> ## Content Index > Fetch the complete content index at: https://nexi.fund/llms.txt > Use this file to discover other available public pages before exploring further. # Noon Energy booked 1.1 GW of storage in four months. AI data centers are next - URL: https://nexi.fund/noon-energy-ultra-long-storage-ai-2026/ - Published: 2026-08-16T12:00:02.000Z - Updated: 2026-08-16T12:00:02.000Z - Description: The reversible solid oxide fuel cell startup booked a 1 GW joint venture with Sabanci Renewables for AI data center power, on top of April's Meta reservation. What has to go right for the chemistry to keep the promise. - Author: Nexi.fund Labs - Tags: Energy & Climate, #mode-1, #hook-statistic, #track-E Noon Energy has signed deals for roughly 1,100 megawatts of long-duration storage in four months: with Meta in April, and now with Sabanci Renewables for a 1 GW joint venture. Its only product is a battery that runs for more than 100 hours. ๐ŸŽฏ Noon Energy's reversible solid oxide fuel cell stores 100+ hours of electricity using carbon and oxygen, not lithium. Sabanci Renewables will co-develop up to 1 GW of ultra-long-duration storage for AI data centers, with commercial deployment possible from 2027. Meta has already reserved up to 100 GWh under a phased deal that starts with a 25 MW project due in 2028. The open question: no storage project with this chemistry has ever run a full data center through a Texas summer. 1 GW ultra-LDES contracted #### Sabanci Renewables JV capacity Signed 13 August 2026\. Noon's reversible electrofuels batteries paired with Sabanci's solar and wind portfolio across ERCOT. ยท *Energy-Storage.News, 2026* 100+ hours storage duration claimed #### Multi-day storage duration The demo unit ran for thousands of hours and held over 200 hours of storage. Lithium cells typically discharge in 4โ€“8 hours. ยท *pv magazine USA, 2026* ## A third of the pipeline, one startup Sabanci Renewables, the renewable arm of the Turkish industrial group, is betting a third of its US pipeline on a startup that has raised just $45 million. The joint venture announced on 13 August orders up to 1 GW of multi-day storage to back AI data center load. Sabanci operates and is building four utility-scale solar plants in the ERCOT market with a combined 790 MWdc of capacity. It is targeting a 3 GW US pipeline over the next five years. The storage joint venture plugs its batteries under that solar, and the two will sell the combination as power purchase agreements or capacity offtake. No project site has been named. That detail matters more than most. ## A battery that runs on air Noon's system is a reversible solid oxide fuel cell. Charge: electricity splits carbon dioxide into solid carbon and oxygen. Discharge: the fuel cell recombines the oxygen back with stored carbon, releasing electricity. Power and energy are decoupled: a power block plus bolt-on tanks of cheap CO2. Everything fits in standard shipping containers. That is the design advantage it presses hardest: "the first fully containerized, modular multi-day storage system to operate for thousands of hours," CEO Chris Graves told Latitude Media in January. Lithium-ion stores 4โ€“8 hours. Form Energy's iron-air batteries target 100 hours. The startup claims 100+ hours with a smaller footprint and, it says, lower cost per unit of energy. ## Why hyperscalers are signing now AI data centers need firm power around the clock, and grid interconnection queues run years long. Renewable generation alone is intermittent. Multi-day storage is the piece that turns "cheap solar" into "always-ready power." Meta was first. In April it reserved up to 100 GWh of Noon capacity under a phased deal beginning with a 25 MW / 2.5 GWh project scheduled for completion by 2028\. Meta's VP of Energy, Nat Sahlstrom, said the agreement "advances that goal with a storage technology that delivers grid resilience and firm power." The Sabanci deal adds a second anchor, and this one owns solar. That is the difference between a pilot and a pipeline. ## The race: three chemistries, one constraint Noon is not alone in chasing the same 100-hour prize. Form Energy raised $750M and started deploying iron-air batteries in October. Antora raised $550M in August for carbon-block thermal storage. Energy Vault signed a 1.25 GW AI data center deal in Texas. The Sabanci tie-up lands Noon at 1 GW before Form Energy's biggest contract. All of them share the same constraint: none has proven multi-day storage economics against a full, hot-climate data center load. The validation window opens in Texas, and everyone is racing toward it. ## What "cents per kWh" actually requires Graves frames the pitch in unit-economics terms: 24/7 renewable power "for cents per kWh." That only happens if three things hold at once. The battery must hit nameplate round-trip efficiency for years. The CO2 tanks must stay cheap at GWh scale. And Sabanci's solar must generate surplus power during off-peak hours to charge them. ERCOT helps: no capacity market inertia, open prices, and a grid that already leans on scarcity. It also punishes the unprepared: storage in Texas means locking in revenue against a volatile real-time market, not a contract. Noon has raised more than $45 million from At One Ventures, Emerson Collective, Clean Energy Ventures, Aramco Ventures, Prime Impact Fund, Elemental Impact, Sabanci Climate Ventures, D3 Jubilee and the California Energy Commission. That is a seed-stage stack for a company selling gigawatts. ## The part press releases skip Long-duration storage has a public-relations history longer than its delivery record. Every chemistry has slipped a timeline. Sourcing, permitting and grid interconnection are all multi-year gates that no press release can close. Then there is the physical question: a solid oxide fuel cell running continuously at high temperature, cased in a shipping container next to a data center substation, for a decade. The demo ran for months. A project runs for years. Those are not the same sentence. ### What makes sense for a 1 GW AI buildout by 2028? ๐Ÿ”ฎ **Noon Energy books 500+ MW of firm offtake with at least one hyperscaler or utility by end-2027.** Probability: 55%. Two anchor customers are already signed (Meta, Sabanci), and the 25 MW pilot gives a 2028 test window. #### โœ… Arguments for Two anchor buyers already committed Containerized design shortens site construction cycles ERCOT's price structure rewards multi-day capacity **Confirmation criteria:** a named project reaching financial close. #### โŒ Arguments against No site, no financial close, no nameplate year-1 data $45M raised against a gigawatt-scale supply chain Form Energy and Energy Vault are further along in deployment **Disconfirmation criteria:** the Sabanci JV gone silent for two straight quarters. ### Development scenarios #### ๐ŸŸข Optimistic scenario (25%) The 25 MW Meta pilot clears interconnection early, the Sabanci JV names a Texas site this year, and Noon converts capacity reservations into manufacturing orders. **Implications:** multi-day storage becomes a financeable asset class, and the Sabanci pipeline pulls 3 GW of firm-power projects forward. #### ๐ŸŸก Base case (55%) Noon delivers the 2.5 GWh pilot by 2028, proves round-trip economics at postage-stamp scale, and signs one more offtaker. Deployment stays project-by-project into 2029. **Implications:** Noon survives as a niche firm-power supplier; hyperscaler PPAs remain the deciding market signal. #### ๐Ÿ”ด Pessimistic scenario (20%) Solid oxide degradation cuts round-trip efficiency below contract inflation, or the capital gap between $45M and a gigawatt supply chain stalls the Sabanci JV. Form Energy and Energy Vault consolidate the market first. **Implications:** the tech survives in labs, ERCOT keeps buying 4-hour lithium, and Noon re-routes toward industrial heat. ๐Ÿ“Š **Key signals to track** Sabanci JV names a Texas project site and reaches financial close. Meta pilot clears interconnection and hits 2028 schedule. Noon raises a Series B for manufacturing capacity. Form Energy signs a deal larger than 1 GW. As we wrote in August, Form Energy's $750M iron-air bet is the battery the grid wrote off. Antora's $550M carbon-block round came the same week as the Sabanci JV. Multi-day storage has gone from a niche curiosity to the most crowded trade in the energy transition. The question is no longer whether 100-hour storage works. It is whether the first mover to prove it at data-center scale gets the whole market. [ Noon Energy makes 1GW US AI infrastructure agreement Primary account of the 13 August Sabanci Renewables joint venture, the 1 GW capacity target and the December deal structure. Energy-Storage.News ](https://www.energy-storage.news/multi-day-energy-storage-startup-noon-energy-makes-1gw-us-ai-infrastructure-deployment-agreement/?ref=nexi.fund) The most complete industry-press account of the Sabanci JV structure. [ Noon Energy, Sabanci Renewables ink JV for 1 GW long-duration storage for AI data centers Adds the technology detail: 100+ hour reversible solid oxide fuel cell systems and up to 100 GWh of storage potential. pv magazine USA ](https://pv-magazine-usa.com/2026/08/14/noon-energy-sabanci-renewables-ink-jv-for-1-gw-long-duration-storage-for-ai-data-centers?ref=nexi.fund) Confirms both the chemistry and the commercial deployment timeline starting 2027. [ Noon Energy partners with Sabanci Renewables for up to 1GW of co-located storage Data center trade press framing of the deal, including reference to the April Meta reservation and the 25 MW project due 2028. DatacenterDynamics ](https://www.datacenterdynamics.com/en/news/noon-energy-partners-with-sabanci-renewables-for-up-to-1gw-of-co-located-long-duration-storage-and-renewable-capacity/?ref=nexi.fund) Corroborates the Meta precedent and the pilot-scale first step of the Sabanci venture.