For the purposes of this conversation, I'll define "nowhere close as follows (and willing to put money on the prediction, if you'd like to negotiate such):
"Primary" means simple majority. >50% ARM market share for consumer PCs as defined below.
"Consumer CPU" refers to non-server application PCs. Workstations, desktops, laptops, and DIY. Personal and professional use (including enterprise purchases, so the full fleet of corporate laptops at any company are included). Some current examples would include things such as the Talos POWER workstations and ARM workstations.
For "PC", the definition gets a little rough, and I'd be willing to negotiate this. I'd say anything where the primary interaction mode is mouse and keyboard. With this definition I would exclude most tablets, but something such as the Microsoft Surface would be included. The test that includes the Surface and excludes most other tablets is that the Surface is sold with a keyboard and trackpad. If the keyboard and trackpad is not included or required, then it's not a PC. This rule would also exclude something like a NAS, where the hardware is basically that of a tower desktop, but the use is headless; this would be a server. Like I said, this can get a bit loose. Happy to negotiate a firmer definition.
As for "nowhere close", I'd say that we won't see ARM as the primary consumer CPU for the next decade easily. I'd be willing to bet $1,000 US on this.
If you wanted different timelines, I'd put $5,000 on 5 years, and $500 on 15 years.
Apple's market share won't push ARM past 50%. Enterprises move slowly and make up a lot of the PC market. If you assume a 3-year hardware replacement cycle, then ARM would have to come to dominance at least 3 years before the end of a betting period for it to have a chance of meeting the end of the betting period. Thus, for the 5 year timeline, ARM would have to be dominant by 2022. For the 10-year, by 2027.
Separately, in writing this response, I realize that you could easily make the argument that a "consumer CPU" includes cell phones and tablets. This is a very different conversation.
> "Consumer CPU" refers to non-server application PCs. Workstations, desktops, laptops, and DIY. Personal and professional use (including enterprise purchases, so the full fleet of corporate laptops at any company are included).
This is a totally arbitrary distinction. We can’t just ignore the fact that consumers are doing the bulk of their computing on smartphones nowadays.
> We can’t just ignore the fact that consumers are doing the bulk of their computing on smartphones nowadays.
> > Separately, in writing this response, I realize that you could easily make the argument that a "consumer CPU" includes cell phones and tablets. This is a very different conversation.
I think you can see that I didn't ignore this fact in my post.
I made a definition (as requested) to clarify my position and based on my interpretation of the original post in this thread. Based on this interpretation, it seemed like that poster was referring broadly to PCs. The definition was for the purposes of a prediction.
In defense of my interpretation, I would consider it unreasonable to assume that the OP meant to include smart phones in their consideration of "primary consumer CPU", due precisely to the fact that you and I have both mentioned, that many individuals use a smart phone for the bulk or even all of their computing.
"Primary" means simple majority. >50% ARM market share for consumer PCs as defined below.
"Consumer CPU" refers to non-server application PCs. Workstations, desktops, laptops, and DIY. Personal and professional use (including enterprise purchases, so the full fleet of corporate laptops at any company are included). Some current examples would include things such as the Talos POWER workstations and ARM workstations.
For "PC", the definition gets a little rough, and I'd be willing to negotiate this. I'd say anything where the primary interaction mode is mouse and keyboard. With this definition I would exclude most tablets, but something such as the Microsoft Surface would be included. The test that includes the Surface and excludes most other tablets is that the Surface is sold with a keyboard and trackpad. If the keyboard and trackpad is not included or required, then it's not a PC. This rule would also exclude something like a NAS, where the hardware is basically that of a tower desktop, but the use is headless; this would be a server. Like I said, this can get a bit loose. Happy to negotiate a firmer definition.
As for "nowhere close", I'd say that we won't see ARM as the primary consumer CPU for the next decade easily. I'd be willing to bet $1,000 US on this.
If you wanted different timelines, I'd put $5,000 on 5 years, and $500 on 15 years.
Apple's market share won't push ARM past 50%. Enterprises move slowly and make up a lot of the PC market. If you assume a 3-year hardware replacement cycle, then ARM would have to come to dominance at least 3 years before the end of a betting period for it to have a chance of meeting the end of the betting period. Thus, for the 5 year timeline, ARM would have to be dominant by 2022. For the 10-year, by 2027.
Separately, in writing this response, I realize that you could easily make the argument that a "consumer CPU" includes cell phones and tablets. This is a very different conversation.