This is a scarcity compounder, not a moonshot. TSMC should stay central to AI infrastructure because it owns the hardest manufacturing and packaging bottlenecks, but its size, capital intensity, and geopolitical
discount limit how far the multiple can expand. I expect strong revenue compounding, mild multiple compression from scale, and still enough earnings-power growth to roughly double
enterprise value over five years.