Stupid observation but especially in semis when you gotta truly change your mind a lot, there gets a point where you gotta just flip your opinion if you either feel victory lappy or very confident in your belief. Auto and LSCC is mine like asap. Just GFL. It always happens, you have a thesis go right, and you get to have like a gloat quarter, and then in your moment of max comfort you gotta get paranoid.

I like everyone else think NVDA numbers next year are too low, but I feel weirdly complacent in that one. Should still work.

Jensen's keynote thread: 1/n I think the comparison of "how much things have changed in 5 years" was a good demo, but it really doesn't really flex what *really* has changed IMO

Nvidia Earnings Thread: $NVDA

Let's see the most important AI company in action.

First: *NVIDIA SEES 2Q REV. $11.00B PLUS OR MINUS 2%, EST. $7.18B

1/n Guidance:

Revenue $11.00B +/- 2% vs FactSet $7.17B
Non-GAAP gross margin 70.0%, +/- 50bp vs consensus 66.6%

There's no way EPS isn't a fucking face smash, wow there is just no way to have seen that candle coming.... jfc

Time to time, I flip through Sumco's earnings presentation and I'm like honest this is pretty good. These are forecasts, but their top-down from a wafer perspective is helpful

First is global data traffic server shipments in millions of units - notice the 2023 down year

$AMD

Q4:
Rev $5.60B vs FactSet $5.51B
gross margin 51% vs guidance of 51%

Q1 Guidance:
Rev $5.3B +/- $300M vs FactSet $5.50B
YoY the Client and Gaming segments are expected to decline, partially offset by Embedded and Data Center segment growth.
Non-GAAP gross margin ~50% Segments vs consensus - so a slight DC miss, slight client miss, gaming beat, embedded beat

Data Center $1.66B vs FactSet $1.72B
Client $903M vs FS $1.02B
Gaming $1.64B vs FS $1.53B
Embedded $1.40B vs FS $1.33B

IEDM Interesting Graphic Thread

(I won't supply explanations - just squirreling this away for later) Data center percentage of direct compute

I am almost positive that what's going to happen is that semiconductor OEMs in their desire to clean the channel is going to overshoot it - and then there will be a "shortage" again lmao. Not like '20 - but a quarter or 2 where there's a massive catch up. The expectation is that Q4/Q1 is going to be the mother of all channel scrubbing, everyone is going to be draconian.

I don't know when things get better - but if I had to guess 2H23 looks like as good as any time, presume the fed hike / strong dollar can't go to infinity

TSMC notes 1/x

Revenue growing ~36%, gross margin up to 60%. GM driven by FX + cost improvement.

Revenue by node: 28% of revenue is now 5nm, 7nm is 26% 2/x

revenue by platform! This is interesting! Last Q HPC was largest, this Q smartphone snaps back. I would guess that's something of a Nvidia impact. This also doesn't bode super well for HPC - its supposed to beg growing faster than this.

Also IoT is bonkers - accelerating!

ir.appliedmaterials.com/static-files/7…

"Chips have more than 50 miles of interconnect wiring" resistance increases exponentially as wiring scales

interesting - does this mean more wiring is needed (am a real dumbass in this way)

so 15 eps x 22 = 330 stock - a bit more than a 5 year double. sounds about right rev cagr - here are some of the higher growth segments, blends to 7-10% rev

Let's tell a little story about how I came to semicap as an investor. At my old firm, we had a list of qualitative things we wanted to see in biz before we bought

secular growth
oligopolistic+ industry
pricing power
positive FCF and returns that to shareholders 1/ every day I would pretty much pound the pavement looking for these attributes. seriously.

ASML was the first one i was like wow this is monopoly, secular, has nutty pricing power, great FCF + buysback shares. Then I looked next door and saw the rest of semicap.

The earnings call is one of the oddest calls I have ever been on. 1) Chairman is on the call 2) the defensiveness of the call is absolutely wild.

It seems like a shame performance by Paul. I will post some of the weirder stuff in a thread.

https://twitter.com/_fabknowledge_/status/1484981175462424577

I think that someone at Soitec at least read my piece. I love how they use "conspiracy theory"

at least on the point of influence on the board. Yo I laid out why the independents were not independent. 2 of the 5 are very in the pocket lol.

What looked like a series of power moves actually was France nationalizing Soitec in plain sight! This story has it all, board drama, management rioting, and importantly China.

thread of the full events going on at Soitec right now 🧵 1/ Soitec shares got pummeled after the announcement of a CEO change. Most investors were extremely confused by the action. Management even sent a letter that they strongly disagree with the Board. What is happening!?

bloomberg.com/news/articles/…

TSMC reported earnings last night and their results shocked most investors. Why? Buckle up and let's talk about one of my favorite companies in the world: TSMC. 1/ the biggest surprises were revenue guidance and capex guidance.

Let's start with revenue guidance - TSMC guided to ~16.6b-17.2b in Q1 or 6% over Q1 consensus. TSMC believes they can do "mid to high 20s" revenue growth

importantly they are reporting ACCELERATING REVENUE!

We are on the edge of earnings season and it's feeling palpable. Two headlines that I thought were interesting.

VAT Group preannounced ahead of street orders accelerated meaningfully. 20% of orders were 1 time in nature - but that's a meaningful acceleration ex that. Not quite a pre-release but $ONTO put out a total backlog number to the tune of ~$500 million. $100 million of it just for packaging for 2.5d/3d packaging. $100m in orders for new JetStep Lithography system (PCB litho iirc)

Bodes well for $CAMT as well ofc.

Semiconductors can be a hard industry to understand, especially for investors without a technical background.

Here are a few of the books that really helped me understand semiconductors better. In order of approachability - a 🧵 1/ In the beginning there was Fairchild.

I think that is the best way to start. Makers of the Microchip by Christophe Lécuyer is the best book to begin with IMO.