🔥Stalecollected in 5m

Nscale-Microsoft 1.35GW AI Compute Deal

Nscale-Microsoft 1.35GW AI Compute Deal
PostLinkedIn
🔥Read original on 36氪

💡1.35GW MSFT deal + Rubin flagship boosts AI infra capacity

⚡ 30-Second TL;DR

What Changed

LOI with Microsoft for 1.35GW AI compute delivery from H2 2027

Why It Matters

Scales AI training infrastructure massively, enabling next-gen GPU workloads for hyperscalers like Microsoft.

What To Do Next

Plan GPU cluster migrations to microgrid-backed sites like Monarch for 2027 Rubin readiness.

Who should care:Enterprise & Security Teams

🧠 Deep Insight

Web-grounded analysis with 9 cited sources.

🔑 Enhanced Key Takeaways

  • The deal involves approximately 430,000 NVIDIA Vera Rubin GPUs, structured as a multi-year compute services agreement combined with a long-term data center lease[6].
  • Nscale is deploying Caterpillar G3500 series natural gas generator sets to achieve 2GW of power generation by H1 2028, powering the NVIDIA Vera Rubin DSX AI Factory reference design[2][4].
  • Announced at NVIDIA's GTC 2026 conference on March 16, 2026, establishing Nscale as the global flagship deployment partner for NVIDIA Vera Rubin architecture[4][5][6].
  • The Monarch site in Mason County offers proximity to Ashburn, Virginia, and Chicago for low-latency AI workloads connected to major cloud hubs[6].
  • Nscale establishes Nscale Energy & Power division and acquires American Intelligence & Power Corporation as part of the Monarch campus deal[8].

🛠️ Technical Deep Dive

  • NVIDIA Vera Rubin NVL72 GPUs deployed in systems engineered with the NVIDIA Vera Rubin DSX AI Factory reference design[2][4][5].
  • Caterpillar G3500 series natural gas generator sets for onsite power generation, scaling to 2GW by H1 2028[2].
  • Microgrid supports up to 8GW+ onsite power potential, certified by West Virginia state as first AI microgrid[2][4][5][6].

🔮 Future ImplicationsAI analysis grounded in cited sources

Monarch campus will expand Nscale's total capacity beyond 2.35GW
It builds on Nscale's existing over 1GW operational capacity with the new 1.35GW Microsoft deployment[4][5].
Deal addresses AI data center power constraints through independent microgrid
Onsite generation via Caterpillar generators scales to 8GW potential amid McKinsey's forecast of 156GW global AI demand by 2030[2][4].
Positions Nscale as key US AI infrastructure player
Combines acquisition, LOI, and flagship NVIDIA deployment to create one of the largest dedicated AI compute installations[2][3][6].

Timeline

2026-03
Nscale announces LOI with Microsoft for 1.35GW AI compute at GTC 2026
2026-03
Nscale acquires Monarch Compute Campus and American Intelligence & Power Corporation
📰

Weekly AI Recap

Read this week's curated digest of top AI events →

👉Related Updates

AI-curated news aggregator. All content rights belong to original publishers.
Original source: 36氪