- The amazing GPU's they sell are used to run software to learn things and solve problems
- The data centers that are being build will keep running for years using better and better software to learn more and more things about the world, technology, etc.
1. When the purchases have been made and the data centers are running, who will NVIDIA sell to next?
2. When the first mover advantage has passed and many companies offer chips with similar capabilities for lower prices or a new technology like RISE 2.0 or whatever unknown approach will come out, what will happen to NVIDIA?
I think those are very reasonable concerns. For now it seems like the demand significantly outstrips the supply so they will have massive sales for years to come, but what shape the future takes is still up in the air. Nvidia's advantage is less on the chip side and more in their developer ecosystem (IMO), which is something that competitors will struggle to replicate (something like 4m developers are trained in Cuda and there really isn't anything comparable elsewhere).
Would note that their biggest customers have all developed their own chip technology to run on their own platforms so once that scales up the industry could look very different. Whether that just means they sell few chips to the Amazons/Alphabets/Microsofts of the world, or whether they develop something that can properly compete.
I did a post a few weeks back (The Kings of Cloud: https://www.marketlabnewsletter.com/p/kings-of-cloud) which shows Street estimates for $382 billion in 2028 sales for Google Cloud, Microsoft's Intelligent Cloud and Amazon's AWS. Presumably some of that would come at the expense of Nvidia, since going DIY won't make sense for a lot budding AI companies.
I wonder about two things related to NVIDIA:
- The amazing GPU's they sell are used to run software to learn things and solve problems
- The data centers that are being build will keep running for years using better and better software to learn more and more things about the world, technology, etc.
1. When the purchases have been made and the data centers are running, who will NVIDIA sell to next?
2. When the first mover advantage has passed and many companies offer chips with similar capabilities for lower prices or a new technology like RISE 2.0 or whatever unknown approach will come out, what will happen to NVIDIA?
I think those are very reasonable concerns. For now it seems like the demand significantly outstrips the supply so they will have massive sales for years to come, but what shape the future takes is still up in the air. Nvidia's advantage is less on the chip side and more in their developer ecosystem (IMO), which is something that competitors will struggle to replicate (something like 4m developers are trained in Cuda and there really isn't anything comparable elsewhere).
Would note that their biggest customers have all developed their own chip technology to run on their own platforms so once that scales up the industry could look very different. Whether that just means they sell few chips to the Amazons/Alphabets/Microsofts of the world, or whether they develop something that can properly compete.
I did a post a few weeks back (The Kings of Cloud: https://www.marketlabnewsletter.com/p/kings-of-cloud) which shows Street estimates for $382 billion in 2028 sales for Google Cloud, Microsoft's Intelligent Cloud and Amazon's AWS. Presumably some of that would come at the expense of Nvidia, since going DIY won't make sense for a lot budding AI companies.
Wow, great report! Jam packed with info and hilarious memes.
The Big Short meme with “listening to college grads explain what they’ve been up to in networking events” had me cracking up.
I just wrote a blog post relating to Millennials giving Gen Z advice for the future too.
Good stuff!