@manch thanks for singling me out to be the biggest NVDA bull. Ha ha, don't forget this is a RE forum. I am a big bull on bay area real estate instead of any stock. At the end, no matter it is NVDA or AMD or INTC or GOOG takes lead, I am going to win as they are all in bayarea. I am just not as brave as you to take so much debt, so I do not have any spare money for stock anymore...
Thanks for the audio, it is very helpful. Get back to NVDA, I am going to be very concise and very high level here:
We are talking about an once a century opportunities
- Chips can be used to communicate or to compute. All previous technologies are driven by communication. Now it is time for computation to take over.
- It is more than sequential vs parallel. It is about how you are created and how the universe is functioning.
- We use to able to model intelligence communications by languages + instructions which is intrinsically sequential.
- Now we can model how you perceive the reality (as neuron by AI or Deep Learning) and the reality itself (as all physical laws of matter and energy by VR). Both are all intrinsically in parallel.
- Really, AI is just a small part of the "parallel revolution". It is going to eat everything.
Jensen Huang is a real crazy guy
- Investor hated him as he invested all the profit into his crazy visions and many of them failed.
- He is a visionary, a fighter and will win no matter what. He is like Andy Grove of Intel, Bill Gate of MS as a cofounder CEO. Intel, AMD lack this.
NVDA is a software company, far more than just a hardware company
The same architecture everywhere
- From huge processor used in data center to little tiny chip (jetson) used in any small devices and anything in between.
- CUDA (GPGPU), not optimized but can be used any where in any industries. That is the unfair advantage for NDVA.
- In every usage of the "parallel revolution", competitors could always optimizes in a particular usage, like Google's TPU. That only validates the market. If the market is important enough, NVDA could join to optimize, like TPU in Volta, it will not be too late.
Open source, open platform
- I agree. The moat of NVDA is open source and open platform.
- We are in a different world from a decade ago where patent portfolio may or may not be helpful anymore.
- While many low level stuffs by NVDA is still proprietary, they also open sourced many softwares on top so that the world will codevelop with them.
There will be risk, a lot of risk in investing company like this. That is why I am still stick with RE.