“But it’s the software that’s killing everyone.” He notes even Nervana’s founders initially built their technology by testing on Nvidia GPUs, which to Mosesmann is a sign of the pre-eminence of CUDA. "The ecosystem matters at this stage” in A.I.’s evolution, he says.”
That is a prototypical Geoffrey Moore Gorilla, just as Intel, Microsoft, Oracle, ARM, Apple, and a select few others were. They build and own an open but proprietary architecture that an entire industry standardizes on.
Apple is a bit different as they are more closed than the rest, but they have managed to make it open with IOS, and the rise of the browser vs. the OS has made Apple laptops more accessible, thus the superior stability of their OS could start to succeed and gain large chunks of marketshare against Microsoft that was impossible prior to browser and mobile changing the computing paradigm.
Here we have GPUs changing the computing paradigm. As Nvidia states, the more GPUs you buy the more you save, precisely because GPUs are so much more efficiently per processing unit. The CPU is still necessary, but int he data center, the more concentrated the GPUs via the CPU the more efficient the system.
GORILLA. I said it last year at less than half the price (and yes, a good chunk of the only 3 stocks I held at the time, then to 2, now to X) and it is even more true now.
Prognostication is about spotting patterns. This is a rare pattern. It is though.
In regard to future valuations, 45% earnings growth over 6 years is 9x. I think that is aggressive enough. And yes, that indeed could happen with ray tracing and edge computing to go along with the data center.
And as the article you linked to states, they are building complete systems, with switches (Arista) and data storage (PSTG) (of course they will work with anyone, but that is their current partners with their state of the art) and as the article also mentions, perhaps 6 years ago GPUs would not be considered the ideal solution for AI, but who cares. Neither was Windows, nor x86, but they set the standard.
NVDA is going to have a lot of market volatility. Cryptocurrency is rocking their share price far, far, far, far, more than it is so worth. But there goes the markets passions. Ignore it, this volatility, remember where this company is going as they hold the winning worldwide standard for the key enabling technology for the rest of many of our lifetimes (albeit may be quite a long time for me 
unexplored territory and what not) or at least, if not that, the next decade or two.
Sure, sure quantum computing, other breakthroughs. Not this decade, won’t take over next decade, maybe the following decade it may be a threat.
When you are the proprietary owner of the worldwide technological system of the largest technology paradigm in the history of the world (as it will be larger than just the CPU and the laptop) your long-term investment risk is such that you can ignore volatility and panic.
The issue becomes, not that Nvidia will be a superior investment over the next 6 years, but whether or not there will be even better investments. And yes, yes there will be, of course.
But the challenge is, as it always is, finding these investment systematically instead of just lucking in on and them. And this we have done on Saul’s board and NPI board quite well.
We don’t buy venture capital, we don’t buy vision, we buy into actually real world success stories, with market leading or market innovating companies, and then we make adjustments as necessary.
Nvidia is one that an adjustment will not be made. It simply is a long term winner.
Have I been emphatic enough ;). Oh don’t shoot me.
Tinker