I can see why they wanted to change the "RISC-V is inevitable" line. After you have heard it too many times it starts to sound like end-of-the-world zealots that expect us not to notice that the world has not ended as they have endlessly predicted.
Perhaps "RISC-V is now" will build some excitement. But there is a danger that it will ring even more hollow if we have to say it for 3 years before it rings true. RISC-V is inevitable and, for many use cases, it is soon. But for most people, it is certainly not now. RISC-V is still too slow and too expensive for most use cases outside of micro-controllers.
In the next 2-3 years, all that will change. RISC-V has a real shot at the embedded mid-range where the ARM product line has stagnated somewhat. RISC-V has caught up in performance, has the features, and is perhaps a bit better in terms of vectors and AI. With luck, we will see things like smart TVs and perhaps even phones. Real success in phones may take a few years of slowly chipping away though.
And on the server side, RISC-V has a chance to break-in. It is not going to put Intel, AMD, or ARM out of business just yet but it can start to build share. Specialized AI workflows offer a real opportunity.
For the SBC market, RISC-V is far too expensive given what it competes with. If RISC-V find volume elsewhere, this could change.
On the desktop/laptop, RISC-V is not even remotely a player. This will be the hardest market to break into. But any success ARM has makes it easier for RISC-V to follow. And once they are both there, what is the advantage of ARM again? Certainly not legacy compatibility.
RISC-V is inevitable. And its time certainly "starts" now. But I am not sure that RISC-V is now for most of us.
Perhaps "RISC-V is now" will build some excitement. But there is a danger that it will ring even more hollow if we have to say it for 3 years before it rings true. RISC-V is inevitable and, for many use cases, it is soon. But for most people, it is certainly not now. RISC-V is still too slow and too expensive for most use cases outside of micro-controllers.
In the next 2-3 years, all that will change. RISC-V has a real shot at the embedded mid-range where the ARM product line has stagnated somewhat. RISC-V has caught up in performance, has the features, and is perhaps a bit better in terms of vectors and AI. With luck, we will see things like smart TVs and perhaps even phones. Real success in phones may take a few years of slowly chipping away though.
And on the server side, RISC-V has a chance to break-in. It is not going to put Intel, AMD, or ARM out of business just yet but it can start to build share. Specialized AI workflows offer a real opportunity.
For the SBC market, RISC-V is far too expensive given what it competes with. If RISC-V find volume elsewhere, this could change.
On the desktop/laptop, RISC-V is not even remotely a player. This will be the hardest market to break into. But any success ARM has makes it easier for RISC-V to follow. And once they are both there, what is the advantage of ARM again? Certainly not legacy compatibility.
RISC-V is inevitable. And its time certainly "starts" now. But I am not sure that RISC-V is now for most of us.