Semiconductors are out, software is in. Why that’s not about to change anytime soon

Workday fell about 40% in Q1 2026 before rising more than 55% in Q3; it gained over 2% this week despite a broader tech sell-off. Over three months, the software ETF IGV climbed more than 13%, while the semiconductor ETF SOXX fell more than 4%, reflecting a rotation toward software and AI applications. Investors remain wary of an AI capex slowdown, and a Wells Fargo strategist said a 5%–10% market drawdown could occur as expectations reset.

Analysis

The key question is not whether agents replace screens, but who controls the trusted data, permissions and workflows agents need. That favors platforms that can remain the system of record or orchestrate work across tools; it is less clearly positive for vendors whose economics depend on human seats or familiar interfaces. Workday’s HR/finance domain may be sticky, but its recent rebound raises the bar: AI features must support retention, pricing or expansion, not merely reassure investors that the product remains relevant. ServiceNow’s ability to coordinate workflows across applications is a differentiated exposure; Microsoft has a distribution advantage through existing work surfaces. These are conditional advantages, not proof of incremental AI revenue.

Near term, software’s relative strength versus semiconductors can persist as portfolios rotate away from capex beneficiaries, but the move is vulnerable to earnings disappointments or renewed AI infrastructure spending. Over 1–3 months, watch earnings commentary for software net retention, seat/pricing trends and measurable AI attach, alongside hyperscaler capex guidance. Over 6–18 months, a peak in AI build-out could compress semiconductor expectations while also reducing demand for software if customers defer budgets. The contrarian risk is that investors are treating “applications win” as a clean rotation: AI can expand software utility while simultaneously weakening per-seat monetization.

Workday’s cited private-market discount is a manager’s assessment, not independently established valuation evidence. After the sharp recovery, avoid extrapolating the rebound without checking forward revenue multiples and estimates. A broad software-over-semiconductor expression is cleaner than assuming every application vendor wins; falsify it if software estimate revisions weaken or the SOXX/IGV relative trend reverses.

Key Decisions for Investors

• Consider a staged long IGV / short SOXX relative-value position over 1–3 months rather than chasing the semiconductor selloff outright. Keep risk defined: reduce or exit if IGV underperforms SOXX persistently or software earnings revisions turn negative; the pair is exposed to a renewed capex-led semiconductor rally.
• For single names, prefer NOW over WDAY as an AI-workflow expression, but size modestly until evidence shows AI-related customer expansion or durable retention. Reassess on earnings disclosures about workflow demand and monetization; do not rely on product announcements alone.
• Treat WDAY as a watch/hold-on-confirmation candidate, not an automatic value buy after its rebound. Verify current valuation versus growth, net retention, seat/pricing trends and AI-related bookings before adding; thesis weakens if guidance or customer metrics deteriorate.
• Monitor hyperscaler capex guidance and software earnings revisions over the next quarter. A capex slowdown without corresponding application revenue growth would undermine the rotation thesis and favor reducing the software-over-semiconductor pair.