Goldman Sachs Forecasts ~$3.8 Trillion in AI Infrastructure Spending Tied to $NVDA, $AVGO, and $AMD Systems...

Goldman Sachs projects roughly $3.8 trillion in artificial intelligence infrastructure spending tied to systems from $NVDA, $AVGO, and $AMD through 2028.
Based on the breakdown of capital expenditure and power capacity:
- $NVDA: ~$2.7 trillion (63% of CapEx, 47% of gigawatts)
- $AVGO: ~$876 billion (21% of CapEx, 31% of gigawatts)
- $AMD: ~$228 billion (5% of CapEx, 6% of gigawatts)
While $NVDA continues to capture the clear majority of total dollar spend, the divergence between capital expenditure share and deployed power capacity underscores the rapid adoption of custom AI silicon—including solutions co-designed with $AVGO—from major hyperscalers $META, $GOOGL, $AMZN, and $MSFT.