Look at when China-Taiwan relations deteriorated. Relations were improving at record pace under the KMT in the 2008-2016 period. The CPC had basically recognized the de facto independence of Taiwan, even having Xi Jinping meet the KMT Chairman in 2015.
That rapidly deteriorated with the election of the DPP in 2016, who immediately took an extremely hawkish view on China, invited the USN to cross through the Taiwan Strait in a FONOPS (which, prior to this, had been established as territorial waters under the status quo), started paying US politicians hundreds of thousands of dollars to speak in Taiwan, coincidentally started seeing more aid from the US, forced TSMC to expand their US footprint (with many complaints from TSMC), and followed the US in placing export controls to China. Years of progress under the KMT unraveled.
You have to understand the context here: under the KMT, it's agreed that the two parties will disagree on who "rules" the territories of China. It's also implicitly established that neither party will seek to relinquish their claim on the other's territories. For both the CPC and the KMT, this is a matter of ideology and policy. Knowing that this ideological block isn't going anywhere, CPC-KMT discussions led to the conclusion that, fine, we won't agree, but we also won't do anything about it. Neither military intruded over the status quo median line, neither military provoked the other with missiles or fighters or whatnot, and it was established that the issue was one of minimal importance compared to economic development and peaceful codevelopment. China knows that taking Taiwan is basically impossible, and Taiwan has no aspirations to retake China.
In comes the DPP, arms swinging, with support from the US, and says that the KMT is clearly siding with the CPC on this issue and is clearly going to seek reunification with the CPC. Reunification is against KMT policy for obvious ideological reasons, but alas. So, the DPP comes in, saying they want de jure independence and to align with the US, fuck China, Taiwanese people aren't Chinese, etc. etc. Obviously, China isn't too happy about this, but things proceed as usual.
Taiwan then declares that the Taiwan Strait is international waters (since, per DPP policy, Taiwan is not China and thus the Taiwan Strait doesn't classify as territorial waters), that they want more weapons from the US, and that they don't want to trade with China. China is unhappy about this, but it exposes a key vulnerability in the concept of international waters: there's nothing stopping China from flying in international waters. So, with the justification of the US FONOPS (i.e. sailing an armed US warship) through the Strait, China starts flying sorties past the median line (which, as established, is now international airspace). China also starts shooting missiles from international airspace crossing international airspace into international airspace, using US FONOPS as justification for this being perfectly reasonable. That's how we ended up here. I'm strongly opposed to the DPP, not strongly opposed to Taiwan. I see the DPP as being intentionally provocative and throwing away a massive economic boon (trade with China) in exchange for the DPP's own ideological goals. It's coming at the cost of opportunities in Taiwan, it's destabilizing the region, and it's pushing Taiwan into the same unstable flip-flopping political situation as the US.
The US is constrained by capitalism and globalization... And years of mismanagement at Intel under Krzanich... and the lack of profitability of Global Foundries.
Intel only recently adopted a foundry model: previously, their fabs were only used to manufacture Intel chips... I'm sure you can imagine some of the issues there, but it helps that Intel is a massive company. Intel really bit off more than they could chew with 10nm and started to lag behind.
GloFo used to be AMD (until it was spun off for profit because AMD needed money... GloFo gave up on 7nm because it was seen as too expensive.
As for Samsung? Nobody really knows why Samsung's technology sucks, but it sucks. Something wrong with their FinFET process in general I guess.
TSMC isn't a decade ahead. They're maybe 5 years ahead of SMIC and maybe 2 years ahead of Intel/Samsung. They're only so far ahead of SMIC because SMIC isn't allowed to import EUV machines from ASML since the US decided that China was getting too close to toppling American dominance in semiconductors and AI.
The main thing limiting SMIC is the lack of EUV machines, but Huawei is expected to pop one out soon based on the rumours being spread on Chinese forums. That's the story. TSMC doesn't have some magic sauce, they have scale, billions of dollars in government support, and a slight technological edge. If anything, TSMC's magic sauce is that the most desirable job in STEM in Taiwan is to become an engineer at TSMC: they attract top talent in a way that Intel doesn't.