Thank you for your service comrade, the CCP has deposited ten cents into your account.

The natural instinct of all parties involved has been to move toward fab independence from TSMC. The idea of sanctions on US technology, basically eliminates US technology over time. The US is the only country ever to deploy such a strategy in an attempt to protect the US semiconductor business. Companies like Qualcomm oppose this ludicrous US policy and so does the Dutch firm ASML. What is going to happen is that China will develop self-sufficiency and revoke US products and they won’t need ASML technology.

ASML technology is the lynchpin of the fab world. Their EUV systems are necessary to make the narrow gated fab, along with a host of other substrate washes and other products made in various countries which have stupidly decided to join the US sanctions. IT WON’T WORK!! It never has worked. What will happen is that US products and sanctioned products will die on the shelf eventually. The only gain will be short term.

While narrow gate is the hot topic, 14 to 28 nm is the money chip.

Let’s remember that Intel missed the handheld market entirely for the simple reason they have been the most sanctioned company in the world by the US Gov. The US government blocked Intel chip sales to China for years and China went on to develop its own core capacity. Intel was virtually destroyed by politicians. The idea that INTEL will suddenly be relevant in the global chip market is a pipe dream. Asia is removing US made components, not embracing them. Apple makes there own processors to avoid US sanctions in Asia. It seems so simple that disrupting the division of labor has only caused chip shortages and increased costs. Further, China is moving along the supply chain and developing capacity and their own substrates. This scares the US suppliers because they know this is not rocket science and China has the capability to become self-sufficient.

For example if you raised cattle and bought your fence from the USA but the US wanted to harm your cattle industry so they sanctioned wire fence, what would happen? The rancher would buy fence elsewhere or make their own. Eventually China will go after the global fence market with highly competitive products. The US would have to apply more protectionism to their fence and the US beef producers would have to raise prices and could not export their products. Protectionism always fails. Always.

Another issue with fab. IBM is using Samsung FAB. What prevents Samsung from letting IBM develop their servers and then developing their own. Intel has the same problem. Is Qualcomm going to use an Intel Fab plant or Samsung? No. TSMC is successful because it only does FAB. However US sanctions will ultimately destroy TSMC because of Trump stirring global paranoia. Chips will also become more expensive and no Asian vendor is going to want to be negotiating with US politicians to use US chips in their consumer products.

Even INTEL is embarking on RISC-V, to avoid US sanctions. Can you imagine? I like the idea but it is a little late to the dance since China has a number of RISC-V chipmakers.

But the cat is out of the bag on all this now. Rivers of paranoia are flowing and that means continued chip shortages and a shakeout of chipmakers. Look at Japan and Korea. Since the US push is on for fab independence, Japan and Korea feel left out of the equation. They will more closer to greater Asia and China and away from US dependence.

At present ASLM and TSCM are holding the chipmaking world hostage. So in the long run it is in the best interest to develop competing technologies and fab works to break this dependence and the US hegemony. That is exactly what is going to happen over the next several years. China will throw the most money at this and has the big sway as the largest buyer of chips. In the long run, Intel can make the smallest baddest chip in the world, but if they can’t sell it in Asia, they have NO MARKET and their chip has no purpose. TSMC will also avoid making chips from any company that has a proposed fab operation. Business is robust and they don’t need to cultivate any competitors.

Astonishing how US politicians can so wreck a smooth flowing semiconductor industry that has thrived on division of labor and cooperation. The US is going to lose 20% of the global semiconductor market over the next three years.

Intel CapEx;

2015 = $7.3 b
2016 = $9.6 b
2017 = $11.8
2018 = $15.2
2019 = $16.2
2020 = $17.0
2021 = $25 ish

Shows $76.3 billion on the books 2015 through 2020 that’s around one year of Intel revenue. On a depreciated basis every year credited to plant and equipment use and/or maintenance (assuming physical site ‘construction’ is sufficient size depreciated at 30 years under Rocks Axoim subsequently requires debiting Property, Construction (Plant and Equipment) x2 every year sustaining a production worthy operation under Rock the accrual ratio is not one to one.

Subsequently Intel in the next I will use 6 years here instead of 5 of 7 years needs to earn minimally $152 billion before R&D, operating cost and profit. TSMC is in the same situation.

The realization is that Intel and TSMC need to double revenue and the question is over what period of time on the ability to stretch the capabilities of lithography on the existing investment in plant and equipment. Certainly within 5 to 6 years Intel and TSMC must double revenue then the question becomes is that objective sooner?

Here’s the a’ha on this conundrum of economics immensely simplified for Intel and TSMC. Imagine an equally distributed ‘standard’ cost curve over the prior 30 years of accelerated lithography advancement. Ignore for the moment High K and strained silicon resets in that cost curve. Think ‘as standard’ and continuous and the the bottom of the downward slopping cost curve is reached at 7 nm and now industry; Intel and TSMC and others, are on an upward slopping cost curve into process saturation and through a jump gate into whatever comes next. Essentially the prior investment in what was is now the total cost of the remaining investment in what can be.

Time compresses as the physics or process saturates at infinity.

Mike Bruzzone, Camp Marketing