AI infrastructure spending is capital-hungry, and right now the market is charging companies dearly to raise equity — Corning lost 12% on a $2 billion offering and its optical peers fell in sympathy, while Oracle's scrapped $7.5 billion sale still sent th

Verdict: The Dilution Short Has Real Precedent — But No Trigger Is Live Yet

The strongest case for this short is a real and recent asymmetry: per the Yahoo Finance report, Corning lost 12% on a $2 billion offering while Coherent fell 11% and Lumentum 9% in sympathy, and insiders at all four names were net open-market sellers for the filing period ended June 30, 2026 — roughly $35.0M out of Corning, $39.6M out of Oracle, $9.3M out of Coherent and $64.9M out of Lumentum. The strongest case against is that the fundamentals argue demand is accelerating, not cracking: Lumentum grew revenue 83.2% with gross margin up 13.7 points to 41.7% for the fiscal year ended June 27, 2026, Coherent's net margin jumped 10.5 points to 11.3%, and Oracle posted a 25.4% net margin — shorting names whose income statements are improving is fighting the tape. The backtest evidence is real but window-dependent: 238.3% total return with an 83.3% win rate over 12 trades on the five-year COHR pairing, versus -6.2% on just three trades in the most recent two-year window. No entry is live today — every name's price is below its 9-day EMA trigger and the LITE and ORCL MACD histograms are still positive — and no offering is currently announced, so the setup waits rather than fires. A fresh dilution headline at any of the four names, or price reclaiming the 9-day EMA with ADX above 20, is the single fact that flips this verdict.

Trade now: the setup is coiled but not armed — here is the distance to each trigger

This idea is a rules-driven setup waiting on its entry conditions, and as of the September 15 data, none of the four tickers has all conditions live. The repeat is identical across COHR, GLW, LITE and ORCL: price must cross above the 9-day EMA, RSI (14) must be below 50, the MACD histogram must be negative, and trend strength (ADX, 14) must be above 20. Right now the momentum half is in place — RSI sits at 38.1 on COHR, 36.9 on GLW, 38.7 on LITE and 38.3 on ORCL, all comfortably under 50 — but every price-level trigger is still out of reach. COHR closed at $266.5 versus an EMA trigger of $287.24, a gap of $20.74. GLW closed at $143.6 versus $156.89, $13.29 short. LITE is the furthest away: $835.03 versus $905.14, a $70.11 gap, and its MACD is still positive at +20.5, so two conditions fail there. ORCL ($144.32 versus $151.98) has trend strength already above 20 at 24.0 but MACD still positive at +2.55. On risk, the strategy defines it mechanically: a fixed-risk sizing engine with a stop at a 2.4% loss and a take-profit at 4.9% per position, capped at 25% of the book. In price terms from current levels, that is roughly $260.1 down and $279.55 up on COHR ($143.6 to roughly $140.15 / $150.63 on GLW), so the effective reward-to-risk is about 2:1 on any position that triggers. Once a position is 30 days old, an exit also fires if price crosses back above the 9-day EMA. Support levels matter for the alternative entry leg: COHR's nearest support is $262.62, GLW's $141.63, LITE's $830 and ORCL's $139.86. What does waiting mean concretely? No order goes in today. The strategy's own history supports patience: over the 60-month window the traded COHR pairing produced 12 trades with an 83.3% win rate and a 238.3% cumulative return against a 20.3% maximum drawdown — results that only accrue by taking entries when the conditions actually align, not early. The nearest candidates to arm are GLW, where trend strength sits at 19.3, just 0.7 points under its 20 threshold, and ORCL, where only the MACD sign remains wrong. For a fuller picture of LITE's fragile fundamentals — gross margin improved to 41.7% year over year but net margin sits at negative 230.1% — the catalysts timeline below matters too.

Dilution Announcements as a Repeatable Short Trigger

The thesis rests on a real asymmetry: when AI-linked optical names need equity, the market punishes them hard. Per the Yahoo Finance report, Corning lost 12% on a $2 billion at-the-market offering while Coherent fell 11% and Lumentum 9% in sympathy — that is a 12-to-6 percent repricing in a single day, far more violent than typical sector drift. Oracle's scrapped $7.5 billion sale still knocked the stock lower per Investor's Business Daily, confirming that supply events are being punished even when the deal does not complete. The signal does not require you to be right about AI demand long-term; it only requires the market to keep charging a premium for equity access over the next few weeks, which the cited events suggest it does. The fundamentals make these companies unusually dilution-dependent. Oracle's FY2026 free cash flow is -$23.7 billion against $31.3 billion in cash and $122.3 billion in long-term debt, with capex of $55.7 billion — the cash engine cannot yet fund the AI buildout without new capital. Coherent's FY2026 free cash flow is -$1.0 billion after $1.1 billion of capex on $7.1 billion of revenue, and its shares outstanding jumped 27% year over year (from 157 million to 196 million), confirming the group is already relying on the equity market. Corning sits at the 1st percentile of its sector on free cash flow percentile measures, meaning there is little cushion if rates stay elevated. The quant evidence supports the short thesis with a real backtest. On a COHR-focused window over five years, a strategy keyed to these four names produced a 238.3% total return with an 83.3% win rate and a 20.3% maximum drawdown, drawn from 12 completed trades. Over the most recent 12-month window, the same framework returned 1.8% on a single trade with near-zero drawdown. The exits are disciplined: positions stop out at roughly -2.4% and take profit near 4.9%, so the edge is realized in small, frequent wins rather than a single bet. Ownership and insider posture corroborate the bearish signal. For the June 2026 filing period, insiders at all four names are net open-market sellers — Oracle insiders sold $39.6 million net, Lumentum $64.9 million net, Corning $35.0 million net, and Coherent $9.3 million net. That is not a single-company quirk; it is a synchronized disposition across the exact basket the thesis targets, with 20 ORCL and 16 GLW filers respectively participating. When the people closest to the business are selling into a market that already punishes new supply, the short signal has confirmation from two independent sources. The group's weakest fundamental changes align with a capital-consumption…

Scores

• Conviction score breakdown: 52
• Thesis support: 70
• Trade readiness: 35
• Risk quality: 60
• Backtest evidence: 55
• Fundamentals trend: 40

Watch items

• GLW — ADX (14)
• ORCL — MACD (12,26,9) histogram
• COHR — Close vs 9-day EMA
• LITE — Insider net open-market selling
• GLW — Next dividend event
• COHR — Close vs nearest support