Hacker News Reader

Comment by ipnon | Hacker News Reader

TSMC is in a league of its own. Only Samsung currently has the capacity to compete on a similarly international level. There are a variety of startup competitors from mostly China and the US, but TSMC is so technologically advanced, and the demand for ever faster chips is so great, that they will easily retain the dominant global position for at least the duration of the 2020's.

View on HN

It might seem like Samsung is caught off balance by TSMC's recent progress, but they are still very much in this race.

I haven't worked for Samsung in ~7 years, but last I recall they were aggressively pursing all manner of value optimization angles. They also have an existing presence in the US, whereas TSMC is going to be about 5 years out on their US factory. Keeping American workers around to inject ideas into your manufacturing process can be invaluable if you have typically been operating the business like a Borg cube (i.e. relentless standardization throughout, potentially at the cost of more complex value-add proposals).

Fabs seem like a classic example where the market will trend to a monopoly. Each generation makes fabs themselves more expensive along with the incremental R&D . As each generation largely obsoletes the previous, the price of being second goes up.

Drexlarian nanotechnology?

Also a perfect case study of where the free market worldview and the nation-state worldview bump heads. The free market tends towards a single global monopoly for these highly capital intensive industries in the name of economic efficiency, but this inevitably erodes the sovereignty and independent action of countries that have to depend of the good graces of wherever this monopoly has its HQ.

Until a startup with an innovation in fabrication itself comes along disrupting the industry and rendering all those incumbent investments suddenly obsolete.

Happened again and again. In free markets, of course. Even if they may trend to, only controlled markets reach monopoly.

Maybe after twenty to fourty years this will happen. So far, this hasn't happened in this field.

There are plenty of counter examples, and I'd dare say the majority of notable historical "true" monopolies were disrupted due to government intervention rather than disruption. While intel has had a process lead for decades, they've never had a monopoly on fabrication.

Do you have some examples? Any monopoly I know of was government granted, usually through regulations. I actually know of much more examples of disruption...

Now the market can arrive to a state of monopoly due to some powerful player, but as long as the startup mechanism is functioning, competition will arrive and disrupt.

Standard Oil, US Steel, AT&T are the best known examples in the US. None of these (including AT&T) were government granted. They also share a few key traits

1) They dominated their "primary" markets

2) Their primary market was deep, and competing with them was cost prohibitive for any economic entity as in lose the GDP of a small country expensive.

Ok, I can’t talk about the first two because, frankly, it’s been a while.

But AT&T - you kidding me?! It was illegal to even bring your own phone device to their network! It was a classical example of government protected monopoly.

A free market will break its own monopolies. A market corrupted by government regulation will need government intervention.

https://en.wikipedia.org/wiki/Kingsbury_Commitment has a useful summary of the 5 attempts made by the government to curb AT&T's market power.

With regards to it being illegal to connect to the phone system with your own device... this was a notion pushed by AT&T's legal, and operations teams. https://en.wikipedia.org/wiki/Carterfone has some useful context on the regulatory decisions which compelled AT&T to allow third party devices on "their" network.

Ultimately large monopolies contort the institutions, governments, and legal systems they are surrounded by. You wouldn't believe that a Trillion dollar corporation would argue for legislation that wasn't in their favor.

Yeah, but it is the legislation that is preventing the competition and thus making them a monopoly, otherwise they would be disrupted by nimbler, faster startups of the day for which a monopolized market is an irresistible target (little competition and large profits).

This doesn't really work for natural monopolies, a small phone operator doesn't connect you to many people, a small steel mill doesn't have the economy of scale as a large mill, small oil operators couldn't build their own oil distribution.

I don't think those are actually natural monopolies. But there are ways to compete when you are small and you can't compete on the usuals (price, distribution, etc). It takes creativity though.

I remember when the local telecom monopoly in my town held ski-high Internet prices, regular people started laying up Ethernet cables in the air between high-rises. A new ISP was born that way, who now has completely wiped out the original telecom.

Of course though, there is now a regulation against doing that kind of thing...

It seems that any activity where R&D is a significant chunk trends towards a monopoly.

Reply on news.ycombinator.com