For years, investors assumed AI infrastructure spending was almost unlimited. Cheap financing funded data centers, venture capital backed compute providers, and semiconductor companies enjoyed expanding orders.
That assumption is now being challenged.
I wouldn't declare an AI capex peak just yet. Computing demand remains real. What may have peaked is the willingness to finance projects regardless of their returns.
The next phase will separate companies selling essential infrastructure from those simply renting expensive capacity. Cash-rich hyperscalers could even strengthen their positions as weaker competitors struggle to refinance.
My takeaway: Don't blindly short semiconductors, but stop treating AI spending as a guaranteed rising tide. The next winners will be determined by cash flow, pricing power and capital efficiency, not simply AI exposure.
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
