Quick Read
-
MUU shed 27% in a single session as MU dropped 13%, despite the stock still sitting 203% higher year to date.
-
Broadcom’s AI revenue guide missed buyside expectations by 7% and sent AVGO down 20%, while NVDA fell only 6% on Friday.
-
Micron’s June 24 earnings are the binary event. Soft HBM pricing confirms a cycle break, while a clean beat reverses MUU’s damage.
-
It sounds nuts, but SoFi is giving new active invest users up to $1,000 in stock for a limited time, and all it takes is a $50 deposit to get started. See for yourself (Sponsor)
If you held Direxion Daily MU Bull 2X Shares (NYSEARCA:MUU) into Friday’s close, you watched the fund shed 26.65% in a single session, which is what 2x daily leverage does when its underlying has its worst day in years. The underlying, Micron Technology (NASDAQ:MU), closed at $864 on June 5, 2026, down roughly 13% from a prior close of $996. A roughly 13% move in the stock turning into roughly 27% in the fund is the textbook outcome on a clean one-day drop, no volatility drag required. The interesting question is what happened to MU. The fund did exactly what it is engineered to do.
The arithmetic of a leveraged unwind
Before Friday, MU was the chip cycle’s poster child. The stock was still up 203% year to date through June 5 and 715% over the prior twelve months, riding an HBM (high-bandwidth memory) supercycle that the company was telling investors stretched into multi-year contracts through fiscal 2026. Micron’s market capitalization sat at $974.37B after the Friday drop, with CEO Sanjay Mehrotra having raised guidance just two quarters earlier to $18.70B in revenue plus or minus $400M for fiscal Q2 2026, alongside a non-GAAP gross margin guide of 68.0%. The Q1 report itself delivered $13.64B in revenue (up 57% year over year) and EPS of $4.78 versus a $3.94 consensus. None of that fundamental picture changed on Friday. The price changed.
For MUU holders, the math is brutal in both directions. The fund resets daily, so a stock that compounds higher day after day drags the leveraged vehicle along at a rate that exceeds 2x cumulatively. The same arithmetic in reverse, a single concentrated drop, gives you exactly Friday. A roughly 13% loss on MU produced a roughly 27% loss on MUU, almost linear. There is no need to invoke volatility drag for one bad day. The drag arrives later, in the chop that tends to follow a violent move like this one.
What actually triggered the selloff
The crack started on Wednesday night, two days before MU got hit. Broadcom (NASDAQ:AVGO) reported after the close and guided Q3 AI semiconductor revenue to roughly $16.0B against an unofficial buyside number closer to $17.2B, a miss of about 7%. CEO Hock Tan added commentary that Google may use multiple chip suppliers, which the market read as a small but real notch out of the hyperscaler concentration thesis. AVGO fell roughly 20% over June 3 to June 5, including a nearly 8% drop on Friday alone.
Friday morning brought the second leg. May nonfarm payrolls printed 172K against an 80K expectation, which revived Fed rate-hike chatter and pushed the 2-year Treasury yield to 4.16%, a 16-month high. The 10Y-2Y spread compressed to 0.38% on June 5, the low of the trailing twelve months, after sitting at 0.51% on May 1. Rate-sensitive growth names led the way down. The PHLX Semiconductor Index fell roughly 10% and the broader semiconductor complex shed about $1.2T in market value. MU itself lost roughly $113B in market cap.
The other AI-adjacent names told you what kind of selloff this was. NVIDIA (NASDAQ:NVDA) fell only 6.2% on Friday to $205.10, a fraction of the MU damage. NVIDIA still posted $81.62B in Q1 FY27 revenue, up 85% year over year, with Jensen Huang describing “the buildout of AI factories, the largest infrastructure expansion in human history”. Contagion hit MU disproportionately because memory is, and has always been, the most cyclical layer of the semiconductor stack. When demand softens at the margin, memory pricing moves first and moves hardest. The bull case requires that supply stays disciplined while demand grows. SK Hynix and Samsung racing to expand HBM capacity is the supply side of that ledger doing the opposite of what bulls need.
SoFi Active Invest is offering a limited-time promotion. Open an account, fund it with $50 or more, and you could receive up to $1,000 in complimentary stock for Active Invest accounts. See for yourself by clicking here now.
The leveraged-ETF tax that comes next
Friday’s drop on MUU is the easy part to explain. What MUU does over the next two or three weeks is where the structural cost of holding a 2x daily product shows up. If MU chops in a range while traders wait for the next earnings print, MUU bleeds value even on net-flat tape because the daily reset compounds losses on alternating up and down sessions. Prediction markets are pricing exactly the kind of chop that punishes MUU. The Polymarket book on MU for the week of June 8 puts 61% probability on a $900 close, with the up-or-down market on June 8 sitting at 51% up and 49% down, a coin flip. End-of-June pricing gives 58% probability that MU closes above $980, which would represent recovery, though not all the way back to the $996 starting line.
Reddit captured the human side of the unwind in real time. The wallstreetbets sentiment score moved from 94 (very bullish) on June 1 to 36 (bearish) by June 5 at 9am ET, the peak activity moment of the entire period. The top post by engagement, with 16,396 upvotes and 903 comments, was a capitulation note from a trader who had ridden the move from a starter account to seven figures and watched it collapse. The post that defined the prior weekend, by contrast, was titled “+6,476.76% gain on MU LEAPS, should I sell?”. That is the shape of a forced unwind, leveraged retail positioning last to give in.
What to watch into June 24
The next real piece of information arrives on Wednesday, June 24, 2026, after the market close, when Micron reports fiscal Q3. The company has confirmed the date. Four things in that release determine whether Friday was an air pocket inside an intact supercycle or the start of a real cycle break.
-
HBM3e average selling prices. If pricing held through May, the contract structure is doing its job. If pricing softened, the SK Hynix and Samsung capacity expansion is already biting.
-
Contract structure commentary. Micron has been shifting toward 3-5 year supply agreements with hyperscalers. The longer the locked-in book, the less the spot cycle matters in the near term.
-
Hyperscaler inventory builds. Broadcom’s guide hinted at slower hyperscaler order growth. Micron will either confirm that read or push back on it directly.
-
Fiscal 2027 CapEx guidance. The first number that tells you how Micron itself reads the cycle, ahead of any analyst model revision.
For MUU specifically, June 24 is the binary event. A clean beat with held HBM pricing and an extended contract book reverses Friday quickly, and the 2x mechanics work for holders again. A guide that confirms the Broadcom data point, with any softness on HBM ASPs, compounds the damage in a way that makes Friday look like an opening move rather than a flush. The forward read is that the market just priced in a real, nonzero probability that the memory cycle has broken, and a 2x daily leveraged single-stock ETF is a poor vehicle to hold through that question being answered.
Want Up To $1,000? SoFi Is Giving New Active Invest Users Free Stock
Looking to grow your money but unsure where to begin? SoFi Active Invest is offering a limited-time promotion—open an account, fund it with $50 or more, and you could receive up to $1,000 in complimentary stock for Active Invest accounts.
From $0 commission trading to fractional shares and automated investing, this app is designed to simplify investing for everyone, whether you’re just starting or already experienced. Its easy to sign up and secure your bonus.
