MRAM · Everspin Technologies, Inc.

Current Thesis

The AI-memory narrative that powered a +300% run (52-week range $5.64–$51.50) is breaking, and the tape is now confirming it. Kerrisdale Capital disclosed a short on 2026-05-19 with a $14 price objective, arguing the rally is disconnected from a business whose largest end market is casino gaming and slot machines not hyperscalers or AI servers with revenue stuck in a $50–65M annual band for years. Price is rolling over: ~$29.25 on 06-04 to a $23.03–$27.99 range on 06-06, closing ~$23.05, still above every published analyst PT ($17.5–$18.5, avg $18) and far above Kerrisdale's $14. This is a late-stage squeeze unwinding, not an accumulation base a chase trap, not a fresh long.

Bull Case

• Q1 2026 (reported 2026-04-29): revenue $14.9M, +14% YoY, beat the $14.63M estimate; MRAM product sales $14.1M, +28% YoY and +5% sequentially.
• Gross margin lifted to 52.7% on better mix and yields; adjusted net income $2.6M vs a $1.2M loss a year prior.
• New $40M, 2.5-year defense subcontract (Toggle MRAM process tech to the US Defense Industrial Base) real diversification away from industrial/embedded, plus $2.1M of other income booked in Q1.
• Q2 2026 guide $15.5–16.5M implies continued sequential growth.
• Roadmap: high-reliability 128Mb/256Mb UNISYST MRAM parts targeted for H2 2026, engineering samples late 2026.
• Cash $40.5M with low debt a survivable balance sheet for a ~$60M-revenue microcap, and a tight float that fuels squeezes on any positive headline.

Bear Case

• Kerrisdale (2026-05-19) $14 PT = ~63% downside from pre-report levels; core claim is that MRAM is substitution-driven niche memory for industrial/embedded persistence, solving a different problem than the HBM/DRAM running AI clusters.
• Revenue stagnant in the $50–65M/yr range for years; 14% YoY growth does not justify a >300% re-rating.
• Insider distribution is one-directional: $3.9M sold last quarter with zero buys, then Director Finch's 310,091 shares / $8.52M (~$27.47) on 06-02-2026 selling straight into the squeeze.
• The headline $0.40 Q1 EPS (vs $0.01 est) was inflated by one-time $2.1M defense other-income; the underlying result was still a $0.3M net loss.
• Price sits above every sell-side PT even after the 06-06 drop to ~$23, leaving the $14–$18 cluster as the mean-reversion target.

Setup & Price Structure

• 52-week range $5.64–$51.50; the May parabola topped well below the high and is now marking down.
• 06-04 ~$29.25 → 06-06 traded $23.03–$27.99 and closed ~$23.05, a wide distribution day that broke the parabolic structure.
• No clean higher-low base has formed; the stock is stretched far above any rising moving average after the run, with momentum now negative.
• Trading above all analyst PTs ($17.5–$18.5) into a short-report-plus-insider-sale backdrop the path of least resistance is toward the $14–$18 zone.
• Structure reads as a post-parabolic markdown with active short pressure, not a pullback to support inside an uptrend.

Catalyst Calendar (next 30 days)

• Q2 2026 earnings: NOT in the window. Q1 was reported 2026-04-29; the Q2 print is expected ~late July (~2026-07-28, est.).
• No FDA/PDUFA, no scheduled product launch in window UNISYST 128Mb/256Mb parts are H2 2026, samples late 2026.
• Open-ended risk, not scheduled: further insider Form 4 sales (pattern active through 06-02) and any Kerrisdale follow-up.
• No dated binary catalyst inside the next 30 days.

What Would Change Our Mind

• A genuine AI/datacenter design win with a named customer (not casino gaming or generic industrial) the one thing that would validate the premium the short report attacks.
• Revenue breaking decisively above the $50–65M/yr band: a Q2 actual above ~$17M or a raised FY guide.
• Insider buying replacing the current distribution pattern.
• Technically, a clean higher-low base that reclaims the 20-EMA after the markdown not a reflexive bounce off a falling knife.
• Absent those, strength here is distribution to exit, and the watch stays defensive.

Correlation Notes

• Trades with the small-cap-ai-momentum / memory-squeeze cohort it appeared alongside on 05-11 (GSIT, NVEC and the speculative IT-mover baskets), driven by retail flow rather than semis fundamentals.
• High beta to risk-on/risk-off retail sentiment; behaves like a squeeze name, decoupled from its own income statement.
• Short-seller-targeted microcap with a tight float single-headline gap risk in both directions, amplified by ongoing insider supply.