Even American companies like Intel are growing more capa/output outside USA than inside for the last 5 years : Israel, Ireland, China, etc
Meanwhile, the below data shows what is going on in Asia.
I'd love to see a domestic semiconductor manufacturing boom, but the data shows it's just not happening anytime soon or else there would be a lot of new fabs and a lot of new tools being invested in America.
10X? BS! America semi is growing faster than any in Asia in the last 2 quarters in 2017. Both Intel and MU are expanding semi capacity in 3D NAND in US, Singapore and China, while other companies like ADI, TI, Global Foundry ( US, NY state ) are expanding.
America has a zero % chance of regaining any kind of manufacturing "lead" in semiconductors of any variety in the foreseeable future. Even if Intel expands in Arizona, or Samsung in Texas, or if Taiwan runs out of water and TSMC also sets up shop in USA...there are still 10x more capacity coming online in Asia.
Inside transfer may not be the only factor Samsung rises to the top. For example, Hynix did not have any inside sale transfer, it is still the second largest DRAM. In the case of Samsung, it was its Wireless devices sale to generate large profit ( larger than Intel's profit ) a few years ago, then they were re-investing this profit into their factory, to generate more profit; besides, their government was trying to help their export in very possible way; meanwhile, in the US, the government ( federal and local state ) in a "hand-free" enterprise mode, while demanding higher wages at all levels, as a result, America is losing the competitiveness in all geographical areas, especially comparing to the developing world like Korea and China. In the memory world, it is appearing to turn the tide recently, America's innovative idea to generate newer products, like X-point memory and others, it will take back the crown from Korea. I am sure America will be in the #1 spot again so long as we are providing a fair and open competitive environment in the market place, with a little government support, like research and manufacture support, America will be great again.
Another consideration in the Samsung and Intel battle for the future number one semiconductor sales ranking is Samsung's internal transfers. With about 25-30% of Samsung's 2016 semiconductor sales being internal transfers to its own electronic systems businesses, a large part of it future semiconductor sales success is dependent upon a continuation of strong system sales by the company.
One of the reasons for the Japanese companies' lackluster semiconductor sales since the 1980s and 1990s was the weakening of their electronic system sales. When the Japanese companies were dominating the top 10 semiconductor ranking so many years ago, internal transfers also averaged about 25-30% at companies like NEC, Hitachi, Toshiba, etc. As thier electronic system sales weakened, so did their internal market demand for semiconductors.
Besides the cyclical nature of the memory market, if Samsung has a serious stumble in its smartphone business, it may once again find itself as the number 2 supplier.
Samsung is ahead last quarter; it is not clear how much longer Samsung can hang on to the #1 position. Intel has said their revenue in the next quarter will be $15.7B, it will be likely higher than that since it was a midpoint estimation. Meanwhile, memory and display revenue has stop rising for Samsung, while Intel's 3D NAND and X-Point are coming on line, will increase Intel's revenue, while decrease Samsung's due to Intel's strong competition product both in NVMe, and X-point ( in memory), it will eat into Samsung's DRAM too ( with X-point).
It is not just a regular memory circle. Supply-Demand imbalance in the memory market can NOT be easily solved at this time because it is a totally new phenomenon due to memory device scaling limitation.
DRAM bit growth is the lowest for the last 20 years. And, DRAM vendors are not making big investment in DRAM fabs anymore. So, over-heated DRAM price will last long.
There is a fierce competition on 3D NAND. Though, 3D NAND is unable to be cost competitive so far due to poor process architecture and manufacturing difficulities. So, it will be difficult to expect abrupt price drop of NAND and, therefore, NAND supply will be also difficult to meet demands.
From now on, customers should pay more than what they wish for. This is the reality of end of Moore's Law.