Replies to post #143779 on Intel Corp (INTC)

[chipguy]

sadly I am also hearing that 10nm is just a shrink of 14nm :(

At these feature sizes nothing is "just a shrink".

When a critical layer starts 100 atoms thick a 30% shrink means one thing.

When a critical layer starts 5 atoms thick a 30% shrink means something
else altogether.

[Unkwn]

Capex comes out of cash flow; free cash flow is cash flow less capex.

I'd rather say Capex is financed by operating cash flow (or financing cash flow in case of raising debt or equity).

Um, ok, so first of all the majority of Samsung's Capex spending is for memory capacity (mainly DRAM but also plenty of NAND flash), so this isn't at all apples to apples.

Do you have any proof for that? As far as I know, Samsung doesn't disclose this. In addition, there are still some synergies between memory and logic semiconductor production. I know it is not the same but it isn't fully orthogonal either.

Third, I think you may be confused as to capex v.s. R&D spending for process development. The vast majority of capex is to put capacity into place, has nothing to do with a company's ability to R&D a new manufacturing node.

I said this a few times already: They can capitalize a big part of their R&D spending for process technology and it is paid through depreciation. It makes sense in this case, since the expenses have more a character of an investment than an expense. R&D for TSMC has been around 1.5 billion in 2014. That basically makes things worse, since that fully adds to Capex for process spending.

Secondly, TSMC's cash from operations ("cash flow") over the last twelve months was $17B, not $12B. Capex in the last year was ~$8.1B, so TSMC is very nicely free cash flow positive.

Ok, I checked that. Last fiscal year it has been at around 14 billion. My twelve billion were wrong because the Taiwan Dollar depreciated since then. If you have a look at TSMC's quarterly operating cash flows you can see that it actually was 20% lower than 9 months before - not a nice trend that I can see here.

Anyway, it's just not a credible argument to say that TSMC can't afford to spend to develop next gen nodes -- it can more than afford to pay for them.

I didn't say they can't afford it. They can't afford to outspend Samsung and Intel. That's a fact you seem to ignore.

The key will be finding customers to pay for it. Keep an eye on the health of the high end smartphone market for an indication of what the future holds :P

If you're asking me, the future of the high end smartphone market will be lower margins. We'll see.