Quick hits - July 08, 2026

$SMH held 580. Harder test setting up now.: $SMH defended the 50d ema and the 580 level yesterday, which is what the bulls needed to see. A tougher test is now setting up. Hold 580 here and the bounce thesis stays alive. Lose it and the gap to 560 reopens.

Goldman at $1T. Jefferies slightly lower. Neither says CAPEX peaked.: Goldman is projecting $1T in hyperscaler CAPEX spend for 2027, and Jefferies came in slightly lower but corroborates the same call: the four major US hyperscalers have very likely not peaked on spend. That's the direct counter to the META peak-capex narrative that's been weighing on chip names all week. If the institutional read is right, the capex receivers still have ceiling to find.

Memory stocks peak 6 months before prices do. Jefferies sees hikes through 2027.: The pattern: memory stocks typically top out around 6 months before memory prices peak. Jefferies is forecasting 40-45% memory price hikes running through 2027. If that timeline holds, the peak in the stocks may still be ahead of us, not behind. Worth keeping in mind before fading $MU or $SNDK on the near-term tape.

Morgan Stanley pushes back on CPO delay FUD.: FUD has been circulating on co-packaged optic delays, some of it from Semianalysis. Morgan Stanley's latest estimates push back on that, suggesting the CPO scale-up timeline is still intact. If the delays are overblown and the narrative fades, names with direct CPO exposure like $ALAB may be discounted on a thesis that isn't actually broken.

$PENG: guidance looks sandbagged, value still here post-earnings: A quick model on the print shows $PENG sitting at good value even after the beat. The read is guidance was sandbagged and they could beat strong again next quarter. If that plays out, the pullback in today's tape may look like the entry in hindsight.

Blue Origin raising at $130B. $RKLB trades at $49B.: Blue Origin is reportedly raising $10B at a $130B valuation. $RKLB is more vertically integrated and currently trades at roughly $49B. Market may not close that gap quickly, but the relative read is hard to ignore.

$VPG near entry, bear case getting pushed back: $VPG has pulled back close to the original average entry price. Long-term thesis is intact and recent research is pushing back on the primary bear argument. Hold, don't chase.

Trump hit Iran. Now he's watching. TACO forming.: Trump struck Iran overnight and is now saying he'll monitor before deciding whether to strike again. Pattern looks like a TACO forming, so if energy names spike hard on the headline, don't chase. Fade the rally if follow-through isn't there.

$QQQ pinned at 709, $SPY needs 750 to hold: Early bounce fading into $QQQ's 709 resistance and the map doesn't look all that bullish. $SPY needs to pull to 750 and hold it as foundation: if it can't, amplification levels point lower toward 740 support. Dealers are still defending the floors but less optimistic. Mega caps still long gamma, which helps, but conviction on this bounce is low right now.

Energy catching a bid. Trump and Iran doing the work.: Trump is floating moves that could push oil higher, and Iran war headlines are already giving energy a bid today. These headline-driven catalysts fade fast, so watch whether energy holds into close or gives it back.

$QQQ into a key test. NDX holding 29k and semis getting the bid.: $QQQ is pressing into an important technical test right now. NDX holding 29k on a headline-driven day is more encouraging than the tape feels. The rotation out of $IGV and into $SMH is the other tell — if semis keep getting the bid while software fades, the setup may be healthier than it looks.

$NBIS reclaims $200. Sentiment may be shifting.: $NBIS just reclaimed $200, and for a name we flagged at support earlier this week, that's a meaningful sentiment shift. Watch whether it holds the level.