Intel Corp (INTC)

[This Causes an Error]

. Maybe TSMC isn't lying this time. But everything they are doing smells funny.

Generating billions in revenue and net income and winning the lion's share of the 14/16nm foundry market in 2016 doesn't smell funny to me. Smells like good execution by a good company.

16nm is NOT equivalent to Intel 14nm.

No, it's not. I have not disputed this.

Their process is markedly inferior to samsungs.

Nonsense.

I would say there is a very good chance that their 10nm is not a true 10nm process.

Their 10nm should be a little denser than Intel 14nm, but that's not really the point. Point is, we very well may see "10nm" TSMC product in the market before you see Intel "10nm."

Intel will get to "10nm" before TSMC gets to its "7nm" but I suspect that TSMC's "7nm" should be essentially equivalent in terms of density to Intel 10nm, if not a little denser.

Intel and TSMC should transition to their 7nm and 5nm processes, respectively in ~2020.

Today Intel has the lead, but from what both companies have said publicly about their future plans, TSMC is running much faster than Intel is, which means that the gap that currently exists should close over then next couple of generations.

But I have watched companies like TSMC and TI and IBM spread disinformation for years on semiconductor process technology.

I think you are not thinking about the massive structural change that has happened in the foundry world. Until the mobile computing revolution, Intel had the lion's share of the leading edge business industry-wide. There simply wasn't enough revenue to go around to make it worthwhile for the foundries to invest at the level that Intel was able to thanks to its massive volumes in PCs.

This changed with iPhone/smartphones. The foundries need to meet the very demanding needs of companies like Apple and Qualcomm, who, combined, need >100 million units of leading edge SoCs each year.

The massive revenue growth that TSMC experienced allowed it to significantly rev up its R&D engine. Take a look at this:



The TSMC you "remember" spent well under $1B/year in R&D; the TSMC of today spends >$2B/yr in R&D.