AI Everywhere, Everything

The AI investment boom is now broad and large enough to be a dominant force for equity markets, bond yields and global economic growth, but its eventual slowdown could deflate the bubble.

• The infrastructure buildout is already comparable in scale to historic investment waves such as the late-19th-century railway expansion, although it is only a few years old versus two decades for the railways.
• AI exposure extends beyond chips, memory, optics, storage and LLM providers to industrial companies, utilities and national equity markets; the 10 largest S&P 500 constituents represent more than 40% of the index and are all AI plays.
• U.S. investment-grade bonds linked to AI megacaps and infrastructure providers rose from roughly 7–8% of the universe to 12.1% over the past year, with the largest increases coming from Alphabet, Amazon, Meta, SpaceX and Oracle.
• Data-center investment is lifting U.S. growth: fixed investment in information processing and intellectual property, the GDP components most affected by data centers, is contributing almost 1.0% to real GDP versus a 30-year average of about 0.4%.
• Hyperscaler capex is estimated at $365 billion in 2025, $762 billion in 2026 and $1.125 trillion in 2027; although spending keeps rising, the implied growth rate falls from 108% to 47%, making the change in momentum a key bubble risk.

Takeaway: Investors should map their direct and indirect AI exposure and preserve diversification, while monitoring earnings versus expectations, economic weakness, yields, hyperscaler capex growth, credit-default swaps and the market’s ability to absorb future mega-IPOs.