The financial realities facing US long-duration energy storage integration

Using a conservative cost assumption of US$180-per-kWh from BNEF’s projections, the incremental capital cost to double duration from four to eight hours is approximately US$72 million. However, at the current PJM capacity pricing, annual revenue only increases from US$7 million to US$8.5 million.

“That’s an incremental yield of about 5%,” Khan explains. “As you can tell, that’s not really helping long-duration storage as much. The duration ladder for the ELCC rating should have a similar slope as the cost curve. That way, developers and investors can feel like the additional investment in duration has a reasonable payback.”

This helps explain why, despite widespread acknowledgement that the grid needs 8-, 10-, and 12-hour systems, the market continues to be dominated by four-hour installations. At least in this case, the economics don’t support the capital intensity required for longer duration.

Captive batteries

The data centre-BESS convergence has generated significant attention, but Khan argues the industry has misunderstood what captive batteries actually provide versus wholesale market participants.

“Captive batteries are selling time—the ability to interconnect faster than the conventional process,” Khan emphasises. Calibrant Energy’s delivery of a 31MW/62MWh BESS at an Aligned Data Centres campus in the Pacific Northwest illustrates that the battery’s primary value wasn’t energy arbitrage but rather providing firm capacity that accelerated interconnection timelines.

Captive BESS provides speed to power, flexibility, and ride-through capability. The anchor is the opportunity cost of interconnection delays, not the levelised cost of storage (LCOS).

In contrast, wholesale market batteries face an entirely different risk profile. “Your biggest counterparty risk is ISO settlement,” Khan notes. “Your duration is going to be more optimised to the market, whereas in a captive situation, you could do more than four hours, even more than eight hours, if you’re going to be an islanded private power network.”

But Khan is also emphatic about maintaining perspective on data centres’ role in the broader storage market, “BESS for data centres is an accelerant, but it’s not the foundation for increased interest in the battery storage market.”

The structural driver remains energy arbitrage enabled by intense solar and wind buildout depressing midday power pricing. “We are seeing power prices go up. Inputs to power production have gone up in the last several quarters, in some instances even outstripping inflation,” Khan observes. “This is going to have a long-term deflationary impact on power pricing in the markets, and so it’s going to have a strong foundation for growth.”

Critically, Khan believes battery storage’s fundamental value proposition is resilient even if data centre growth moderates. “Even if data centres were to slow down for whatever reason, I don’t think BESS would be as disproportionately impacted as, let’s say, gas generation.”

Vertical integration

The industry debate over vertical integration has typically focused on margin capture and supply chain control. Khan offers a more nuanced perspective that reframes integration as primarily a risk management strategy.

“I think owning a critical component and doing really well at that is evidence of good risk management rather than building up,” Khan explains. “That distinction actually matters more to an investor than it does to a developer, but it’s an understanding that’s still evolving in the industry.”

Tesla’s commitment to LG for domestic lithium iron phosphate (LFP) cells exemplifies this approach. “It’s not a pure bet on cell margins,” Khan notes. “They’re riding through the tariffs and any policy insurance or schedule uncertainty, which all three of those things can kill projects.”

Long-duration technology

Khan’s analysis of when alternative technologies can compete with lithium-ion (Li-ion) reveals that true long-duration advantage requires meeting two distinct criteria simultaneously.

“Long-duration is not just being able to discharge for longer, but also the fact that your asset can be warranted for longer as well,” Khan emphasises. “Companies that are looking to break through into the long-duration space have to check both boxes. They have to last longer and they have to discharge for a longer duration, with financeable trade-offs.”

Beyond 10 to 12 hours, when cycling between 100 to 200 cycles annually and being paid primarily for availability, Li-ion’s cycle life advantage becomes “almost worthless,” and the equation shifts entirely to cost of energy. The lower round-trip efficiency (RTE) of many alternative technologies only works economically when charging occurs at negative power prices.

In the 8-to-12-hour range, asset life becomes the differentiator. Khan cites Hydrostor’s advanced compressed air energy storage (A-CAES) technology, which targets 50-year operational life versus Li-ion’s 20-to-25-year lifespan. “A 50-year life without augmentation rewrites the whole calculation for LCOS.”

For 4-hour applications, lithium dominates, but Khan identifies a specific niche where alternatives can command premium pricing: non-flammable chemistries for high-value collocations. “If you’re sitting next to a multi-billion-dollar data centre, their insurer is going to impose stricter requirements for fire safety, and as a result, insurance costs will go up quite a bit. The argument there is insurability, not purely cost.”

Khan points to Eos Energy’s recent disclosures showing pipeline growth exceeding 50% driven by data centre demand as evidence that “insurability and fire safety is becoming more of a binding requirement if you are looking to be collocated with large-scale infrastructure.”

On sodium-ion (Na-ion), Khan is clear about its strategic positioning: “Na-ion is a hedge against lithium rather than a true long-duration play. Even though Na-ion loses on a few accounts against lithium in terms of performance, it’s all about cost hedging and not particularly duration.”

Capacity market reform

Beyond identifying structural flaws, Khan offers a prioritised framework for capacity market reform that would genuinely enable LDES deployment.

First priority: multi-year commitments. “It would be great to have a 10-year price lock on capacity,” Khan states. “We need commitments that can underwrite capital-intensive, long-life assets.”

Second: aligning duration accreditation with cost curves. “The slope of cost and the slope of duration needs to be better aligned because you cannot expect more LDES to come in when those two things are diametrically opposed to each other in terms of value.”

Third: alternative procurement pathways. “Utilities will have to think of procuring LDES outside of just pure capacity to help them build reasonable amounts of resources under their belt before they go all in on long-duration. Walk before you run, in a way.”

Fourth: removing or substantially raising administrative price caps. “Make the collar symmetric at a materially higher ceiling, or remove the requirement completely so that the price signal serves to drive new entry of resources.”

Fifth: accreditation discipline. “We can’t have accreditation moving around every year or every other year. Developers already deal with hundreds of variables that they have to de-risk over a project that’s under development for several years. To the extent that we can have more disciplined accreditation modeling that helps developers get a view of what their asset is worth for the next five years, that is going to be super helpful in underwriting risk.”

Unpriced risks and market evolution

Khan offers several contrarian perspectives that challenge prevailing industry assumptions and reveal how capital is repricing battery storage risk.

“The cost decline assumption is the largest unpriced risk,” Khan states. “Projects with commercial operation dates between 2027 and 2029 priced off continuing deflation of cells and transformers—that is likely not going to be true.”

This assumption requires believing that lithium supply is fully qualified and quantified domestically, that past supply shock issues are permanently resolved, that there’s consensus on supply availability timing, and that transformer lead times will dramatically improve.

“Developers were thinking they’re going to ride the cost curve further from now until 2030. I think they’re going to be in for a little bit of a price shock,” Khan warns.

He also gives his outlook on Li-ion pricing through 2030, “Pricing for Li-ion is not going to go down much further. I think it’s going to saturate, and we’re going to see blips of higher pricing between now and 2030.”

Policy changes in China are acting as the main driver. “Local policy in China means these manufacturers are going to be slowly weaned off state subsidies, and there’ll be more of an open market environment. We will see a lot of consolidation in the space, and pricing will settle as a result.”

Merchant batteries

Khan states about the foundation of merchant battery economics, “Merchant batteries are structurally a self-cannibalising business model. The biggest competition to BESS is BESS.”

The data supports this view. “Merchant value in ERCOT has gone down by roughly 90% between 2023 and 2025,” Khan notes. Yet paradoxically, “in the same period, fair market value of ERCOT assets has gone up.”

“Capital is repricing merchant exposure downward, and simultaneously, contracted exposure is going upward. This is what definancialisation looks like in progress.”

Khan argues the winners will be “those selling contracted availability to a single counterparty—a data centre under an energy service agreement (ESA), a utility under an integrated resource plan (IRP), or a capacity market offering long-tenor agreements—rather than those trading a merchant position.”

While merchant economics deteriorate, Khan points to optimisation capability as the source of durable value.

He cites a Gridmatic study of 30 assets in California that found revenue dispersion ranging from US$1-per-kW-month to US$6-per-kW-month. “Their conclusion was the spread is because of better bidding strategy, not because of equipment, not because of construction quality, not because of who was underwriting the asset. That spread is direct evidence of optimisation.”

This creates an interesting market dynamic. “The industry thinks that optimisation software is going to be the durable value capturer,” Khan observes. Yet this exists in tension with the merchant cannibalization thesis. Optimisation can capture more value from a shrinking pool, but cannot prevent the pool from shrinking as more batteries enter the market.

“We’re seeing two different themes play out between ERCOT and CAISO, and they’re both on opposite ends of the spectrum,” Khan concludes suggesting the market is still working out which model will ultimately prevail.

Khan will be speaking on a resource adequacy and LDES panel at Battery Asset Management Summit USA 2026 with participants Mitchell Bauer, president at developer GCI; Jacob Steubing CCO at IPP Linea Energy; and moderated by Lukasz Cianciara, founder and operating partner at advisory and management firm Silmera. He will also be participating in a panel at the Solar & Storage Finance Summit, taking place at the same location.

Battery Asset Management Summit USA 2026 will be held 15-16 September in Garden Grove, California, hosted by Energy-Storage.news publisher Solar Media (part of the Informa Group). The agenda emphasises addressing the roles of AI, cybersecurity, and second-life applications, broken down into two tracks: Technical Asset Management and Commercial Asset Management. This year, the conference is also co-located with Solar & Storage Finance Summit USA. Visit the official site for more details.