Cramer defends a struggling tech giant and says another stock could soon be a buy

Stocks rose Thursday despite renewed U.S.-Iran tensions and higher oil prices, but tech showed a sharp split: Micron surged 7.5% while Salesforce fell 2.5% after KeyBanc downgraded it to hold. Jim Cramer framed the AI buildout divide as hyperscalers writing the checks (software/hyperscaler beneficiaries) versus suppliers benefiting (e.g., Micron/SanDisk). Meta also pulled back as investors focused on accelerating AI spending tied to manufacturing its own AI chip and expanding compute capacity to 14 gigawatts next year.

Analysis

The market is reading AI spend as a winner-take-most allocation trade, not a broad technology beta trade. That favors the suppliers of compute, memory, and infrastructure over software vendors whose near-term monetization still looks abstract; in other words, cash is moving from promises to atoms. Over the next 1-3 months, that can keep pressure on enterprise software multiples even if the broader tape firms, because investors are underwriting a longer payback period for AI-related opex.

META is the most important second-order signal here: rising capex is less about bloated spending and more about a visible demand curve that justifies new infrastructure. The real beneficiaries are the adjacent supply chain pieces the article doesn’t name directly — power, cooling, foundry, advanced packaging, and memory — which should continue to outperform if capex guidance stays elevated. The risk is that if ad growth or ROI commentary softens, the market will stop treating capex as investment and start treating it as margin leakage.

The aerospace weakness looks more like mechanical post-spin flow than fundamental deterioration, which often creates a better entry point than chasing strength. That dislocation can persist for weeks as passive funds and arbitrage accounts rebalance, but over 6-18 months the standalone asset should trade on aftermarket/service durability rather than the parent’s conglomerate discount. Contrarian take: consensus may be overreacting to software weakness and underpricing how durable AI capex can be if hyperscalers are still seeing constrained demand; the thesis breaks if any major spender signals a capex pause or if software vendors show clear AI-driven revenue acceleration.

Key Decisions for Investors

• Pair trade: long INTC / short CRM for 1-3 months. Expresses the hardware-vs-software capex split with cleaner relative fundamentals. Add on CRM strength or if software sentiment catches a bid; stop if CRM guides higher AI monetization or INTC loses follow-through into the next earnings window.
• Overweight semis and memory via SOXX or SMH rather than single-name software exposure. Best risk/reward is to own the infrastructure basket on pullbacks; the thesis is invalidated if hyperscaler capex commentary rolls over or if semis fail to hold relative strength after the next spend update.
• Tactical long HONA on further post-spin weakness, sized as a technical dislocation trade rather than a conviction fundamental long. Use a staggered entry over the next 2-4 weeks; cut if aerospace backlog or margins deteriorate, because that would turn a flow-driven dip into a real thesis break.
• Reduce or hedge CRM into any bounce until the market gets clearer evidence that AI spend is translating into enterprise software monetization. A simple stop is a sustained move back above the pre-downgrade trading range on improving guidance.
• No actionable trade in PEP or LEVI absent a new catalyst; treat them as watch items only.