He Watched His $1M ETF Investment Crash to $200k. Then He Cashed Out at $7.5 Million

A Redditor’s SOXL trade reportedly turned $1 million into about $7.5 million, despite an 80% drawdown during the 2025 tariff-driven semiconductor selloff. The article says SOXL has since surged to around $234, including a 165% gain in April 2026 and a 971% rise over one year, driven by renewed AI infrastructure spending and strong semiconductor momentum. It also warns that the 3x leveraged ETF remains highly volatile and path-dependent, with tariff risk and hyperscaler capex guidance the key variables to watch.

Analysis

The key second-order effect is not just that AI capex is strong, but that the trade has become mechanically reflexive: hyperscaler spend validates the semiconductor order book, which lifts the basket, which then forces systematic buyers and momentum funds to add exposure, amplifying the move beyond what fundamentals alone justify. That matters most for the highest beta names in the mix; NVDA and AVGO are the clearest leverage points because they sit closest to data-center demand and can re-rate on guideposts before earnings catch up.

The hidden risk is that SOXL is now trading like a consensus macro expression rather than a stock-picker vehicle. Once positioning becomes crowded, the next drawdown is usually triggered by a guidance miss, export-control headline, or a modest capex pause that is trivial for the underlying businesses but severe for a 3x wrapper because of daily reset and volatility clustering. In other words, the upside can persist for months, but the path becomes far more fragile when realized vol stays elevated even if the trend is positive.

The market appears to be underpricing how quickly a tariff or trade-policy shock can reintroduce supply-chain air pockets, especially for TSM-linked manufacturing and any company with China-sensitive demand. The contrarian setup is that the “AI infrastructure supercycle” may still be intact, but the easy money in levered beta is likely behind us; from here, incremental upside likely comes from stock selection and call structures rather than outright ETF exposure. QCOM is the least direct beneficiary and the most vulnerable to being left behind if investors continue to pay up only for pure data-center exposure.

A practical read: the trade remains bullish over weeks-to-months, but the asymmetry has shifted from “buy the dip” to “own upside with defined risk.” If volatility compresses, leveraged exposure can still work; if not, the same mechanics that created the windfall can erase it quickly.

Key Decisions for Investors

• Rotate from SOXL into NVDA and AVGO calls for the next 1-3 months; the basket still has upside, but single-name options better isolate AI capex upside with less volatility decay than a 3x ETF.
• Use a bull call spread in SOXL rather than shares if expressing upside over the next 4-8 weeks; target a 1.5:1 to 2:1 reward/risk and cap exposure because daily reset risk rises sharply after a 100%+ run.
• Pair long NVDA/AVGO vs short QCOM over 1-2 quarters; this isolates data-center strength against names more exposed to handset/enterprise cyclicality and should outperform if AI spend remains the dominant driver.
• Trim or hedge any existing SOXL longs on spikes in VIX or tariff rhetoric; the drawdown profile can reassert within days, not months, if trade-policy headlines hit the tape.
• Monitor hyperscaler capex commentary and semiconductor order-book revisions as the next catalyst window; if guidance stays elevated into the next earnings cycle, add on pullbacks rather than breakouts to improve risk/reward.