My opinion on risks for $NVDA, a ๐งต:
1) Their main customers for high margin AI systems are big tech. They are extremely motivated to develop their own chips ( $GOOG, $MSFT, $AMZN ) and offer it as a service via cloud offerings. The economic moat of $NVDA might not be strong.
- South Koreans cashing out life insurance to go long
- PDT removed
- SpaceX IPO @ $1.7T evaluation
- OpenAI / Anthropic IPO @ ~ $800B
Buffet said that markets are frothy a lot of non-profitable companies IPOing. But this market, can see ATH again in a month. It's hard to trade.
AI infrastructure companies when it comes to energy source:
- nuclear plants take 10 years+, always over budget (find me a nuclear power plant built in the west recently on time and on budget)
- solar / wind / batteries deploy in < 1 year and prices have been going down for years
AI companies now will look to optimize cost / make it more efficient. The business model in terms of capex and opex require it for this AI thing to be sustainable in the long term, which plays into why hardware is a cyclical business.
The next phase of AI will be diffusion and segregation. Diffusion is how AI gets into every nook and cranny of everything. First it'll be businesses that are able to use AI to get massive productivity gains. However during that time, there will be a significant job loss.
My thoughts on AI: so we've seen a lot of hardware stocks rally and I think we're past the exponential rally phase. DRAM and NAND will have other players enter the market fairly soon, like CXMT. In addition, hardware is a cyclical business, regardless of super cycle or not.
Signs of diffusion are companies that are laying off people now, investing in AI in terms of training and usage, followed by productivity improvements. Signs of segregation are companies that are developing tools to label AI traffic and provide options to limit AI interaction.
During that period of uncertainty, I think we will cling on to our humanness. Where we actively consume products and services made by humans and not AI. Similar to car purists prefer an engine vs electric, how people prefer vinyl over mp3s, but more intense.
During this time there will be a period of transition and uncertainty with social unrest. I would consider it a transition period of AI replacing jobs and then new types of jobs would be created. It's unclear how long this would take and how chaotic the progress would be.
My thoughts on AI: so we've seen a lot of hardware stocks rally and I think we're past the exponential rally phase. DRAM and NAND will have other players enter the market fairly soon, like CXMT. In addition, hardware is a cyclical business, regardless of super cycle or not.
If there's a lack of power capacity for AI in the US, chances are they're going to find an off-shore location, at least for R&D training as it's the best risk to reward. They're not going to sit on their ass for 10 years on a nuclear power plant. Canada's hydro is a better option
4.5% on US 10y bonds, 5% on US 30y bonds. Time to pay attention. In my opinion there's 2 way this ends. The good way is to increase assets or inflation.
4.5% on US 10y bonds, 5% on US 30y bonds. Time to pay attention. In my opinion there's 2 way this ends. The good way is to increase assets or inflation.
USDJPY at 160. Japanese 10 year bond at all time high. They have currency swap with the US and the largest foreign Treasury holding. Given the current macro environment, things are about to get interesting.