After Liang Mengsong joined SMIC, did Samsung's chip technology begin to emerge again?
Currently, TSMC and Samsung are the only two companies in the world that produce 3nm processes.
However, while both have hit the nail on the nail at 3nm, in terms of actual performance, one is in the sky and the other is on the ground. TSMC's 3nm contributed 6% of revenue in Q3 2023 and 15% in Q4 2023 and has become one of TSMC's key revenues**.
What about Samsung? It doesn't seem to have customers yet, not even Samsung's Exynos chip, which has opted for the 4nm process instead of the 3nm process.
The industry believes that Samsung's 3nm process has a ridiculously low yield rate, so no one dares to use it, and no one wants to use it. In other words, Samsung, which has the reputation of mass production of 3nm process, is actually very reckless in technology and cannot be used at all...
Thinking of the 14nm FinFET transistor technology launched by Samsung in 2015 and the 16nm technology launched by TSMC, Samsung's 14nm had an advantage over TSMC's 16nm at that time, and we can't help but sigh.
So Samsung and TSMC have an equal share of Apple's orders, and the A9 processors used at that time used Samsung's 14nm process and TSMC's 16nm process respectively, which shows how strong Samsung was at that time.
In fact, Samsung's 14nm FinFET technology advances thanks to one person, and that is Liang Mengsong. He was one of TSMC's top technicians, but later became dissatisfied with TSMC and chose to leave and join Samsung.
Before 14nm, the industry used MOSFET transistor technology, while TSMC and others have reached 20nm, while Samsung is still at 28nm.
At the time, Liang strongly advocated that Samsung abandon the 20nm process and upgrade directly from the 28nm process to the 14nm process and use the latest Finfet transistor technology.
Samsung trusted Liang Mengsong and took a gamble, going from 28nm directly to 14nm, giving up 20nm, and directly entering the FinFET transistor era, no longer using MOSFETs.
At the end of 2014, Samsung's 14nm finfet process was put into production, and in 2015, it was mass-produced, leading the world, when TSMC, UMC, etc. were shocked, and then TSMC was eager to upgrade before launching 16nm finfet transistor technology, but it has lagged behind Samsung.
Later, Samsung made persistent efforts and launched 10nm and 7nm, and Samsung worked hard to catch up with TSMC. Unexpectedly, in 2017, Liang Mengsong joined SMIC.
Subsequently, SMIC quickly mastered 14nm FinFET technology, while Samsung, after losing the big man Liang Mengsong, has moved to 5nm and even 3nm by 2022, although the process is still moving forward.
But in fact, since the 7nm process, the process has not performed well, 7nm chips generate a lot of heat, Snapdragon 888 is the best example, 5nm has low performance, and has been criticized.
I can also take a look at the picture above, at the 7nm process, the transistor density of TSMC and Samsung is almost the same, one is 097, one is 095, Liang Mengsong was still in Samsung at that time.
But when it comes to 5 nanometers, the gap in transistor density is getting bigger and bigger, and by 3 nanometers, it has been distorted, Samsung's 3 nanometers are just equivalent to TSMC's 5 nanometers, and everyone understands why Samsung's 3 nanometers are not used much.
You can see how important a key technical person is to a company. After Liang Mengsong joined SMIC, SMIC quickly repaired 14nm and checked 7nm. But due to the limitations of the equipment, I don't need to say more.
In fact, we have reason to believe that if SMIC is not restricted, EUV lithography machines will also be bought casually, but now it is not behind Samsung, TSMC, and 3nm can also be realized, what do you think?