Tech Stocks: Echoes of 1998

The article argues that the AI-led Nasdaq rally is tracking the late-1990s internet boom almost identically over roughly 1,000 days, implying the sector remains in the early stages of a generational bull market despite interim pullbacks. It cites Zacks consensus estimates for SanDisk (SNDK) to deliver 3,689.34% earnings growth next quarter, supporting its view that AI-linked earnings momentum is accelerating. The comparison is technical and historical rather than a broad market forecast, limiting near-term market impact.

Analysis

The relevant signal is not the historical chart overlay but the widening gap between AI infrastructure earnings visibility and the promotional quality of marginal beneficiaries. SNDK's implied earnings growth is likely dominated by a depressed comparison base and memory-cycle pricing; the key underwriting variables are NAND contract-price durability, enterprise SSD mix, and gross-margin conversion rather than the headline percentage. If hyperscaler capex remains elevated through the next two reporting cycles, storage suppliers can see disproportionate estimate revisions because operating leverage is high; if capex shifts from training to inference efficiency, commodity memory demand may not support current expectations.

Near term, bullish retail analogy narratives tend to concentrate flows into high-beta, low-liquidity AI proxies such as QBTS, where valuation is detached from a demonstrable revenue ramp. That creates a reflexive upside window measured in days to weeks, but also makes these names the first source of liquidity in any factor unwind. AMZN is a cleaner second-order beneficiary than the speculative cohort: AWS can monetize both model usage and enterprise workload migration, while its retail cash generation absorbs investment volatility; the risk is that AI capex depreciation arrives ahead of incremental AWS revenue, delaying margin expansion over 1-3 quarters.

The contrarian view is that a continuation of the AI cycle need not validate current broad AI equity multiples. The internet precedent ultimately rewarded distribution, cloud-like utility economics, and durable platforms—not every infrastructure vendor or concept stock. A 6-18 month winner set is therefore likely to narrow toward firms with contracted demand, pricing power, and balance-sheet capacity, while crowded unprofitable AI exposures face multiple compression even in a constructive Nasdaq tape.

Key Decisions for Investors

• Initiate a 3-6 month pair: long SNDK / short QBTS, sized beta-neutral. The trade isolates tangible memory-cycle and AI-storage earnings revision potential from sentiment-driven quantum exposure; review after SNDK's next earnings report. Exit if NAND pricing weakens for two consecutive monthly industry checks or QBTS announces independently verifiable commercial bookings sufficient to reset its revenue path.
• Accumulate AMZN on broad AI-led drawdowns rather than chase momentum; target a 6-12 month holding period. Underwrite AWS growth reacceleration and incremental operating leverage, but cut exposure if AWS growth decelerates sequentially while capex/depreciation meaningfully exceeds consensus, as that would undermine the AI monetization case.
• Avoid adding outright QBTS exposure after sharp momentum moves; treat it as a tactical flow vehicle only, with a 2-4 week horizon and hard stop near 15% below entry. There is no disclosed earnings anchor in the supplied data to support a strategic valuation case.
• Set an earnings-quality alert for SNDK: favor adds only if management confirms enterprise SSD demand, NAND pricing, and gross-margin progression—not merely an elevated next-quarter EPS growth rate. A guidance raise driven by mix and pricing offers materially better risk/reward than a beat driven by cost timing or a low base.