There are no alternatives to Nvidia GPUs if you want to efficiently train an AI, they have an effective monopoly.
Personally i would have invested in TSMC/ASML or Intel as this point. Buy the shovel maker.
>>TSMC
Relatively low margins, few good well negotiating customers. Can't grow fast.
>>ASML
What do you think market could have missed there? Good company with simple business model. Very unlikely to grow fast.
>>Intel
Terrible company with long track record of stupid decisions and deception. What make you think they will suddenly turn around and start innovating?
That's why my bets are on them.
Everyone is reliant on ASML and unlikely to change.
Intel has fabs and is too big to fail, the us government won't let it happen.
I'm making a safe bet on the long term, not a speculative one based on a technology that will or won't be worth that valuation.
After all, if it's election day, my analysis says FooCorp is worth $50 under a Biden administration and $100 under a Trump administration, and the candidates are neck-and-neck, that means the shares are worth $75. But tomorrow the shares will have either gained or lost $25.
But as detailed recently in Going Infinite, Jane Street's losses around the 2016 election were because they didn't know outcome was worth the example $50 and which was $100.
Figuring out which ones, and to what extent, is an unorthodox but effective path to wealth.