Falling battery costs mean solar no longer trades price for reliability. That rewrites energy economics, industrial strategy, and emissions trajectories.
For more than a decade, solar energy has come with an asterisk.
Yes, it’s cheap.
Yes, it’s clean.
Yes, it’s fast to deploy.
But it’s intermittent.
That single word has done an extraordinary amount of work. It has justified gas as a “bridge fuel”, slowed grid reform, propped up fossil incumbents, and allowed policymakers to treat solar as a partial solution rather than a foundational one.
The latest analysis from Ember suggests that excuse has quietly expired.
Based on real-world auction data from Italy, Saudi Arabia, and India, Ember finds that long-duration utility-scale battery storage now costs around $125/kWh all-in, translating to a levelised cost of storage of roughly $65/MWh. Pair that with today’s already cheap solar, and you get dispatchable electricity for around $76/MWh .
Not in theory. In practice.
That changes the conversation.
The cost threshold that mattered more than people realised
Battery storage has been falling in price for years, but not all cost declines are equal. Some are incremental. Others cross thresholds that unlock new behaviour.
This one does the latter.
Ember’s analysis shows that outside the US and China, large, long-duration battery projects are now being built at around $125/kWh all-in. Roughly $75/kWh comes from core equipment, largely manufactured in China, with about $50/kWh for installation, grid connection, and EPC services that are mostly local.
Crucially, this isn’t driven by a single factor. Yes, battery cell prices have collapsed, with LFP cells now around $40/kWh in China. But the bigger story is system maturity.
Modern batteries last longer. Twenty years is now standard.
They’re more efficient. Around 90% round-trip AC-to-AC efficiency is normal.
They’re lower risk. Auctions and contracted revenues reduce financing costs.
Those improvements alone have cut the Levelised Cost of Storage (LCOS) by more than a third compared to earlier generations, even before accounting for falling equipment prices.
This is what “boring infrastructure” looks like. And boring infrastructure scales.
Dispatchable solar, not mythical baseload
It’s worth being precise about what this does and does not mean.
This is not about turning solar into perfect, year-round baseload power. That would require significant overbuild and far more storage to cover seasonal variability.
What it does mean is something far more useful.
Shifting roughly half of daily solar generation into evening and night-time hours aligns solar output much more closely with real demand. At an LCOS of $65/MWh, that adds about $33/MWh to the cost of electricity. With global average solar prices around $43/MWh in 2024, the result is dispatchable solar at roughly $76/MWh, and falling.
That is cheaper than new gas generation in most markets. It is dramatically quicker to deploy. And it avoids locking countries into volatile fuel imports.
The energy system doesn’t need ideological purity. It needs arithmetic that works.
Batteries are grid infrastructure, not just energy shifters
One of the most persistent analytical mistakes in energy modelling is treating batteries as single-purpose assets.
Energy shifting is only part of the value stack.
Grid-scale batteries also provide:
Frequency response
Voltage support
Congestion management
Capacity adequacy
Black-start capability
Each of those services carries its own revenue stream. When stacked together, they materially improve project economics and reduce risk.
Ember’s LCOS calculation explicitly excludes these additional revenues. In the real world, they lower the effective cost of storage even further.
This matters because it reframes storage from a “cost add-on” for renewables into a core piece of grid infrastructure. Once batteries are treated like substations rather than accessories, planning assumptions change quickly.
Long-duration storage closes the remaining gaps
Short-duration batteries handle daily variability extremely well. Seasonal and multi-day challenges require additional tools.
Those tools are arriving.
Advanced compressed air energy storage, such as Hydrostor’s A-CAES systems (I will publish an interview with Hydrostor on my Climate Confident podcast early in the new year), is moving from demonstration to deployment. Pumped hydro is being repowered and modernised rather than newly built. Thermal storage, sodium-ion batteries, and iron-air chemistries are expanding the storage palette.
The key point isn’t that batteries solve everything. It’s that solar plus batteries now solve most of the problem, and long-duration storage cleans up the rest.
The system no longer hinges on speculative technologies. It hinges on deployment speed and market design.
When electricity stops being scarce
Most energy debates are still trapped in a scarcity mindset. How do we ration? How do we balance? How do we keep the lights on?
But the combination of ultra-cheap solar and increasingly cheap storage breaks that frame entirely.
When clean electricity becomes abundant, reliable, and locally produced, the question stops being how do we manage energy? and becomes what do we do with it?
Take water. Desalination has always been technically feasible and politically awkward, because burning fossil fuels to make drinking water during a climate crisis is a grim trade-off. Remove the energy cost, and that dilemma evaporates. Solar-powered desalination turns water security into an infrastructure problem rather than a moral one, particularly for sun-rich regions already facing chronic drought stress.
Then there’s heat. Roughly half of global final energy demand is heat, not electricity. Cheap, stable clean power makes high-temperature heat pumps, electric boilers, and industrial electrification economically obvious rather than heroic. Entire classes of “hard-to-abate” emissions quietly become “hard to justify continuing”.
Zoom out further and you hit industry. When energy is cheap and predictable, location decisions change. Manufacturing no longer needs to cluster around fuel supply or ports. It clusters around skills, logistics, and demand. Energy-intensive industries can co-locate with solar and storage, decoupling growth from fuel imports and price volatility.
Digital infrastructure shifts too. Data centres, AI workloads, and high-performance computing are currently framed as climate liabilities. In a world of abundant clean power, they become flexible demand. Compute follows the sun. Workloads move in time as well as space. Carbon intensity falls without throttling innovation.
Perhaps the most underestimated impact is social. Distributed solar paired with storage bypasses fragile grids entirely. Regions that never had reliable electricity leapfrog straight to firm clean power, with knock-on effects for healthcare, education, and local industry. Energy poverty doesn’t decline gradually. It collapses where deployment is allowed to scale.
And underneath all of this, something even less visible happens. Fuel logistics fade from relevance. No tankers. No pipelines. No strategic reserves. Once built, the marginal cost and marginal risk of producing energy approach zero. Energy security stops being about geopolitics and starts being about maintenance and software.
This is what the Ember numbers are really pointing to. Not a better grid. A different baseline.
What this means for grids, emissions, and industry
For power systems, dispatchable solar flattens price volatility, reduces the need for gas peakers, and simplifies long-term planning. It replaces fuel risk with capital planning.
For emissions, the impact is immediate. Solar plus storage displaces coal faster, avoids methane leakage from gas backup, and delivers lifecycle emissions that fossil alternatives cannot touch.
For industry, the implications are structural. Stable, cheap, clean electricity enables electrification, co-location, and long-term contracting without exposure to fuel shocks. Energy stops being a constraint and becomes an advantage.
The real bottleneck is no longer technology
None of this is waiting on breakthroughs.
The bottlenecks now are:
Market design that undervalues flexibility
Interconnection queues built for a different era
Grid codes written around synchronous generation
Political reluctance to confront incumbents
These are governance problems, not engineering ones.
Solar didn’t just get cheaper. It grew up.
Ember’s analysis makes one thing clear. Solar energy is no longer cheap electricity that happens to arrive at inconvenient times. Paired with storage, it is dispatchable, predictable, and economically compelling at scale .
That removes the asterisk.
From here on, the question isn’t whether solar plus storage can underpin modern power systems. It’s how quickly institutions adapt to a reality that has already arrived.
This article was originally posted on TomRaftery.com
Photo credit Jonathan Cutrer on Flickr