Chinese semiconductor industry

Just to continue mah rant for one more post ...

This is a fight.

If this is like warfare, then what did Carl van Clausewitz said?

It was all about space, time, mass.

Mass is your center of gravity. That is where your power comes from. That power in this battle for the Americans come from their US companies and patents. The Chinese center of gravity is their market.

How is America going to destroy China's center of gravity to win the war? How can America destroy China's market for IC products?

I don't know. But I do know how China will try to go about to destroy America's center of gravity in the IC industry. Destroy the enemy's center to win the battle.

Space, from time space and mass, space is the battlefield. Whether on the plains, in the mountains, at sea, or in the jungle, that is space and that is your battlefield.

The battlefield here is the entire worldwide IC supply chain and all its complexity. It is not easy getting a handle on it. No one really completely knows. All we can really do is look at the final product, the chip.

Time is the campaign. That has started. China is way past the point of no return in this fight.



That is one big difference between China and America in this IC fight.

One side is most definitely passed the point of no return.

The other side still dithers on how much more to ban, or even scale back the tariffs.

Which side we think gonna win?

FairAndUnbiased said:

This is funny, as if nobody in China in all of history can ever learn to program EDA software, a code that comes from our alien lord's and masters on high as a gift to us humble humans.

Click to expand...

Challenge is that some quirks and optimization were only known by greybeards possessing very specific skills in signal processing.

caudaceus said:

Challenge is that some quirks and optimization were only known by greybeards possessing very specific skills in signal processing.

Click to expand...

what does signal processing have to do with a simulation software? digital logic circuits are typically not noise limited, 1.8 V switching is small for analog and mixed signal microprocessors (3.3-5V) but 1.8V is still gigantic compared to typical noise levels in electronics.

FairAndUnbiased said:

what does signal processing have to do with a simulation software? digital logic circuits are typically not noise limited, 1.8 V switching is small for analog and mixed signal microprocessors (3.3-5V) but 1.8V is still gigantic compared to typical noise levels in electronics.

Click to expand...

caudaceus said:

https://twitter.com/i/web/status/1455987598258688000

Click to expand...

one thing I noticed about these types is that they can make the most mundane garbage sound like it is alien tech. that is why kickstarter scammers (and Theranos, Enron, etc) are so successful at conning them - they do not have the capability to separate hype from the reality. If EDA software requires insane greybeards to develop then the average seniority level at EDA companies wouldn't be, as they said on your linked Twitter thread... recent MSEE graduates.

FairAndUnbiased said:

one thing I noticed about these types is that they can make the most mundane garbage sound like it is alien tech. that is why kickstarter scammers (and Theranos, Enron, etc) are so successful at conning them - they do not have the capability to separate hype from the reality. If EDA software requires insane greybeards to develop then the average seniority level at EDA companies wouldn't be, as they said on your linked Twitter thread... recent MSEE graduates.

Click to expand...

So ultimately you believe that despite lack of "senior talents" local EDA can achieve technical frontiers? That's good to hear.

FairAndUnbiased said:

one thing I noticed about these types is that they can make the most mundane garbage sound like it is alien tech. that is why kickstarter scammers (and Theranos, Enron, etc) are so successful at conning them - they do not have the capability to separate hype from the reality. If EDA software requires insane greybeards to develop then the average seniority level at EDA companies wouldn't be, as they said on your linked Twitter thread... recent MSEE graduates.

Click to expand...

Usually what they mean with greybeards are people that know where in the huge existing codebase they can adjust things and how those changes propagate through the already huge codebase.

If Chinese EDA companies start writing from scratch they actually can implement things faster. Software design hasn't stand still the last couple of decades but the chance of management approval a complete rewrite of a huge existing codebase and primary source of income product is pretty much nihil. Don't think those so called greybeard have some kind of locked away math of physics.

caudaceus said:

So ultimately you believe that despite lack of "senior talents" local EDA can achieve technical frontiers? That's good to hear.

Click to expand...

yes. software is software. As long as you have a foundation to build on and people capable of learning, you can move forward. Let's compare learning software to learning... petrochemical engineering.

To start learning how to program EDA software, you need a computer, valued ~500 USD. So if the will to develop is there, and there's an existing company or university group that makes EDA software (no matter how primitive) then you can develop it. You can read about it in a book, then start trying for yourself. That is why there are EDA startups possible.

In contrast, here's what you need to just get started learning about a real oil refinery... an oil refinery (valued ~1 billion USD). Reading about a distillation column is one thing, but the books don't tell you what an actual distillation column looks like and how gigantic it is. Oil refineries are incredibly profitable. But how come nobody is building an oil refinery startup? Because 1-10 billion USD just to get started building is a gigantic barrier to entry.

FairAndUnbiased said:

yes. software is software. As long as you have a foundation to build on and people capable of learning, you can move forward. Let's compare learning software to learning... petrochemical engineering.

To start learning how to program EDA software, you need a computer, valued ~500 USD. So if the will to develop is there, and there's an existing company or university group that makes EDA software (no matter how primitive) then you can develop it. You can read about it in a book, then start trying for yourself. That is why there are EDA startups possible.

In contrast, here's what you need to just get started learning about a real oil refinery... an oil refinery (valued ~1 billion USD). Reading about a distillation column is one thing, but the books don't tell you what an actual distillation column looks like and how gigantic it is. Oil refineries are incredibly profitable. But how come nobody is building an oil refinery startup? Because 1-10 billion USD just to get started building is a gigantic barrier to entry.

Click to expand...

OTOH digital EDA is a duopoly between Cadence and Synopsys so there's that at least what Quora told me..
Granted that their market valuation isn't really big like recent unicorns and latest food delivery app.

caudaceus said:

OTOH digital EDA is a duopoly between Cadence and Synopsys so there's that at least what Quora told me..
Granted that their market valuation isn't really big like recent unicorns and latest food delivery app.

Click to expand...

In the semiconductor industry, America dominates semiconductor design just like TSMC and Samsung dominate semiconductor manufacture. But I do hear this sector is being capable of being disrupted by Chinese companies.