Busy couple of weeks on results front
-
KLA Corporation has been on the move recently, +22% YTD due to accelerated CoWoS packaging capacity build out at TSM, who have moved plans to build the largest packaging facility in the world and install 2,000 machines. Plus the further shrinking of logic and DRAM to the N2 node means increased need for KLA's process control, metrology and inspection solutions.
As we have discussed previously, the market doesn't fully understand the drivers of KLA revenue. Their services revenue is driven by an installed based of 50k machines and customers pay a reoccurring fee for ongoing support. KLA has a 96% renewal rate making this segment very sticky. Further > 50% of said machines are over 18 years old so clearly any China sanctions are not directed at this level of technology.
TSM has signalled 'significant capex' in the area of advanced packaging, an area which KLA dominates.
The company is also repositioning itself towards silicon carbide and gallium nitride management systems where the costs involved are very high and there is a drive for zero defects.
@Adam-Kay said in Busy couple of weeks on results front:
KLA Corporation has been on the move recently, +22% YTD
Apologies for the muppety question: is that +22% in the last 16 days or the last 365 days?
-
Thanks.
-
A quick update on the AI Chip restrictions.
The limits on the export of advanced AI chips were established through an interim final rule, they will not take effect for one year and carry a 120-day comment period. This time allows the Trump administration to modify the regulations if it sees fit. If Trump’s officials make no changes, the rules will come into force in January 2026 as written.
Note: Nothing in the amendments addresses any countries access to cloud-native AI such as that offered by many companies around the globe.
Have a good wend, all
Adam
-
Yes,solid day and solid start to the year.
-
Hi Wibble and thank you. It is very much a Team effort. The market is very unpredictable over 'weeks', any short period in fact, whereas over the longer term prices reflect business fundamentals. You only need to look at KLA Corporation which I have mentioned recently. The stock price softened due to investors sentiment around China, however were wrong about the magnitude of the impact. And more recently it appears they will actually grow earnings over the coming years due to their expertise in the sub 3nm node. Something if you look back, we have been saying for the last 2 years.
Long term, good businesses will reward patient investors via earnings and cashflow growth coupled with compounding (increased dividends and reinvested profits).
We will be making a couple of changes in the coming weeks, once the IC has met to discuss them further.
It is always nice to see new highs but as always past performance is no indication of future returns.
-
A quick update on the AI Chip restrictions.
The limits on the export of advanced AI chips were established through an interim final rule, they will not take effect for one year and carry a 120-day comment period. This time allows the Trump administration to modify the regulations if it sees fit. If Trump’s officials make no changes, the rules will come into force in January 2026 as written.
Note: Nothing in the amendments addresses any countries access to cloud-native AI such as that offered by many companies around the globe.
Have a good wend, all
Adam
@Adam-Kay said in Busy couple of weeks on results front:
Nothing in the amendments addresses any countries access to cloud-native AI such as that offered by many companies around the globe.
I had this very thought a few weeks back, the hardware doesn't have to be in China to be used by the Chinese, so the restrictions seem fairly easily circumvented.
-
Absolutely correct Ron, as it relates to rogue states(use)This is why it's impossible to police.
-
Hi
US markets will be closed today due to it being Martin Luther King Jr Day. FX markets are open and valuations tomorrow will reflect those movements where applicable.
Regards
-
Most of the ones I have on a watchlist are climbing today, not Apple but all the rest.
I reckon 9/10 chance that Trump will be good for our investments. But the 1/10 option may involve actual apocalypse rather than just some numbers dropping on a spreadsheet
. -
I take a view that Govt policy or governing style have a short term effect as investors speculate on the impact. Long term it's business fundamentals. The US want to increase wealth and are pro business.
Trump policy:
deregulation and tax cuts. America first.
Drill Drill DrillUK policy:
High taxes
strong labour laws
being a good global citizen(ethics). At any cost
equality -
Netflix reported earnings tonight. Record growth, record subscriber numbers and a very big pop after hours to circa usd 1,000.
-
Sam Altman, Open Ai chief joined President Trump to announce the Stargate Project, A $500B data centre/super computer initiative spanning 4 years. Key partners are NVDA, ORCL, OpenAi, ARM, MSFT and Softbank. We will deploy $100B immediately. Altman went on to say 'this infrastructure will secure Americas leadership in AI, create in excess of 100k jobs and create massive economic benefits for the entire world.
The build out is in progress now, starting with Texas. The entire stack will operate on Nvidia hardware and software.
-
Well, it's interesting looking back at the commentary around certain tech names vs the wider noise, naysayers and speculation. I can assure you those that were negative will continue with the same comments. Classic human nature, an inability to change ones mind or simple envy. Take your pick.
The area where there is the most misunderstanding is, the built out of AI is required to make machines a LOT smarter than they are today. A machine 100X smarter than a human. We aren't even 1X yet. Further the foundational models build on one another. You can't build chatgpt 10 without building 5 through 9. Further there is simply not the GPU or HBM or the rack scale capacity or the power supply to build it even if you wanted too. Stargate looks to be around 5M GPU, it will vary because one must assume chips being used which haven't been built like Ruben next. Masa Son speculated 20M would be needed to run the known models being built today. This is just the chatgpt, llama, Claude, PaLM models. Many times more will be needed for robotics, autonomous cars, enterprise and sovereign solutions.
-
The build out(progress) is an investment in the future, no different to any other major infrastructure. The goal is it will pay off significantly for the early investors. And the point I'm making is, we don't have to defend our position because it is evidence based. Evidence that shows clearly that the main players in this space have a very long runway for growth. A runway that is clearly much bigger than even the most bullish actually thought.
