IBM's miss looks company-specific, but management's commentary is more interesting than the headline. If customers are genuinely pulling forward spending on servers, storage and memory ahead of expected price increases, that suggests AI infrastructure demand remains robust rather than weakening.

I wouldn't call this a full software-to-hardware rotation yet. Enterprise IT budgets are finite, so near-term spending can temporarily favour infrastructure before shifting back to software once capacity is deployed. Companies with slower AI monetisation may also face greater scrutiny.

My view: this is a tactical rotation, not a structural one. AI hardware, especially memory and networking, could continue to outperform in the coming quarters if supply stays tight. Longer term, however, hardware spending ultimately needs software and AI applications to generate returns, so I wouldn't write off quality software names after one earnings report.