Everspin Technologies stock tumbles after short seller report By Investing.com

Everspin Technologies fell 8% after Kerrisdale Capital disclosed a short position and argued the stock’s rally is disconnected from fundamentals. Kerrisdale said Everspin is not meaningfully exposed to AI infrastructure, cited stagnant revenue of roughly $50 million to $65 million for years, and set a $14 price target, implying about 63% downside. The firm also noted heavy trading volume and insider selling during the stock’s surge of more than 300%.

Analysis

The key issue is not just that MRAM is expensive; it is that the market has started to price it as a proxy for AI memory capex despite no evidence that its product is being pulled into the bottleneck applications that matter. That mismatch is vulnerable to a fast de-rating because semiconductor sentiment is highly reflexive: once a short thesis lands, liquid small-cap names with crowded ownership can unwind 20-40% in days before fundamentals even need to reassert themselves.

The second-order effect is that this resets the relative-value framework across the entire “AI memory” basket. If investors lose confidence in adjacent beneficiaries, capital will likely rotate toward names with direct exposure to HBM, DRAM, or advanced packaging and away from legacy/industrial memory substitutes; that should compress multiples fastest in companies whose growth is substitution-led rather than demand-creating. The insider selling noted during the rally also matters because it gives long-only holders a narrative bridge to justify profit-taking rather than defending the position.

Catalyst timing is asymmetric: the downside can arrive immediately on follow-through from the short report, but the upside rebuttal would require visible design-win evidence, a material revenue inflection, or management guidance that proves AI adjacency within one to two quarters. Without that, the stock is likely to trade more on positioning and short interest than on operating data. The contrarian risk is that small-cap semis can remain detached from fundamentals longer than expected if retail momentum and thematic inflows keep rotating into “AI themer” names, but that tends to be fragile once the first earnings miss or guidance pause appears.

Best risk/reward is to express the view as a tactical short or put spread rather than outright directional exposure, because the main edge is timing, not terminal value. If the market keeps rewarding thematic scarcity, a relative-value short versus a true AI memory beneficiary should outperform a naked short by reducing beta to the sector tape. The cleaner setup is to fade MRAM into any strength while preferring longer-duration longs in names with real AI memory content and near-term confirmation of demand.

Key Decisions for Investors

• Short MRAM on any post-report bounce; target a 15-25% drawdown over 2-6 weeks, with a stop above the prior momentum high because squeeze risk remains elevated in crowded small caps.
• For cleaner risk management, buy MRAM 1-3 month put spreads rather than outright puts; downside can accelerate quickly, but implied volatility should stay bid after the short disclosure.
• Pair trade: short MRAM / long a direct AI-memory or packaging beneficiary with visible hyperscaler exposure for 1-3 months; this isolates the thesis that MRAM is a narrative beneficiary, not an earnings beneficiary.
• Trim exposure to other low-fundamentals, AI-adjacent small-cap semis into strength; if MRAM breaks, the re-rating can spill into the broader thematic basket over the next several sessions.
• Watch for the next 1-2 quarterly reports and any customer concentration commentary; absent a concrete AI-related design win, use any confirmation rally to add to the short rather than cover it.