Dave Sekera on AI's Parabolic Run, Rotation to Value & Undervalued Sectors

Morningstar strategist Dave Sekera says investors should rotate out of tech and growth stocks into value names, arguing the AI trade is losing momentum and that memory stocks are overvalued. He specifically flags Micron (MU) and SanDisk (SNDK) as examples of the stretched memory trade. The note is likely to pressure sentiment in AI- and memory-related equities, but it is analyst commentary rather than a company-specific catalyst.

Analysis

This is less a fundamental call on AI than a positioning call on crowded duration exposure. If large-cap growth de-rates even modestly, the second-order winner is not just “value” broadly, but balance-sheet quality, cash-return names, and defensives that have lagged due to index crowding and passive flows. The real transmission mechanism is factor rotation: systematic de-grossing in momentum/low-vol sectors can create a multi-week air pocket in the most consensus growth names, while less-owned value baskets can rerate without needing major earnings revisions.

The memory complex looks vulnerable because it sits at the wrong intersection of expectations and elasticity. Supply has already been responding to high prices, so any sign of demand normalization can compress forward multiples faster than spot fundamentals deteriorate; that makes this more of a 1-3 month valuation unwind than a years-long semiconductor bear case. The bigger second-order loser is the broader AI supply chain: if investors start demanding proof of monetization, adjacent beneficiaries like equipment, interconnect, and select datacenter plays can see multiple compression before any revenue slowdown shows up.

Contrarianly, the market may be underestimating how quickly AI capex remains self-funded by hyperscalers, which could keep the trade alive longer than valuation purists expect. The key risk to fading growth is that relative earnings revisions may still favor tech for several quarters, and any macro soft landing backdrop would re-ignite the same leadership the strategist is calling tired. In that scenario, the rotation becomes a shallow factor correction rather than a durable regime shift, and the best shorts are the most levered, least profitable “AI pick-and-shovel” names rather than the mega-cap platforms.

Key Decisions for Investors

• Rotate 25-50% of any overexposed tech/growth basket into high-quality value/defensive cash generators over the next 2-4 weeks; prioritize names with 4-6% FCF yields and low earnings dispersion risk.
• Short SNDK on a 1-3 month horizon via equity or put spreads; use a modest-sized position because the risk/reward is skewed by multiple compression rather than immediate fundamental deterioration.
• Pair trade: long value ETF / short QQQ for 6-8 weeks to express factor rotation with lower single-name risk; stop if QQQ reclaims recent highs on accelerating breadth.
• For AI-adjacent exposure, trim the weakest balance-sheet names first and keep core platform longs only where capex is translating into visible revenue acceleration; the market is likely to punish “story” before it punishes earnings.
• If seeking convexity, buy 2-3 month put spreads on the most crowded growth leaders rather than outright shorts; this limits squeeze risk if the rotation proves temporary.