Ultra-long-duration storage developer Noon Energy and Sabanci Renewables — the clean-energy arm of Turkish conglomerate Sabanci Holding — announced on August 13, 2026 a strategic agreement to deploy up to 1 GW (100 GWh) of ultra-LDES capacity powering AI infrastructure with 100% renewable, around-the-clock electricity. The parties will jointly develop projects integrating Noon’s reversible electrofuel battery, structured as power purchase agreements or capacity offtakes, with the first commercial deployments possible as early as 2027. Noon’s technology replaces lithium with abundant carbon and oxygen chemistry to deliver more than 100 hours of storage. Sabanci Renewables currently operates 790 MWdc of solar in ERCOT and plans to grow its U.S. portfolio to 3 GW within five years — the anchor market for this partnership. The deal underscores how the home battery cost per kWh equation is being rewritten for the multi-day storage applications that AI data centers demand.
Overview of the Technology / News
The announcement is significant for two reasons: the customer and the duration. The customer is the hyperscale AI sector, whose data centers require firm, dispatchable, carbon-free power 24 hours a day — a load profile that four-hour lithium-ion batteries fundamentally cannot serve alone. The duration is 100+ hours, a category known as ultra-long-duration energy storage (ultra-LDES) that spans the multi-day gap between renewable abundance and scarcity, far beyond the daily cycling that defines conventional batteries.
Noon Energy’s technology is an "electrofuel" battery: it reversibly converts energy into and out of a storable fuel using carbon and oxygen, two of the most abundant elements on Earth, rather than lithium, cobalt, or nickel. Sabanci Renewables brings the capital, the development pipeline, and its ERCOT operating footprint. Together the two are betting that AI’s insatiable, reliability-critical load is the first large commercial market for multi-day storage — a bet with implications far beyond any single project.
Why This Development Matters
AI is reshaping the economics of clean energy in real time. Hyperscalers and data-center operators have signed a wave of power deals for 24/7 carbon-free energy, and the binding constraint is no longer cost per megawatt-hour but the ability to match renewable supply to demand on an hour-by-hour, and increasingly day-by-day, basis. A lithium-ion battery that discharges in four hours cannot cover a windless, overcast stretch of several days; an ultra-LDES system designed for 100 hours can.
The strategic logic also reflects a shift in who buys storage. The first decade of utility-scale storage was driven by grid operators and utilities seeking frequency and capacity services. The next decade is being driven by corporate and industrial buyers — led by AI — who need firm clean power as a procurement requirement, not an ancillary benefit. That demand base is larger, faster-moving, and less price-sensitive to duration than any regulator-driven market that came before.
Technical Deep Dive
The core engineering trade-off in energy storage is between power, energy, and cost, and every chemistry chooses two of the three. Lithium-ion is a power-dense, energy-dense, but material-expensive chemistry: its active materials and manufacturing set a floor on the home battery cost per kWh that has fallen dramatically but cannot reach the sub-$20/kWh territory that multi-day storage requires. Noon’s electrofuel approach attacks the cost side by storing energy in the bonds of an abundant carbon-based fuel, decoupling the energy capacity — which scales with cheap tankage — from the power capacity — which scales with the fuel cell stack.
This decoupling is the single most important concept in long-duration storage. In a lithium-ion system, adding energy means adding cells, so cost scales roughly linearly with duration. In an electrofuel or flow system, energy is stored in inexpensive bulk medium, so extending from 10 hours to 100 hours adds little marginal cost — only more tank volume. That is why a 100-hour lithium battery is economically absurd, while a 100-hour electrofuel or iron-air system is the entire point. The efficiency trade-off — electrofuel and iron-air systems run at roughly 50-60% round-trip efficiency versus 90%+ for lithium — is acceptable for infrequent, multi-day discharges where capex per kWh dominates the levelized cost.
For AI data centers specifically, the reliability math is unforgiving. A 99.99% uptime requirement means the power supply must ride through not just daily peaks but multi-day renewable lulls. That is why hyperscalers are stacking complementary assets: lithium-ion for sub-daily balancing, ultra-LDES for multi-day resilience, and firm generation (gas, nuclear, or geothermal) as a last-resort floor. The same "right battery for the right duration" logic applies at the household scale, where off-grid battery system sizing distinguishes a daily-cycling home battery from a multi-day backup need.
Real-world Applications
The headline application is 24/7 carbon-free power for AI data centers, a market that has grown from near-zero to a defining force in electricity demand within a few years. ERCOT, where Sabanci Renewables operates, is a leading venue because Texas has become a hub for both data-center development and cheap, abundant renewable generation — though its grid is notoriously short on multi-day firming. An ultra-LDES fleet in ERCOT could shift weekend wind and solar surpluses into weekday AI loads.
The applications extend well beyond AI. Island grids and remote mining operations that currently burn diesel for days at a time, industrial facilities needing resilience against multi-day outages, and transmission-constrained regions seeking to defer new lines are all natural customers for 100-hour storage. As the technology proves itself in the demanding AI market, expect it to cascade into these adjacent sectors — much as lithium-ion matured in consumer electronics and EVs before conquering the grid.
Industry Impact / Market Implications
The partnership is a direct competitive signal in the long-duration race. Noon Energy is now positioned against Form Energy (iron-air), which recently raised a $750 million Series G, as well as flow-battery players like Invinity (vanadium redox) and ESS Inc. (iron flow), and compressed-air and liquid-air systems from Hydrostor and Highview Power. Each occupies a slightly different niche, but the common thread is a land-grab for the multi-day storage market that BloombergNEF projects to exceed 100 GW by 2040.
The AI-driven demand surge also has a secondary effect on conventional storage. As hyperscalers sign offtakes for ultra-LDES, they simultaneously pull forward lithium-ion deployments for daily balancing, tightening the cell supply chain and reinforcing the learning curve that keeps lowering home battery cost per kWh. The two storage tiers are complements, not substitutes — a point that is easy to miss but central to understanding where the market is heading.
Future Outlook
The near-term milestone is unambiguous: prove that an electrofuel system can be banked and delivered at commercial scale by the 2027 target. That requires the technology to clear the same hurdles Form Energy is now navigating — manufacturing ramp, demonstration of multi-day dispatch, and, most importantly, securing project financing from lenders who understand an asset that cycles only a handful of times per month.
Looking to 2028-2030, the likely outcome is a differentiated storage stack: lithium-ion for hours, iron-air and electrofuel for days, and hydrogen or pumped hydro for seasonal needs. AI’s appetite for 24/7 clean power is the catalyst accelerating this stratification. The strategic lesson for the broader market — from utility planners to a homeowner comparing home battery cost per kWh — is that duration is now a first-class design variable, and the cheapest battery is the one matched correctly to the length of the gap it must bridge.
