Jim Cramer says investors aren't ditching tech — they just want cheaper stocks

Jim Cramer argues investors are not exiting AI/tech, but rotating out of high-multiple names as bond yields rise, favoring cheaper valuations. He contrasts MongoDB at ~52x forward earnings, down ~13% despite beat-and-upbeat guidance, versus Dell at ~16x, up ~16% on strong results. He adds Nvidia at ~17x forward earnings is viewed as cheaper despite skepticism about the durability of data center spending, implying the AI trade is intact but investors are more selective on price.

Analysis

This is a discount-rate trade disguised as an AI debate. The market is rewarding companies that can monetize AI spend now and punishing those whose valuation depends on several years of flawless execution; that favors infrastructure and hardware names with visible FCF over premium software. The second-order effect is a forced re-ranking inside tech: vendors tied to customer ROI and procurement discipline should take share from narrative-driven platforms, while adjacent high-multiple SaaS names face multiple compression even if their operating metrics are fine.

The near-term catalyst is still rates: if real yields keep rising over the next 2-6 weeks, the de-rating in long-duration tech can extend regardless of earnings beats. Over 1-3 months, the key test is whether hyperscaler and enterprise capex commentary confirms that AI budgets are expanding from experimentation into repeatable deployment; any softness there would hit NVDA first and then the broader data-center complex. Over 6-18 months, the structural winner is the company set that turns AI capex into measurable productivity, while the structural loser is the cohort that needs perpetual multiple support to justify growth.

The contrarian miss is that the selloff may be selective rather than cyclical: AI demand is probably intact, but the market is no longer subsidizing expensive duration. That means "cheap" tech can still work if fundamentals persist, while "expensive" tech can stay broken even with decent prints. The thesis fails if yields roll over sharply or if the next earnings cycle shows a real step-up in AI bookings and consumption that forces estimate revisions higher across the stack.

Key Decisions for Investors

• Long DELL / short MDB as a 4-8 week relative-value pair: express the view that AI monetization with cash flow visibility beats premium software valuation risk; cut if MDB re-accelerates net retention or DELL guide softens.
• Use NVDA as the cleaner AI expression rather than a broad-tech beta trade; buy 3-6 month call spreads on pullbacks if 10Y yields stabilize, since the market is already pricing in some skepticism on spend duration.
• Short a high-multiple software basket versus XLK or SMH for 1-3 months if real yields keep rising; this is the highest-probability way to monetize multiple compression without taking single-name blowup risk.
• Monitor hyperscaler capex commentary and enterprise AI budget revisions as the falsifier: any downgrade in next earnings season is the point to reduce AI infrastructure longs and press the short in premium software.
• If the 10Y yield breaks lower, cover valuation shorts quickly; that would likely trigger a fast factor reversal back into expensive growth and punish the rotation trade.