Lisa Su on the 10-Year AI Supercycle: 3 Key Companies Leading the Charge

The AI Supercycle Begins

• 💡 Lisa Su, AMD CEO, predicts a 10-year AI supercycle, noting the AI Total Addressable Market (TAM) for silicon accelerators may be underestimated, potentially exceeding $500 billion.
• 🚀 The AI revolution is a hardware arms race, driving the creation of new digital infrastructure including data centers, power grids, and semiconductor supply chains.
• ⚠️ Demand for AI compute is so massive that neither Nvidia alone nor Nvidia and AMD combined can meet it, creating significant opportunities for other industry players.

TSMC: The Invisible AI Foundation

• 🎯 Taiwan Semiconductor Manufacturing Company (TSMC) manufactures over 90% of the world's most advanced semiconductors, including those for AI servers, iPhones, and autonomous vehicles.
• 📈 The company reported 54% year-over-year revenue growth in Q2 2025 and is a dividend compounder with over 15% annual growth.
• 🔬 TSMC's upcoming 2-nanometer process technology in 2026 will redefine efficiency for AI chips, while overseas fabs in Arizona and Japan diversify production and reduce regional risk.

Micron: Powering AI Memory

• 🧠 Micron Technology provides crucial High Bandwidth Memory (HBM) and NAND flash chips, which are essential for AI models to run, train, and scale efficiently.
• ✅ Fiscal year 2025 results showed a 49% revenue increase and 41% margins, with its DRAM business (including HBM) growing an incredible 69% year-over-year.
• 💰 A global NAND shortage is creating a new tailwind, leading to stronger pricing power and higher margins for Micron, with next-generation HBM4 chips launching in 2026.

AMD: Challenging AI Dominance

• ⚡ Advanced Micro Devices (AMD) has solidified its position as a top-tier AI powerhouse through a landmark 6-gigawatt AI infrastructure collaboration with OpenAI.
• 🤝 This strategic alliance involves OpenAI holding warrants for AMD shares, aligning their long-term growth interests.
• 🛠️ AMD's MI 350 and MI450 accelerators are designed to compete directly with Nvidia's GPUs, and the company's integration of 2-nanometer technology enhances energy efficiency and scalability.
• 🌐 AMD partners with major cloud providers like Microsoft Azure, Google Cloud, and Amazon AWS, diversifying its client base and securing its role in AI infrastructure.