Quick hits - July 16, 2026

TSM guidance points to acceleration, not a peak: TSM's Q3 guide of $44.6-45.8B implies 37% YoY growth, actually faster than Q2's 33.7%, with the full-year guide sitting above 40%. That's acceleration, not deceleration, and it stacks on top of the capex-raise story already out there today. Doesn't change the trade, just backs up the demand read into Q4.

Semicap basket flagged bullish: LRCX, AMAT, KLAC, AMKR, ASX: Broad bullish call across the semicap and OSAT chain, LRCX, AMAT, KLAC, AMKR, and ASX all named. No specific level or catalyst attached, reads more like a basket riding the same capex wave TSM's guidance points to. Watchlist material, not a trigger yet.

CPO names: delays in the headlines, TAM thesis intact per BofA: Reports of delays hit the CPO (co-packaged optics) names, but BofA reiterated the TAM is still massive and says most of the scale-up is still ahead, not behind. If that holds, recent price action may be more noise than thesis-breaker. Watching for confirmation before chasing any dip here.

BofA's humanoid forecast: 2045 story, wild chip TAM math: BofA dropped some outrageous humanoid robot forecasts out to 2045, but the useful part is the semiconductor demand math buried under the headline. Way too long-dated to trade off today, just one for the back pocket on where the next leg of secular chip demand could come from. File under revisit later, not act now.

HSBC raises wafer equipment forecasts through 2030: HSBC just bumped every year of their wafer equipment forecast out to 2030, another big desk piling onto the AI capex bull case. Nothing to trade off this alone, it's confirmation not catalyst. Filed under: the equipment supercycle thesis keeps getting reinforced from the sell side, not refuted.

$SNDK's NBM contracts could be the real HBF story: Morgan Stanley's read on $SNDK flags their New Business Model contracts, long-term 3-5 year supply deals that lock in better economics than the old spot-driven approach. Layer in HBF, something $AEHR has been talking up on its own earnings calls, and SNDK may be sitting on a catalyst that isn't fully priced in yet. Nothing actionable today, just a name worth keeping on the memory-supercycle watchlist.

$PENG's asset-light neocloud model, explained: There's real confusion around how $PENG's asset-light approach to being a neocloud actually works, same playbook $NBIS runs. Partners put up the capital and own the physical infrastructure, the buildings and servers, while PENG runs the operations layer on top. If that structure holds, it reframes the balance-sheet risk the bear case keeps leaning on. Still theory-crafting until the market actually re-rates it.

$ATAI holders, the Lilly deal changes the option math: Community chatter has Eli Lilly acquiring $ATAI. If that holds up, long-dated calls out to Jan '27 and Jan '28 likely get capped near the deal price fast, the stock doesn't need to run there anymore for those to be in the money. Worth confirming the official terms before assuming that upside is still live, M&A tends to collapse IV and cap premium in one move.

Jobless claims light, retail sales in line: Jobless claims printed 208k against 217k expected and June retail sales came in at +0.2% m/m, right on forecast. No real surprise in either print, shouldn't move the tape much on its own.

$SPY gamma flips negative, 750 is the line: Today's book settled negative and short gamma, basically a mirror image of yesterday's long-gamma setup. Dealers are chasing moves now instead of absorbing them, so don't lean on anything holding cleanly. 752 looks like a magnet but the kind that pulls and doesn't hold. If 750 breaks, the -GEX kicks in and it could accelerate downward fast, don't be a hero here.

Warning shot level held to the tick, big buyer defending: The warning shot level held exactly where it needed to, and it looks like there's a real buyer sitting on it. If that buyer gets overrun, that's the tell sellers finally get a cleaner opening lower. Until then the level's still doing its job.

Growth portfolio underwater since May, no panic yet: Most picks since May are sitting at least slightly red, memory names included, but the read is these were ill-timed entries, not broken theses. Frame it against the worst momentum selloff in 27 years and the drawdown makes more sense. Stomach for volatility, not stock-picking, is what actually separates the survivors here.

$OSS still liked, but not buying the dip yet: $OSS is trading well below its prior highs even though earnings two months ago were A-plus, proof this meltdown hasn't spared anything. Not adding here, but the plan is to keep gauging the name off the fundamentals instead of the recent price action. Watching, not chasing.

$AMZN ascending triangle into OPEX, $257.50 is the GEX node: $AMZN is tracing an ascending triangle, and $257.50 lines up as a GEX node right into tomorrow's OPEX. That expiry resets gamma positioning, so the level could act as a magnet or a launchpad depending on how dealers rebalance. Watching to see if the triangle holds and breaks with it, not chasing ahead of the reset.

ES pivot at 7575 is the whole ballgame, NDX testing 29k: A big buyer stepping in at 7575 on ES is the pivot holding this whole market up right now. NDX is testing support at 29k at the same time, so if that ES level gives way the downside could come fast. SMH's already below 580 support, and if ES loses its pivot, semis probably get dragged lower with it.

NDX closes below 29k, loses the 50d for the first time: Our ES 7575 buyer got turned back for now. Bigger tell: NDX just closed below 29k, its first close below the 50-day. Checking for another ABI signal at end of day, if that fires again on top of this it adds to the case that the breakdown has legs.