Institutional TMT research, made public. AI, semis, data & macro. Daily pre-market brief. Full research reports on Substack
Manhattan, NY
Joined October 2021
- Tweets15.6K
- Following1.2K
- Followers29.2K
- Likes76.8K
*META POSTS ‘NEGATIVE’ FCF FOR THE FIRST TIME IN A DECADE
It's spending every dollar it makes on AI GPUs, memory, data centers, power. Meta's cash flow is now $NVDA, $MU, $TSM and $ASML's revenue. Companies don't spend like this when they doubt demand…bullish AI trade!!!
If you’re a $META shareholder, here’s how this works now:
Same as $GOOGL in 2014, every shareholder gets one new non-voting Class C share for each share they own.
Same total value, split into two…one keeps the vote and the new one doesn't. It works like a 2-for-1 split, Zuckerberg's control doesn't change, and it takes effect January 19.
Just like google. That's why $GOOGL (voting) and $GOOG (non-voting) both trade today.
The difference is that Meta says it wants a non-voting currency to fund AI, acquisitions and engineer pay. With capex heading to $135–145B…
Interesting…more capex coming folks!! Limit up on AI!!!
Nicholas Mugalli retweeted
People are buying FSD with a car attached, not the other way around
Tesla has just released its Q3 earnings results.
Highlights:
• EPS (non-GAAP): $0.48 (vs. $0.45 est)
• Revenue: $28.05B (vs. $28.3B est)
• Adjusted EBITDA: $3.90B
• Active FSD Subscriptions: 1.74M, up 67% YoY from 1.04M
• Net income (GAAP): $787M
• Free cash flow: $1.53B
• Capex: $4.73B
• Cash & marketable securities: $45.45B, up 9% YoY
*ZUCKERBERG: $META EXPECTS MUSE WILL REACH AN EVEN HIGHER LEVEL OF MONETIZATION PER USER THAN CURRENT FOA
*ZUCKERBERG: BILLIONS OF MUSES AT SCALE WILL GET DONE AND WORK TOGETHER
*ZUCKERBERG: THE MOST COMMON USE FOR MUSE IS SHOPPING
*ZUCKERBERG: AVERGE MUSE PURCHASE SIZE IS $200
*ZUCKERBERG: $META’s THREAD HAS OVERTAKEN X IN MONTHLY ACTIVE USERS
If you’re a $META shareholder, here’s how this works now:
Same as $GOOGL in 2014, every shareholder gets one new non-voting Class C share for each share they own.
Same total value, split into two…one keeps the vote and the new one doesn't. It works like a 2-for-1 split, Zuckerberg's control doesn't change, and it takes effect January 19.
Just like google. That's why $GOOGL (voting) and $GOOG (non-voting) both trade today.
The difference is that Meta says it wants a non-voting currency to fund AI, acquisitions and engineer pay. With capex heading to $135–145B…
Interesting…more capex coming folks!! Limit up on AI!!!
If you’re a $META shareholder, here’s how this works now:
Same as $GOOGL in 2014, every shareholder gets one new non-voting Class C share for each share they own.
Same total value, split into two…one keeps the vote and the new one doesn't. It works like a 2-for-1 split, Zuckerberg's control doesn't change, and it takes effect January 19.
Just like google. That's why $GOOGL (voting) and $GOOG (non-voting) both trade today.
The difference is that Meta says it wants a non-voting currency to fund AI, acquisitions and engineer pay. With capex heading to $135–145B…
Interesting…more capex coming folks!! Limit up on AI!!!
WOW!!! $META revenue was $64.6B vs $63.4B expected, above the top of its own guide and up ~26% Y/Y. Ad revenue alone was $63.1B. The engine isn't slowing.
The Q4 guide is even better than anybody expected $73–76B, a $74.5B midpoint vs $73.7B expected. Meta is telling you growth holds into year end.
On capex, Meta only raised the bottom of the range, to $135–145B from $130–145B. That's Meta saying it will spend at least what it planned, not that the budget is blowing out. And it comes with a revenue guide above the Street, which is exactly the trade we said we wanted more spend, more growth. And 1/2 share split…for every shareholder huge!!
Ignore the $3.83 EPS. That's the ~$10B legal charge everyone knew about. Overweight, $1,004 PT.
*META SEES FY CAPEX $135B TO $145B, SAW $130B TO $145B
There it is… $NVDA should be happy?
WTS: $META reports after the bell. The Street wants about $63.3B of revenue, up 24% from a year ago and near the top of Meta's own $61–64B guide, with adjusted EPS around $6.50. Reported EPS will look much worse because of a roughly $10B legal charge tied to the state attorneys general settlement. That charge was flagged weeks ago. It hits the income statement, not the ad business which matter most here.
The numbers that actually move the stock today are capex and the outlook. Meta is guiding to $130–145B of capex this year, with about $41.7B expected this quarter and the stock's worst days in 2026 came when that number went up. For the 4th quarter, the Street is at $73.6B of revenue. If Meta spends more, that guide needs to rise with it. Muse, Meta's new AI agent, is the wildcard after 5M+ downloads in its first three weeks. We're Overweight with a $1,004 price target, 41% upside.
$LRCX posted the biggest quarter in its history $8.22B of revenue, up 54% Y/Y, and $2.20 of EPS, up 75% and stock is down 6%.
Both beat and both landed above the midpoint of guidance. Two weeks ago this stock had the longest losing streak in its history.
Yes, the beats were small, and after the run this group has had, the Street wanted a blowout. But it's a record on top of a record with revenue up more than 20% sequentially. That's not a cycle ending as bears like to frame it.
The number that decides the stock is the December quarter guide. If Lam guides up again on the call, the selloff was a gift.
On to the call!!
WTS: $LRCX reports after the bell and is guiding to its biggest quarter ever: $8.1B of revenue, give or take $400M, up 21% from June's record $6.72B, with EPS of $2.15 against $1.82 last quarter. Consensus sits right on top of that, at about $8.1B and $2.18. Margins are expected near the highs with gross margin around 52% and operating margin around 39.5%.
But the strange part is the stock just had the longest losing streak in its history and sits about 39% below its June record. That was the whole equipment group selling off after Applied Materials declined to forecast industry spending. Nothing about Lam's business has changed. $ASML and $TSMC both told us this week their customers need more tools and are paying more for them. The number that matters most tonight is the December quarter guide. We're Overweight with a $355 price target, 35% upside.
$TSLA revenue was a hair light at $28.05B vs $28.13B but that's the only miss, and it's a rounding error. Everything that matters beat.
Gross margin was 19.6% vs 17.8% expected, almost two full points ahead and up from last quarter's slide. That was the number we said would decide this print, has the bleeding stopped? It hasn't just stopped imo, it's reversed.
Automotive revenue was $20.56B vs $19.83B, a $700M+ beat on cars alone. The 486K deliveries showed up exactly where we said. The small top line miss came from outside autos, most likely energy, which had already flagged lighter storage deployments.
EPS was $0.48 vs $0.44, and that's after last quarter's $0.33 miss. The car business is healthier than the bears wanted to believe. Now the call is about Cybercab, FSD and Optimus. Overweight, $697 PT. Stock should be happy.
WTS: $TSLA reports after the bell with a low bar: about $28.1B of revenue, roughly flat from a year ago, and adjusted EPS near $0.45, down from $0.50. Last quarter EPS was $0.33, a $0.20 miss, so even a clean print would be a step up. The most important number is already out. Deliveries came in at 486,532 vs about 462,000 expected, a 5% beat, and Tesla sold 22,000 more cars than it built.
The line to watch in the report is auto gross margin. The Street expects around 17%, up from 16.3% last quarter. If it holds there, the bleeding has stopped. Energy storage deployed 13.7 GWh, up 10% from a year ago. Capex is running at about $7.8B this quarter on the way to more than $25B for the year. But the call will matter more than the numbers: Cybercab beyond Austin, unsupervised FSD and Optimus' first customer. We're Overweight with a $697 price target, 80% upside.
WOW!!! $META revenue was $64.6B vs $63.4B expected, above the top of its own guide and up ~26% Y/Y. Ad revenue alone was $63.1B. The engine isn't slowing.
The Q4 guide is even better than anybody expected $73–76B, a $74.5B midpoint vs $73.7B expected. Meta is telling you growth holds into year end.
On capex, Meta only raised the bottom of the range, to $135–145B from $130–145B. That's Meta saying it will spend at least what it planned, not that the budget is blowing out. And it comes with a revenue guide above the Street, which is exactly the trade we said we wanted more spend, more growth. And 1/2 share split…for every shareholder huge!!
Ignore the $3.83 EPS. That's the ~$10B legal charge everyone knew about. Overweight, $1,004 PT.
WTS: $META reports after the bell. The Street wants about $63.3B of revenue, up 24% from a year ago and near the top of Meta's own $61–64B guide, with adjusted EPS around $6.50. Reported EPS will look much worse because of a roughly $10B legal charge tied to the state attorneys general settlement. That charge was flagged weeks ago. It hits the income statement, not the ad business which matter most here.
The numbers that actually move the stock today are capex and the outlook. Meta is guiding to $130–145B of capex this year, with about $41.7B expected this quarter and the stock's worst days in 2026 came when that number went up. For the 4th quarter, the Street is at $73.6B of revenue. If Meta spends more, that guide needs to rise with it. Muse, Meta's new AI agent, is the wildcard after 5M+ downloads in its first three weeks. We're Overweight with a $1,004 price target, 41% upside.
WTS: $TSLA reports after the bell with a low bar: about $28.1B of revenue, roughly flat from a year ago, and adjusted EPS near $0.45, down from $0.50. Last quarter EPS was $0.33, a $0.20 miss, so even a clean print would be a step up. The most important number is already out. Deliveries came in at 486,532 vs about 462,000 expected, a 5% beat, and Tesla sold 22,000 more cars than it built.
The line to watch in the report is auto gross margin. The Street expects around 17%, up from 16.3% last quarter. If it holds there, the bleeding has stopped. Energy storage deployed 13.7 GWh, up 10% from a year ago. Capex is running at about $7.8B this quarter on the way to more than $25B for the year. But the call will matter more than the numbers: Cybercab beyond Austin, unsupervised FSD and Optimus' first customer. We're Overweight with a $697 price target, 80% upside.
WTS: $LRCX reports after the bell and is guiding to its biggest quarter ever: $8.1B of revenue, give or take $400M, up 21% from June's record $6.72B, with EPS of $2.15 against $1.82 last quarter. Consensus sits right on top of that, at about $8.1B and $2.18. Margins are expected near the highs with gross margin around 52% and operating margin around 39.5%.
But the strange part is the stock just had the longest losing streak in its history and sits about 39% below its June record. That was the whole equipment group selling off after Applied Materials declined to forecast industry spending. Nothing about Lam's business has changed. $ASML and $TSMC both told us this week their customers need more tools and are paying more for them. The number that matters most tonight is the December quarter guide. We're Overweight with a $355 price target, 35% upside.
WTS: $LRCX reports after the bell and is guiding to its biggest quarter ever: $8.1B of revenue, give or take $400M, up 21% from June's record $6.72B, with EPS of $2.15 against $1.82 last quarter. Consensus sits right on top of that, at about $8.1B and $2.18. Margins are expected near the highs with gross margin around 52% and operating margin around 39.5%.
But the strange part is the stock just had the longest losing streak in its history and sits about 39% below its June record. That was the whole equipment group selling off after Applied Materials declined to forecast industry spending. Nothing about Lam's business has changed. $ASML and $TSMC both told us this week their customers need more tools and are paying more for them. The number that matters most tonight is the December quarter guide. We're Overweight with a $355 price target, 35% upside.
WTS: $META reports after the bell. The Street wants about $63.3B of revenue, up 24% from a year ago and near the top of Meta's own $61–64B guide, with adjusted EPS around $6.50. Reported EPS will look much worse because of a roughly $10B legal charge tied to the state attorneys general settlement. That charge was flagged weeks ago. It hits the income statement, not the ad business which matter most here.
The numbers that actually move the stock today are capex and the outlook. Meta is guiding to $130–145B of capex this year, with about $41.7B expected this quarter and the stock's worst days in 2026 came when that number went up. For the 4th quarter, the Street is at $73.6B of revenue. If Meta spends more, that guide needs to rise with it. Muse, Meta's new AI agent, is the wildcard after 5M+ downloads in its first three weeks. We're Overweight with a $1,004 price target, 41% upside.
WTS: $META reports after the bell. The Street wants about $63.3B of revenue, up 24% from a year ago and near the top of Meta's own $61–64B guide, with adjusted EPS around $6.50. Reported EPS will look much worse because of a roughly $10B legal charge tied to the state attorneys general settlement. That charge was flagged weeks ago. It hits the income statement, not the ad business which matter most here.
The numbers that actually move the stock today are capex and the outlook. Meta is guiding to $130–145B of capex this year, with about $41.7B expected this quarter and the stock's worst days in 2026 came when that number went up. For the 4th quarter, the Street is at $73.6B of revenue. If Meta spends more, that guide needs to rise with it. Muse, Meta's new AI agent, is the wildcard after 5M+ downloads in its first three weeks. We're Overweight with a $1,004 price target, 41% upside.
If you’re a $SPCX or a neocloud investor, save this!!
BofA: Revenue through 3Q totaled $27.1bn, implying 4Q revenue of ~$19.8bn to reach BofAe FY26 $46.9bn up ~36% q/q…We expect investors to focus on SI capex returns as compute scales toward management's 2 GW year end target🔥
If you’re a memory investor, you’ve got to read this report!!
Micron just held its tech webinar today and one line should stop every memory bear in their tracks…Micron expects the shortage to last at least two more years and it can't see when supply and demand will balance.
The reason is that every wave of AI eats more memory than the last. Training took HBM per GPU up 3.6x from H100 to B300. Inference doubled memory per server in three years. AI agents are growing context about 30x a year. And robotaxis will need 10x+ the DRAM of today's cars.
The roadmap micron showed earlier is built for it. 1-delta DRAM ramps in 2H27 with ~25% better density. G11 NAND crosses 400 layers in 2027 at ~2x G9's density. And by 2028, vHBM stacks memory directly on top of the processor, the shortest path to compute that exists.
We know Samsung and CXMT push 4F2, Micron says it's a dead end and is going straight to 3D DRAM. If they're right, their rivals are spending billions down the wrong road. We think they're right. Overweight $MU…
If you’re a $SPCX or a neocloud investor, save this!!
BofA: Revenue through 3Q totaled $27.1bn, implying 4Q revenue of ~$19.8bn to reach BofAe FY26 $46.9bn up ~36% q/q…We expect investors to focus on SI capex returns as compute scales toward management's 2 GW year end target🔥
OMG!!🔥
JP Morgan just put a super bullish note on $SPCX rising its price target to $500 (from $240) which would be a $6.6 trillion mkt cap or the largest company in the history of humanity
Elon said SpaceX value is probably double of earth’s value & now started to look real!
Brad Gerstner lays out the case why $SPCX is going to eat $CRWV alive and that’s the reason he sold his $650M of CoreWeave a month ago
“The collateral that CoreWeave is putting up which is really only GPUs is not good enough anymore. They used that in the past to get cheap debt”
OMG!!🔥
JP Morgan just put a super bullish note on $SPCX rising its price target to $500 (from $240) which would be a $6.6 trillion mkt cap or the largest company in the history of humanity
Elon said SpaceX value is probably double of earth’s value & now started to look real!
OMG!!🔥
JP Morgan just put a super bullish note on $SPCX rising its price target to $500 (from $240) which would be a $6.6 trillion mkt cap or the largest company in the history of humanity
Elon said SpaceX value is probably double of earth’s value & now started to look real!
Morgan Stanley: $SPCX +9% guides revenue to $33-37bn. With greater visibility into the compute contract pipeline over the next 6 months mgmt set a range w a $35bn midpoint which is ~52% above consensus of $23bn & $500bn-$1tn ARR as Musk expects compute ramp & GrokBot adoption🔥🤯
*HUANG: $NVDA GROWING AI PRESENCE IN EUROPE WITH 12,000 COMPANIES IN IYS ECOSYSTEM
*HUANG: $NVDA INVESTED $10B OVER THE PAST YR AND 14 STARTUPS THIS WEEK
🇪🇺
SCOPE: From Anthropic’s draft IPO filing as per WSJ:
> Net loss, H1 2026: $80B+
> Of that, ~$72B is an accounting charge on convertible notes it used to raise money
> That’s a paper loss (market to market) not cash that left the bank
> The remaining ~$8B covers everything else
The other side:
> Revenue in H1: $16.3B
> Q3 revenue reportedly $16.8B, with adjusted profit and positive cash flow
Huge:
> At least $518B committed to computing and infrastructure over ~10 years
> About 80% can’t be cancelled
> Anthropic plans to spend $111 billion on $GOOGL cloud computing contracts, $110 billion with $AMZN Web Services, and $31 billion on $MSFT "regardless of actual usage."
> Anthropic agreed to spend up to $86 billion renting Nvidia chips from smaller AI cloud providers Nscale and Lambda through 2033 and 2034, respectively, in addition to other cloud deals.
So now that $80B is not issue, its whether revenue can keep growing fast enough to pay for half a trillion dollars of computing... BULLISH af
