Louis_t

SOXX fell 22.1% in July and took in a record 6.92 billion of net inflows the same month. first half of the year was 7.75 billion total, so July alone nearly matched six months. that is not capitulation, that is people buying a drawdown with both hands. leveraged version of that is a very different risk profile though, size accordingly.

quiet period ended yesterday and there is 69.4 trillion won of net cash sitting on the balance sheet. market is front running the buyback announcement, already +6% so part of the news is in. still adding a starter, that cash pile is the floor.

spacex prints its first ever public quarter tonight, then 911.5m shares unlock thursday, about 116b worth. it was 108.37 on friday against a 231 to 236 street target. two hard dates in three days is not a valuation debate, it is a supply test. staying in cash until friday 🫡

intel foundry still lost 2.1 billion last quarter and they raised 2026 capex to over 20 billion anyway. group operating margin did flip positive though, and eps came in at 42 cents against 22 expected. i hold a small position as a turnaround lottery ticket, not as my semiconductor exposure

qqq back near highs going into the busiest earnings week of the year. tesla and google tonight, msft and meta next week. this is the bellwether so how it holds after these prints tells you a lot about the tape. staying long but with tight stops

Goldman just doubled the SanDisk target to 2,200 on NAND tightness. the numbers are getting silly but the supply story runs into 2027. small position, tight leash, this thing moves 10% on a random Tuesday

$Corning(GLW.US) quietly sitting at record highs after locking in that multi billion Amazon fiber deal for AI data centers. the picks and shovels plays are eating good this cycle. optical fiber is the boring backbone of the whole AI buildout and nobody talks about it 🔌

$Marvell Tech(MRVL.US) quietly up 3 percent while everyone watches the big names. UBS and Cantor both raised targets on the CXL story, small dividend coming next week too. underrated AI infra play imo, been slowly accumulating

the number that actually matters isnt the revenue, its the 84.6% gross margin and the ~86% guide. memory was always the cyclical dog of semis, low margin, boom-bust. if MU can hold mid-80s margins because AI HBM is sold out, this stops being a cycle stock and starts re-rating like a structural winner. that 100B of contracted revenue across 16 customers is the part i keep rereading.

$SpaceX(SPCX.US) first dip and my finger already hovering the buy button. Starship cadence is the whole thesis, one red candle changes nothing

Oracle holding steady, the cloud and AI capacity story keeps it bid even on red tape days. slow but steady, i like names that dont panic

RKLB getting added to the Nasdaq 100, that is the kind of milestone that reprices a stock. small cap space going properly mainstream now