MSFT & Meta Add $100B to AI Data Center Surge

🔑 Enhanced Key Takeaways

• •Meta's $600 billion AI infrastructure commitment by 2028 is the largest single corporate capex plan in history, including the 2,250-acre Hyperion campus in Louisiana delivering 5 gigawatts of compute power anchored by nuclear arrangements[1].
• •Meta has purchased over 1.3 million GPUs from Nvidia, AMD, and Google TPUs, and secured $27 billion in funding from Blue Owl Capital for data center construction[1].
• •Amazon leads 2026 capex with $200 billion, followed by Google at $175-185 billion, contributing to hyperscalers' collective $650-700 billion data center spend[2][4][5].
• •Microsoft's AI data center investments tie closely to OpenAI, with cumulative investments over $13 billion since 2019 and OpenAI representing 45% of its cloud backlog[4][6].

🛠️ Technical Deep Dive

• •Meta's Hyperion campus: 2,250 acres in Louisiana, $10B buildout, 5GW compute power with nuclear power deal[1].
• •Prometheus facility in Ohio: Expected online in 2026[1].
• •Meta GPU procurement: >1.3M units from Nvidia, AMD, Google TPUs[1].
• •New AI data centers: Supersized for high energy/water use, designed for AI training with vast GPU clusters[7].

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline