Consolidation of future application
Recent consolidation progress is going so crazy, mostly because of the IC industry is becoming less and less profitable in every individual application fields. My perspective is that this is not just a consolidation of business itself, but a huge change of IC industry business model. There won’t be so many application fields that needs different IC chips, no matter it’s analog or digital. In analog world, more and more functionality will be replaced with digital as long as it can. It’s inevitable just like in digital world, more and more functionality will be replaced by software, as long as it can.
If going down this way, let’s give it a wild guess. All the applications will consolidate to 2 major fields, those are CPU and GPU. The analog parts, such as power, DAC/ADC, sensors and even MEMS parts will be integrated to CPU or GPU as need. I think there would still be GPU because some cases, more and more, GPU shows performance advantages over CPU while keeping the flexibility which is still in the software domain. And most importantly they represents 2 different programming think flow: serial vs. parallel. Maybe in future, like AMD gambled but failed currently, CPU and GPU will merge into 1 single chip, but they will be technically 2 separate parts.
Many may argue that ASIC chips or even FPGA will show significant performance boost comparing against xPU in some particular cases, or consume much lower power. But my reason is what if the xPU is 100x faster, then why should I care if I lose 10x in speed anymore? And what if the xPU is 100x lower power, then why should I care if I lose 10x in power?
Definitely, there will still be some companies provide ASIC chips in few cases, as rare as panda bear, such as finance and hedge fund, energy or something that are so rich that they don’t give a damn about price/performance trade-off and only want to push the performance to the limit so they can make more money than their compatitors.
Also there will still be different types of CPU or GPU that targeting different optimization directions: high performance and low power.
I might be just too pessimistic, however I judge from the following reasons.
- IC design is still in a very slow moving pace which uses design methodologies from the last century.
- Less young guys are choosing this professional.
- Software is so good that it has all the advantages to replace hardware here and there, and eventually everything.