Semiconductor short positions ran into earnings that beat expectations

6/25 US market: opened high, plunged, then recovered. QQQ swung between 705 and 706 intraday, VOO was basically flat. But the semiconductor sector was severely torn apart—Nvidia fell below $200, Micron/SanDisk led the decline. Funds were buying puts on equipment/computing power chains while locking in long-term bullish positions against the trend. Looking at the following three unusual trades together paints a picture of "short-term bearish, long-term bullish, with bears buying at the point of getting their faces slapped."

$Lam Research(LRCX.US)
Direction: Bearish (suspected hedge)
Structure: Primarily Buy Put + small Buy Call for top insurance
Strike: Put $360 (8/21) · Call $440/$420
Size: Put $9.53M / 2619 contracts (4 concentrated buys from 09:43–09:59), Call only about $1.06M total
DTE: 57 days
Insight: The largest single short position of the day, but bought in the early session on 6/25—subsequently, the company's earnings exceeded expectations and guidance was raised, closing +7.2%. The $9.5M Put was bought in the early session on earnings day and got slapped by a +7% move, making a directional active short less likely. It's more akin to institutional downside insurance for long core positions. Simply following the short now risks being bitten by fundamentals; watch if the Put continues to be added after the rebound to discuss direction.

$NVIDIA(NVDA.US)
Direction: Bullish (long-term)
Structure: Buy Call · LEAPS stock replacement
Strike: $300 (2028-06-16)
Size: $4.92M / 1930 contracts (three stacked buys in the late session 15:29–15:54)
DTE: 722 days
Insight: At the moment when Nvidia fell below $200 and competition concerns peaked, three late-session trades poured $4.92M into $300 Calls expiring in 2028—this isn't betting on a rebound; it's using LEAPS to lock in a two-year bullish exposure, avoiding short-term volatility. The logic supporting this trade: short-term negatives (AMD/Qualcomm competition, Micron catching up) hit valuation sentiment, but long-term computing power demand hasn't been disproven. The entry barrier is high (two-year, deep out-of-the-money); ordinary positions are better off watching the $200 spot level rather than directly copying the LEAPS.

$Broadcom(AVGO.US)
Direction: Bearish
Structure: Buy Put
Strike: $310 (2026-10-16)
Size: $1.82M / 1352 contracts (single late-session buy at 15:41)
DTE: 113 days
Insight: Broadcom had just surged due to the OpenAI joint chip development news, then fell −0.8% that day. A single $1.82M late-session buy into the $310 medium-term Put (~18% below current price) is a medium-term bearish bet on "good news exhausted + semiconductor highs." This looks more like a true directional bet than the Lam Research one: bought after the news was realized, not on an earnings day, deep out-of-the-money medium-term. If following the downside, use $310 as the lower observation level and $382 above as a clearer stop-loss logic.

These three trades can be seen as a layered view of "medium-term bearish on semiconductor equipment/chips + long-term bullish on AI computing power leaders"—bears concentrated in segments where good news has been realized or expectations exceeded, while bulls used LEAPS to lock in Nvidia's two-year options exposure. Corresponding to the broader market, 6/25 was a structural pullback of "index flat, semiconductors under pressure alone," not a systemic shift.

Next focus points:

$NVIDIA(NVDA.US) whether the $200 round-number level can hold;

$Lam Research(LRCX.US) whether the $360 Put continues to be added after the rebound;

July Philadelphia Semiconductor Index (SOX) component earnings season

—Any one of these three turning points will determine which side of this "short-term bearish, long-term bullish" picture materializes first.