water and Power concluded the system operated as intended and identified a utility use case for the technology. Again, scale matters.

A successful megawatt-hour-scale installation proves something different from a gigawatt-hour fleet. But flow technology’s core proposition is durable: separate power cost from energy cost.

ELECTROLYTE TANK A ──┐
                     │
                     ▼
                ELECTROCHEMICAL
                    STACK
                     ▲
                     │
ELECTROLYTE TANK B ──┘

POWER CAPABILITY
≈ size / number of stacks

ENERGY CAPACITY
≈ amount of electrolyte in tanks

MORE DURATION CAN MEAN:
BIGGER TANKS
NOT AN ENTIRELY NEW BATTERY SYSTEM

Thermal Storage: Store Heat, Not Electrons

Electricity storage discussions often assume electricity must remain electricity while it waits. That is not always necessary.

Thermal storage converts electricity into heat and stores the heat in a material such as molten salt, water, ceramic bricks, rocks or other media. The heat can later be used directly for industrial processes or converted back into electricity.

Direct heat use is especially interesting because industrial heat represents a large energy demand. If a factory needs steam tomorrow morning, storing heat overnight may be more rational than storing electricity in an electrochemical battery and converting it to heat later.

DOE includes thermal systems among its major long-duration storage technology families. The advantages can include inexpensive storage media, long duration and reduced dependence on battery minerals.

The disadvantage is conversion efficiency when the system must return heat to electricity. This again reinforces the central rule: the correct technology depends on the final service.

Pumped Hydro, Compressed Air and Gravity

The largest installed electricity-storage technology in the United States is not lithium-ion. It is pumped-storage hydropower.

EIA data still shows roughly 23 gigawatts of pumped-storage capacity in the U.S. fleet.

Pumped hydro is simple in principle. Use excess electricity to pump water uphill.

Release it downhill through turbines when electricity is needed. The reservoir is the battery.

Its advantages are scale, long life and proven operation. Its constraint is geography.

Suitable sites require elevation difference, water, civil works, permitting and environmental acceptance. Compressed-air energy storage follows a similar physical philosophy.

Electricity powers compressors.

Compressed air is stored in caverns or engineered vessels. Later the pressure drives machinery to regenerate electricity.

Gravity-storage concepts use heavy masses instead of water. These technologies are often discussed as if they are alternative battery chemistries.

They are better understood as infrastructure projects. The economics depend less on chemistry and more on geology, excavation, construction and asset life.

Hydrogen and the Seasonal Frontier

The farther storage duration extends, the harder it becomes for a conventional battery to compete economically. A battery built for one hundred hours contains a great deal of energy capacity that may sit unused most of the year.

At seasonal scale, chemical fuels become attractive. Electricity can produce hydrogen through electrolysis.

Hydrogen can then be stored and later used in turbines, fuel cells or industrial processes. The obvious disadvantage is efficiency.

Electricity-to-hydrogen-to-electricity loses much more energy than charging and discharging a lithium-ion battery. But round-trip efficiency is only decisive when cycling is frequent.

A system intended to cover rare week-long supply gaps may value storage cost and shelf life over daily efficiency. DOE therefore includes bidirectional hydrogen within its long-duration storage portfolio.

Hydrogen should not be described as a universal battery replacement. It is a candidate for the extreme-duration end of the storage ladder.

Why AI Changes the Storage Market

Article 006 established the central infrastructure problem of AI: data centers want enormous quantities of electricity continuously. Grid resources do not all behave continuously.

Solar output disappears at night.

Wind varies.

Transmission can become constrained. Gas generation requires fuel and pipeline capacity.

Nuclear plants provide firm output but take years to build. Short-duration lithium batteries can shift several hours of energy.

They do not automatically solve a multi-day weather event. This is why Reuters reported in March that rising AI demand is accelerating interest in long-duration storage. Large technology companies and utilities are beginning to examine iron-air, flow batteries and other systems as part of firm-power packages.

The logic is not that a data center runs directly on one experimental battery. It is that long-duration storage can become one layer inside a larger power portfolio.

GRID
  │
  ├── NUCLEAR / FIRM GENERATION
  ├── NATURAL GAS
  ├── SOLAR
  ├── WIND
  ├── TRANSMISSION
  ├── 2–4 HOUR LITHIUM BATTERY
  ├── 10+ HOUR STORAGE
  └── MULTI-DAY STORAGE
          │
          ▼
      DATA CENTER

NO SINGLE RESOURCE HAS TO DO EVERYTHING

Efficiency Is Not the Whole Economic Story

Storage technologies are often ranked by round-trip efficiency. That is useful.

It is not sufficient.

A lithium battery may return a very high share of the electricity used to charge it. An iron-air or hydrogen system may return significantly less.

But if the long-duration system uses much cheaper energy-storage media, its total cost for a rare multi-day event may still be competitive. The correct economic question is:

What does it cost to provide the required grid service over the life of the asset? That includes:

power equipment;

energy reservoir;

round-trip losses;

cycle life;

maintenance;

augmentation;

land;

interconnection;

insurance;

fire protection;

materials;

financing;

and how often the system is actually dispatched. National-laboratory research reinforces this point. The value of long-duration storage varies sharply by region, renewable penetration, generation mix and transmission availability. Some systems may value 20 to 40 hours of storage strongly; others may not. There is no universal duration threshold that makes a technology valuable everywhere.

The Critical-Minerals Question

Storage is also industrial policy.

Lithium-ion supply chains depend on lithium, graphite, cathode materials, separators, electrolytes and enormous Asian manufacturing ecosystems. LFP reduced dependence on nickel and cobalt.

Sodium-ion may reduce dependence on lithium. Iron-air shifts active materials toward iron and water.

Flow systems depend on different electrolytes, tanks and pumps. Thermal storage can use extremely abundant materials.

No technology eliminates supply chains. It changes them.

This matters because energy security is increasingly about avoiding one indispensable material or manufacturing geography. Sodium-ion offers one possible route to diversification.

But there is an irony.

China is already moving quickly to dominate sodium-ion manufacturing as well. Changing chemistry does not automatically change industrial geography. Technological sovereignty requires manufacturing capacity, not merely abundant raw materials.

The Storage Portfolio Grid

The future grid is unlikely to select one storage winner. It will probably behave more like a financial portfolio. Different assets will be held for different risk conditions.

Technology Likely duration niche Core advantage Core limitation
Lithium-ion Minutes to ~4 hours, sometimes longer Mature, efficient, fast, scalable Duration cost and thermal-management concerns
Sodium-ion Short to medium duration Abundant sodium, stationary fit, supply diversification Lower energy density, immature global supply chain
Iron-air Multi-day / ~100 hour Very low-cost abundant active materials Large footprint, lower efficiency, emerging deployment
Vanadium / iron flow Several hours to long duration Power-energy decoupling, long cycle life Pumps, tanks, electrolyte economics, lower density
Thermal Hours to days Cheap storage media, excellent for heat loads Electricity reconversion losses
Pumped hydro Hours to days Mature, huge scale, long asset life Geography, permitting, civil works
Compressed air Long duration Large energy reservoir potential Geology / infrastructure dependent
Hydrogen Days to seasonal Very long duration and sector coupling Low round-trip efficiency, infrastructure intensive

The portfolio approach also changes how storage should be judged. A technology does not need to dominate annual installations to become strategically important. It may only need to solve the rare event that the dominant technology solves poorly.

The SURVXCOM Storage Fitness Test

Before calling any storage technology a breakthrough, ask what job it is actually being hired to perform. This point matters because the technology should be evaluated as part of the surrounding system rather than as an isolated claim or capability.

1. Duration

How many hours—or days—must the system discharge at rated power?

2. Power

How quickly must it respond and how much instantaneous power is required?

3. Cycling Frequency

Will it charge and discharge daily, weekly or only during rare events?

4. Round-Trip Efficiency

How much energy is lost between charging and discharge?

5. Energy Cost

How expensive is each additional hour of stored energy?

6. Power Cost

How expensive are the inverters, stacks, turbines, pumps or conversion equipment?

7. Cycle Life

How quickly does performance degrade under the intended use?

8. Safety

What are the fire, pressure, chemical, thermal and operational hazards?

9. Materials

Does the technology depend on scarce, geographically concentrated or geopolitically sensitive materials?

10. Siting

Does it require special geology, water, large land area or specific climate conditions?

11. Commercial Maturity

Is the system a laboratory result, pilot, delivered project, commissioned plant or bankable fleet?

12. Grid Service

What specific problem does it solve better or more cheaply than alternatives?

The right energy-storage question is not “Which battery wins?” It is “Which technology fits the duration, duty cycle and risk the grid actually needs?”

What to Watch Next

1. Sodium-Ion Shipments

Watch whether CATL reaches its announced 2026 delivery targets and whether large European and Chinese contracts convert into commissioned projects.

2. U.S. Sodium Manufacturing

Watch Peak Energy, ESS/Alsym and other domestic sodium programs for actual cell production, cost and supply-chain localization rather than prototype announcements.

3. Form Energy’s Manufacturing Ramp

Watch installed and commissioned megawatt-hours—not only contracted portfolio—as the Weirton, West Virginia factory scales.

4. Iron-Air Reliability

Watch field performance, round-trip efficiency, availability and maintenance under real multi-day dispatch conditions.

5. ESS Project New Horizon

Watch Salt River Project and Google’s 5 MW / 50 MWh iron-flow pilot for validated operating data.

6. Vanadium Flow Scale

Watch Invinity’s Copwood commissioning and the engineering-to-order transition for the proposed GWh-scale Swiss system.

7. DOE 2026 Storage Research

Watch sodium solid-state, soluble iron electrolytes, vanadium flow improvements, sodium-flow systems and transmission-storage valuation emerging from the current national-laboratory portfolio.

8. Thermal Storage

Watch commercial industrial-heat projects and whether thermal systems begin displacing electrochemical storage where heat is the actual end use.

9. Pumped Hydro

Watch licensing and construction timelines. Technical maturity means little if projects cannot be permitted and built.

10. Data-Center Storage Procurement

Watch whether hyperscalers begin signing large non-lithium storage contracts directly to support 24/7 power requirements.

11. Storage Market Design

Watch whether capacity markets and grid-planning rules compensate multi-day storage for reliability value rather than only short-duration arbitrage.

12. Hybrid Portfolios

Watch projects combining lithium-ion with flow, iron-air, thermal or other long-duration systems. Hybridization may become more important than a chemistry winner.

The Grid Needs More Than One Kind of Battery

Lithium-ion changed the grid because it made storage modular. A battery plant could be built in containers rather than mountains.

It could respond in milliseconds.

It could move solar power from noon into the evening. It could stabilize frequency and absorb excess energy.

That achievement is real.

The mistake would be assuming the first successful grid battery must also solve every future grid problem. The next challenge is longer.

Overnight.

Through a storm.

Across multiple cloudy days.

Through a transmission emergency.

Across a period when wind output collapses. Or across seasons.

Different durations reward different physics. Sodium may win where stationary systems want familiar battery architecture without lithium.

Flow batteries may win where developers want long cycle life and cheap added duration. Iron-air may win where a huge reservoir of rarely used backup energy matters more than efficiency.

Thermal storage may win where the load ultimately needs heat. Pumped hydro may remain unbeatable where geography permits it.

Hydrogen may become useful when the storage horizon is too long for batteries to remain economical. The grid is not choosing one technology.

It is learning to match technologies to time. And that may be the most important storage breakthrough of all.

The future of energy storage is not one chemistry replacing another. It is a layered system in which each technology earns its place by solving the duration problem it was actually built for.

Related SURVXCOM Reading

Critical Technology Hub & Reading Path

Start with the hub: SURVXCOM Critical Technology Hub. This article is part of SURVXCOM’s 30-piece cornerstone tree explaining the systems beneath technological power. Primary lane: Energy, Materials & Industrial Systems.

Continue in the Critical Technology Stack

Across the SURVXCOM Ecosystem

Related SURVXCOM lanes: When the Systems Fail — Preparedness and resilience when infrastructure becomes unreliable.

Primary Research and External Sources

Source discipline: EIA is used as the source of truth for current U.S. storage deployment. DOE and national-laboratory sources frame technology classes and long-duration system value. CATL, Form Energy, ESS and Invinity announcements document company technology, contracts and project status; they are not treated as independent proof of commercial economics. Contracted gigawatt-hours are not counted as operating capacity. Delivered equipment is distinguished from commissioned grid operation. Long-duration storage is not assumed to be economically superior in every region; grid value depends on duration, generation mix, transmission, cycling frequency and market design.

Leave a Reply

Your email address will not be published. Required fields are marked *