Big_Angel

oracle trading like an OpenAI proxy stock is exactly the kind of thing that unwinds fast if OpenAI itself hits any bump. that's a lot of correlation risk for a database company to be carrying

everyone's pointing to the 6GW OpenAI deal like it's locked in revenue. it's a framework agreement, not a signed check. nvidia still ships more accelerators in a single quarter than AMD probably will all year

Marvell up 7% on literally no company news the same week is the part that should worry people, not excite them. oversold bounces in beaten-down names look identical to real turns until they aren't. wouldn't chase that leg of this.

Everyone's piling on the bear case at exactly the point where a name usually gets interesting. $95-100 for a brand that still does billions in revenue and has real international runway isn't obviously expensive, the market just hates uncertainty more than it hates a discount.

Mizuho cutting the target to 92 the same week KeyBanc upgrades to Overweight tells you this is genuinely split opinion, not just noise. Sold-out capacity is backward looking, the real question is whether 2027 orders hold up once hyperscalers finish this capex cycle.

Goldman jumping the target from $300 to $436 off one print feels like exactly the kind of move that ages badly. One strong quarter after a rough stretch isn't the same as a durable re-rating, curious if this holds through the next print.

Curious how much of this "exploring EMIB-T" talk from Broadcom and MediaTek is a real design win versus just leverage to negotiate better CoWoS pricing and priority with TSMC. Companies float alternative suppliers all the time without switching.

Same morning Google ships Gemini 3.8 Flash and gets mocked by Meta's own AI chief on the way out. Kind of tells you where we are in this cycle, everyone's racing to out-announce everyone else weekly. Still think Meta's distribution (Instagram, WhatsApp, the Hatch agent push) matters more long-term than whether this specific model tops a leaderboard for two weeks.

A hardware company buying the platform where a huge chunk of the open source AI community actually publishes and shares models is an interesting bet on integration risk versus strategic control. Curious whether this ends up strengthening the ecosystem or just centralizing another layer of it under one vendor.

$95B backlog sounds huge until you ask how many quarters it actually takes to convert into revenue at current build rates. Order books grow every cycle right before capacity catches up, the number that matters is the conversion pace not the headline size.

If MSCI actually goes through with this the passive-flow selling pressure could be a bigger deal for MSTR near term than anything Saylor says in a press response. Curious how much index ownership is even at stake here.

Upsizing from $15B to $20B mid-raise usually means demand was there to take more, but it also raises the question of whether $15B was ever going to be enough in the first place. If they're back at the well again in another year I won't be surprised.

Genuinely curious how much of that $25B guide raise is already priced into the stock after a 266% run this year. Raising guidance after a huge rally is a different signal than raising it off a flat base, the bar for the next print just got a lot higher.

Nvidia and Marvell both got punished on margins even with beats, so the read-through I keep coming back to is that memory suppliers are on the other side of that trade right now, they're the ones with pricing power while the AI chip names are eating the cost. That dynamic doesn't reverse overnight.

today's +5% already has some of Wednesday priced in ngl, curious if it holds or if this turns into a sell the news setup like some of the other launch events did

Genuinely curious how much a CEO handover actually moves a company this size versus just being a headline event. Apple's org is deep enough that Ternus stepping in probably doesn't change quarter to quarter execution much, the real test is capital allocation decisions a year or two out, not day one.

The 16 billion AI number is the one everyone is anchored on, but I'm more curious about VMware subscription conversion since that's been the quieter growth driver nobody talks about on earnings day. If both numbers land, this print should actually hold up unlike the last two chip earnings this week.

Curious whether the market actually cares more about Ternus as CEO or about the foldable itself. My guess is the leadership change matters way more long term but gets almost no attention this week because everyone is focused on hardware. Watching what he says in his first public moment more than the phone.

Analysts are split on this one which is kind of interesting, some calling it a real catalyst and others saying the event itself is underwhelming since the hardware was already shown. I think the disagreement is actually the tell, if it was obviously good or obviously bad everyone would agree already. Watching the FSD demo specifically, not the car design.

1.4 trillion is more than Meta's entire market cap so it was never a real number, just an opening anchor. What I'm actually watching is whether more states pile onto the lawsuit, that would be the thing that actually moves the stock further.

Actually didn't expect SK Hynix specifically to be the one looking at Intel's packaging. They compete directly with Micron and Samsung in HBM, so if EMIB genuinely saves them cost versus CoWoS, that's a real business reason, not just PR for Hot Chips week. Watching to see if this turns into an actual order next year.

Everyone calling this a bear market after one bad week from Samsung. Feels early. SK Hynix and Nvidia reportedly locked in a multi year DRAM and HBM supply deal per an Edgewater note this week, which is not the kind of thing companies sign if they think demand is rolling over. Watching to see if this holds.

Meta was around 590 on Friday. closed 565 on Monday.

so the market saw a trial with a trillion dollar headline on it and took 4% off. that's the whole reaction 🤷

either everyone decided the number is theatre, or nobody has actually worked out what a forced teen redesign does to ad revenue.

i lean toward the first. but 4% off isn't the kind of discount that makes me want to size up.

watching, no action.

not touching my nvda before AMD reports. whatever prints tonight i still own the toll road, so i just wait 💤

google cloud grew 82% to 24.8b which is genuinely absurd for that scale, but reported eps was flattered by unrealised gains and operating eps is nearer 2.90. half the posts i see are quoting the headline number. read the composition first, then decide if it is cheap