NVDA, TSMC Face Semis Rout as Capital Rotates to Software

May 20
1 min read

The scale of AI infrastructure commitment has moved beyond what near-term monetization can absorb.

The return timeline is now the position.

NVIDIA is carrying a $1 trillion order book — with half that revenue concentrated in five clients whose capital expenditure has compressed their own free cash flow to near zero.

Meta has committed $200 billion to a single Louisiana site, 4,000 acres, targeting 5 gigawatts of compute capacity, against products not yet generating the returns required to service that outlay. JPMorgan Chase, by contrast, is reporting 10% to 30% productivity gains from early-generation AI tools — while deliberately maintaining vendor-agnostic infrastructure to preserve negotiating leverage across the capital stack.

The remaining three slides show how the capital formation around this buildout is being structured — and where balance sheet risk is, and is not, being absorbed.

Spending at conviction and earning at scale are not the same condition.

For portfolios carrying exposure across this infrastructure cycle: at what point does the gap between committed capital and demonstrated return become a valuation variable that cannot be deferred?

Disclaimer: This article is for informational purposes only and is not investment or professional advice. Information and views are from public sources we believe to be reliable, but we do not guarantee their accuracy or completeness. Content is subject to change. Readers should exercise their own judgment and consult a professional advisor. Any action taken is at your own risk.