Cybersecurity and AI Chip Stocks Are Soaring, While Software Stocks Are Crashing. Blame IBM.

IBM pre-announced Q2 results showing revenue up 1% to $17.2B and adjusted EPS of $2.27, but both missed expectations (consensus $17.85B revenue and $3.02 EPS). The shortfall sent IBM shares down as much as ~26% and points to the worst day in the company’s history, while also signaling a broad capex reallocation toward AI infrastructure (servers/storage/memory) and away from parts of enterprise software. That shift boosted AI memory names (Micron +5% and Sandisk +5% intraday; Micron +694% YoY and Sandisk +3,750% YoY in the article) and cybersecurity providers (CrowdStrike/Okta up 10%+), but weighed on enterprise software like ServiceNow (-4%) and Microsoft (-1%).

Analysis

The clean read-through is not “IBM is broken,” but that enterprise budgets are rotating toward infrastructure and away from discretionary software spend. That favors the picks-and-shovels of the AI stack — especially memory/storage suppliers like MU and SNDK — because even a modest reallocation can have outsized pricing power in tight-supply components. It also keeps cybersecurity spend relatively protected versus general app/software budgets, which is why CRWD and OKTA can outperform even if broader enterprise IT spending is flat.

The second-order loser set is broader than IBM: NOW and, to a lesser extent, MSFT risk multiple compression if investors infer that CIOs are delaying new workflow and seat-expansion deals to fund AI/cyber priorities. This is a timing issue first, not necessarily a demand destruction story; the key question over the next 1-3 months is whether billings slow or merely slip a quarter. If the next round of enterprise results shows stable RPO/backlog but softer current-quarter conversion, the market should fade the initial panic.

Contrarian view: the move may be overextended if the market treats one preannouncement as a universal enterprise signal. The strongest falsifier is any confirmation that server/storage purchases were a one-time pull-forward ahead of price increases; that would mean MU/SNDK strength is front-loaded and vulnerable to a second-half air pocket. The more durable thesis requires follow-through in capex comments from other large software and infrastructure vendors over the next earnings season.

Key Decisions for Investors

• Long MU / short NOW for 1-3 months: express the budget rotation from workflow software to AI infrastructure. Risk/reward improves if NOW guidance or billings commentary softens; thesis breaks if NOW confirms no change in deal closure rates.
• Long SNDK on pullbacks, with a 6-12 month horizon: the best setup is continued inventory tightness plus pricing leverage. Falsify the trade if management signals demand was pulled forward and channel inventory rebuilds.
• Long CRWD / OKTA versus short IBM as a relative-value pair for the next earnings cycle: security remains a non-discretionary line item while IBM is more exposed to budget mix shifts. Exit if IBM stabilizes and security spend commentary normalizes.
• Avoid chasing MSFT weakness here; use it as a watch item. Revisit only if Azure/enterprise software commentary shows actual billings deterioration, not just sentiment drag from sector rotation.