TSMC 3 nanometers are in short supply! Has achieved absolute victory over Samsung and Intel (Korean Media)

[BruceA]

True and TSMC has no competition at N3 so the margins will be higher than previous nodes. I have never heard big TSMC customers openly support TSMC price hikes before. Exciting times.

I would argue that margins on the leading node are tough even with price hike. You build a new building outfit it with new tools ( expensive ). The technology has more steps as well so cost goes up a lot too. Then you got depreciation plus tie consumables. As noted by another post, it’s the legacy nodes like N7 that are / should be really making money as the tools depreciated and the efficiency and cost reductions of 5-7 years high volume learning are applied and you of course give some of that to your customers as wafer discount from five years ago. You lure customers from older nodes with PPAC competitiveness to peers and you have TSMCs great business model.

The challenge is China adding a lot of legacy, and Samsung and Intel hoping to take some leading and legacy. Also Intel which total volume is huge hoping to bring it all in house and add some additional customer makes for some competition if they can fix their culture and execute.

Even with all that, by any metric TSM is still cheap at current valuations. Home run you chase AI, want steady singles TSM

 

[KevinK]

I believe Intel's business model allows those people to thrive at Intel and cause long term damages without even knowing it.

It‘s just classic Innovators Dilemma. The Intel near-monopolistic business model / business hid horrible mistakes, plus deterred new disruptive business opportunities that eventually boxed Intel in (Smartphones, GPU/AI). But you can also credit them with one success - They successfully took down Sun and IBM’s server /mainframe business, because those guys also ran into the same Innovators Dilemma problem. And IBM’s one successful attempt to go downmarket, the PC, eventually ended up consuming them as Wintel eventually sucked all of the value out of the PC market, and IBM could never establish enough added value within their ecosystem. Then, a few years later, Intel, with a little help from AMD and Linux, came gunning for IBM, with a far lower cost basis.

 

[MKWVentures]

@Daniel Nenni . What percentage of 3nm wafer is going to Mobile vs PC vs Data centers today. Mobile and PC are below there historic TAM peak. Only Data center is really growing.... correct? and the datacenter growth is GPU/Accelerators right?

 

[hist78]

It‘s just classic Innovators Dilemma. The Intel near-monopolistic business model / business hid horrible mistakes, plus deterred new disruptive business opportunities that eventually boxed Intel in (Smartphones, GPU/AI). But you can also credit them with one success - They successfully took down Sun and IBM’s server /mainframe business, because those guys also ran into the same Innovators Dilemma problem. And IBM’s one successful attempt to go downmarket, the PC, eventually ended up consuming them as Wintel eventually sucked all of the value out of the PC market, and IBM could never establish enough added value within their ecosystem. Then, a few years later, Intel, with a little help from AMD and Linux, came gunning for IBM, with a far lower cost basis.

Intel did try to use the merge and acquisition to bring new idea, products, and talents into Intel. Unfortunately, for variety of reasons it didn't bring sufficient impacts to Intel or can't find a good position in the Intel organization chart, such as:

Altera: $16.7 billion, 2015, spinoff from Intel soon.

Mobileye, $15.3 billion, 2017, spinoff from Intel in 2022.

McAfee, $7.68 billion, 2010, spinoff from Intel in 2017.

Habana Labs, $2 billion, 2019, Intel's hope to compete against Nvidia's AI products. Not showing good result "yet".

Infineon Wireless Solutions, $1.4 billion, 2010, sold to Apple in 2019.

 

[hist78]

Don’t see how TSM doesn’t eventually become the most valuable company in the world at some point. I am more sure about this than I was about nvidia in 2016.

At this moment ( June 17, 2024, 1:15PM EDT), Broadcom is the #10 most valuable public companies in the world while TSMC is at #8 position. #1 Apple, #2 Microsoft, #3 Nvidia, #4 Alphabet (Google), #5 Amazon, #7 Meta Platforms (Facebook), and #10 Broadcom all work directly with TSMC on their important products and services.

 

[Mooredaddy]

At this moment ( June 17, 2024, 1:15PM EDT), Broadcom is the #10 most valuable public companies in the world while TSMC is at #8 position. #1 Apple, #2 Microsoft, #3 Nvidia, #4 Alphabet (Google), #5 Amazon, #7 Meta Platforms (Facebook), and #10 Broadcom all work directly with TSMC on their important products and services.


View attachment 2022

7 of the top ten are TSMC customers indeed. With the others being Berkshire, Aramco, and TSMC itself . All roads lead through CC Wei's office

 

[Daniel Nenni]

@Daniel Nenni . What percentage of 3nm wafer is going to Mobile vs PC vs Data centers today. Mobile and PC are below there historic TAM peak. Only Data center is really growing.... correct? and the datacenter growth is GPU/Accelerators right?

Apple is the biggest N3 customer in HVM thus far so mobile by a large margin. Apple will always be TSMC's biggest customer.

 

[hist78]

7 of the top ten are TSMC customers indeed. With the others being Berkshire, Aramco, and TSMC itself . All roads lead through CC Wei's office

Three of the top ten most valuable public companies (Nvidia, TSMC, Broadcom) in the world are semiconductor companies. I didn't expect it just 2~3 years ago.

 

[MKWVentures]

The most successful semi companies in the world work with TSMC. They have enabled this whole revolution with LEADING EDGE semi foundry.