Meta Signs 100 GWh Supply Agreement for Noon Energy’s Ultra-Long-Duration Storage

Why It Matters

The agreement gives Meta a reliable, low‑material‑risk power buffer for AI workloads, while showcasing a scalable alternative to lithium‑ion storage for the broader hyperscale data‑center market.

Key Takeaways

• Meta reserves up to 1 GW/100 GWh ultra‑long‑duration storage
• Noon Energy’s solid‑oxide system can store energy for 100+ hours
• Pilot starts with 25 MW/2.5 GWh, slated for 2028 completion
• Technology uses carbon, oxygen, avoiding scarce lithium‑ion materials
• Improves data‑center resilience amid renewable integration pressures

Pulse Analysis

The surge in artificial‑intelligence workloads is forcing hyperscale operators to rethink power strategies. Traditional lithium‑ion batteries, limited to four‑hour discharge windows, cannot economically sustain the multi‑day demand spikes that AI training and inference generate. Meta’s commitment to Noon Energy signals a shift toward ultra‑long‑duration storage (ULDS), a technology that can bridge the gap between intermittent renewables and the constant power needs of data centers, reducing reliance on diesel generators and curbing carbon footprints.

Noon Energy’s approach hinges on a reversible solid‑oxide fuel cell that stores electricity in a carbon‑based medium. This design decouples power (MW) from energy (MWh) by adding storage tanks, allowing facilities to scale capacity without proportionally increasing power output. Because the system relies on abundant elements—primarily carbon and oxygen—it sidesteps the supply‑chain volatility tied to lithium, cobalt, and nickel. The pilot’s 100‑hour endurance, demonstrated in a containerized format, showcases the feasibility of multi‑day storage at scale, positioning Noon as a potential cornerstone of the emerging ULDS supply chain.

For the industry, Meta’s 100‑plus hour storage commitment underscores the commercial viability of ULDS for mission‑critical infrastructure. Data‑center operators facing stricter sustainability mandates and grid reliability concerns can look to such technologies to smooth renewable intermittency while maintaining service‑level agreements. As more hyperscalers adopt similar solutions, economies of scale could drive down costs, accelerate deployment, and reshape the energy‑storage market away from metal‑intensive batteries toward more sustainable, long‑duration alternatives. Meta’s early move may give it a competitive edge in AI compute reliability and operational cost efficiency.