Till the late 90s, Intel was a thriving company. After that, all the smart people left, most of them rich from the phenomenal growth of the stock in the preceding years. Since then, smart people haven't been joining Intel. Stock option grants have become worthless and Intel only pays about market average. Therefore, Intel went from being the employer of choice to a second option. In the past, you had smart people who could cut through the layers of process in corporate communications and get their ideas across. Not so any longer. With the mediocre engineers that started working there, people started hiding behind the process and it became much like a third world bureaucracy -- lots of process and little forward motion. That is still the situation today. It's a chicken and egg problem for Intel. If they can get their stock growing again, better engineers will choose to join them which will rekindle the spirit of innovation that once used to be there. But to get their stock growing again, they need better engineers than the ones they currently have.
I don't have enough insight to agree or disagree with your specific points. But from the outcome I am seeing from Intel, I see some sence in what you say. Intel spends 5-10X times it rivals in R&D but not coming out with substantially competitive poducts to justify that much spending. Apple, Samsung, QCOM are doing uphill tasks better than Intel is doing downhill task.
But I think the main problem plagueing Intel today is not the egineering performance. I think Intel today has already surpassed ARMH in performance/watt through R&D muscle power. The problem with Intel is miserable PR and marketing performance. A right guy at the top will change the things downside up.
It's a systemic problem that you cant just root out which short term fixes. It needs a longer term approach to resurrecting the company which includes paying better for the better engineers, something Intel hasn't culturally been wont to doing. They've traditionally taken the stand that their corporate image is so good that they'll automatically attract the best engineers. Unfortunately, that has changed over time. Their lack of R&D ROI is a symptom of this underlying problem. Also, their lack of being able to deliver on commitment. They make the right speeches, get their ducks lined up the right way, but the forward momentum is just not there.
The higher Intel RD cost is driven by their investment in advanced processes. Today, the high cost is driven by 14nm investments and related to the way Intel aggressively charges off the process development costs.
Apple and Qualcom don't have these process development costs since the costs are shouldered by their foundries. The foundries are not making the large investments in processes that Intel is. Samsung has a large retail product element to its business that makes their RD a smaller overall percentage.
You have described the foundry dilemma. Who pays for the risk of developing costly advanced processes and building expensive fabs. Ultimately foundry customers pay.
Apple and Qualcom have problems building their products and Intel has the fabs and is developing products that will match the market to fill their fabs.
Why not earlier? Intel fabs have been running capacity constrained for many years and they could not build the MP3 player chips for customers for new customers without abandoning current customers. Intel did not know that 22nm yield would be as good as it is and built an extra fab as insurance against low yields. That extra fab is the current 25% overcapcity that Intel is now charging off and converting early to 14nm.
In your very limited and very finite wisdom can you point to a product of Intel's 'poor' engineers ? Frankly all the words above were a complete waste of time reading as they don't reflect anything that resembles reality in the real world.