AI Capex Hype Meets Sticky Old Software

AI leaders are currently promoting a narrative that traditional software is obsolete and that aggressive AI adoption is mandatory, while committing roughly $600-650 billion of AI-focused capex to data centers, chips, and infrastructure. That combination of rhetoric and spending is now a defining feature of the broader technology sector backdrop.

Historically, similar “old tech is dead” narratives around the dot-com boom and the cloud transition failed to eliminate entrenched enterprise software. Incumbents like Microsoft (MSFT), Oracle (ORCL), and SAP (SAPd) adapted, layered on new delivery models, and retained substantial profit pools even as high-flying capex beneficiaries saw far more volatile outcomes.

With AI framed as an existential shift that will automate nearly all jobs, current messaging creates a sharp contrast between AI-focused technology companies and traditional enterprise software vendors. Yet core products from Microsoft (MSFT), Oracle (ORCL), SAP (SAPd), and Adobe (ADBE) remain deeply embedded in workflows, with switching costs and process risk limiting the pace of any wholesale replacement.

In prior cycles, the capital-intensive layer of the stack attracted extreme valuations during the build-out, while “legacy” software assets were periodically discounted on fears of rapid obsolescence. When disruption timelines proved longer and more incremental, cash-generative traditional software often held or regained favor as markets reassessed the durability of installed bases and maintenance revenue.

If the present AI build-out follows that pattern, the gap between apocalyptic rhetoric about non-AI software and the slower reality of enterprise change could again become important. In such a scenario, demand for traditional and only partially AI-enhanced software at firms like Microsoft (MSFT), Oracle (ORCL), SAP (SAPd), and Adobe (ADBE) would remain resilient rather than collapsing in line with the more extreme automation claims.