A $10,000 position in the MicroSectors FANG+ 3X Leveraged ETN (NYSEARCA:FNGU) at Thursday’s $32.16 close was worth about $8,392 by Friday’s $26.99 close, a 16% single-session drop on June 5, 2026. Stretch the window back one week and the damage gets uglier. FNGU went from $34.76 on May 29 to $26.99 on June 5, a 22% slide in five trading days. The note is still up 28.1% over twelve months and up 6.85% year to date, which tells you exactly the kind of instrument FNGU is. It pays you to be right about direction, and only on the days you actually hold it.
The arithmetic of a 3x daily reset meeting a real selloff
FNGU is a 3x daily leveraged exchange-traded note issued by Bank of Montreal, tracking the NYSE FANG+ Index, an equal-weighted basket of ten mega-cap technology and internet names. When that basket falls roughly 5% in a session, FNGU is engineered to fall about 15%. On Friday it fell a little more than that, because the index was top-heavy in the names doing the most damage.
NVIDIA (NASDAQ:NVDA | NVDA Price Prediction) fell 6% on June 5, closing at $205.10, a single-day market cap loss the custom data pegs at roughly $279 billion. Broadcom (NASDAQ:AVGO) dropped 8% to $385.73, capping a 19.5% two-day collapse from $479.23 on June 3. Meta Platforms (NASDAQ:META) fell 6% to $593. Microsoft (NASDAQ:MSFT) gave up 3% to $416.67, and Alphabet (NASDAQ:GOOGL) was the calm one at down 1%, closing $368.53. Equal-weight on paper, concentration in practice when the weakest names happen to be the largest line items by absolute dollar moves.
What actually broke this week
Two catalysts stacked on top of each other. The first was Broadcom. The company reported Q2 FY26 results Tuesday evening with non-GAAP EPS of $2.44 against a $2.3972 consensus and revenue of $22.19 billion, then guided Q3 to approximately $29.4 billion in revenue with AI semiconductor revenue of $16.0 billion, more than 200% growth year over year. On absolute terms, that is an extraordinary number. On the implied trajectory the market had baked in, with AI semis having compounded from $5.20 billion to $8.40 billion to $10.80 billion in the prior three quarters, $16.0 billion fell short of the whispered bogey. The stock did the thing whisper numbers do to stocks.
Then came payrolls. Friday’s jobs report printed 172,000 against an 80,000 estimate, sending the 2-year Treasury yield to 4.16%, a 16-month high, and revived the rate-hike conversation the market had stopped having. The 10Y-2Y spread compressed to 0.38% on June 5, the lowest reading in the trailing twelve months, down from 0.49% on May 5. Short rates ripping while long rates barely budge is exactly the cocktail that punishes growth stocks priced on long-duration cash flows. FANG+ is nothing but long-duration cash flows wearing trench coats.
Why FNGU amplifies in both directions
Three things about the structure are worth saying out loud. The first is daily reset compounding. A 3x note tracks the daily move of its index, not the cumulative move. In a sustained directional trend, that compounding works in your favor. In a chop tape, it bleeds value even when the underlying ends up flat. The math is unforgiving and it does not care about your conviction.
The second is concentration. FANG+ holds ten names. When NVIDIA and Broadcom together represent a quarter of the move on any given day, you are buying a concentrated AI semiconductor bet with some social media and search bolted on, dressed up as a diversified tech basket.
The third is that FNGU is an ETN, not an ETF, which means it is unsecured senior debt of Bank of Montreal. The credit risk is small in practice and large in theory, and it is real, even if nobody talks about it on the way up.
The setup heading into next week
The fundamental backdrop the bulls keep pointing to has not changed in any material sense. NVIDIA’s Q1 FY27 print on May 20 showed Data Center revenue of $75.246 billion up 92% year over year, with networking up 199%, and CEO Jensen Huang calling the buildout "the largest infrastructure expansion in human history." Broadcom’s $16.0 billion AI semi guide is a disappointment only in the sense that it is not a bigger disappointment. The mechanism that drove FANG+ higher in the first place, hyperscaler capex chasing AI compute, is intact.
What changed is the price of money. The 10-year Treasury yield closed at 4.47% on June 4, in the 93rd percentile of its trailing twelve months, and the front end ripped higher Friday on the payrolls beat. The VIX closed at 15.40 on June 4, in the lower quartile of its twelve-month range, which means the option market had not yet repriced volatility when this drop happened. Friday’s session likely pulled VIX higher off the floor.
Three things to watch from here. AI capex commentary from the hyperscalers into the next earnings cycle, where the question is whether the $125-145 billion Meta capex range and the run-rate spend at Microsoft and Alphabet hold up or get trimmed. Alphabet’s $80 billion stock sale execution, which removes a meaningful chunk of supply at a moment when the marginal buyer is more rate-sensitive than a month ago. And the SpaceX IPO on June 12, which will tell you whether retail still has appetite for the highest-beta tech story available, or whether the absorption is uglier than the roadshow expected.
Polymarket sentiment captures the ambivalence. The crowd gives NVIDIA a 91.5% probability of closing above $170 by end of June but only a 57.5% probability of clearing $200, the price it sat at before Friday. Alphabet’s June 8 direction market sits at 60.5% down. Microsoft’s two-day direction trades at 62.5% down. The market is pricing more pain near term and partial recovery by month end. That is exactly the regime where a 3x daily reset note is most punishing, because chop kills compounding.
If you held FNGU into Friday, you already know what happened. If you are looking at the 28.1% one-year return and thinking about getting in, the indicator that matters is the trajectory of the 2-year yield and the tone of AI capex commentary in July. The story that drove FNGU higher is the same story that broke it this week. Whether the next month rhymes with May or with Friday is the only question that matters, and the honest answer is that nobody knows yet.
Top Gaining Stocks
Top Losing Stocks