Hacker News Reader

Comment by hardware2win | Hacker News Reader

Ive invested in them 1yr ago and as the time is passing Im feeling more and more confident that it was really good bet

Im excited and cant wait to see next 3 years

why would you believe Intel can actually catch up to the far-ahead competition?

TSMC aren't standing still, they obviously aren't stupid and they have shitloads of money. Intel probably scores zero on those 3 points.

The far ahead argument is something I read weekly on investing boards

Some claim tsmc is decade ahead, other say it is 3 to 5 years, etc, etc.

From my understanding both are in a race that will be decided in next 10-14 months and when it comes to the future, Intel has priority on next gen EUV (HighNA), so there is a chance

Various medias report that Intel has 1-2 year advantage due on tech like PowerVia

So, it is not that clear how far ahead TSMC is

_____

>TSMC isnt stupid

Yes, they arent, so arent their customers who want geo diversified supply chain and risk amortization (Earthquake? War? Drought?)

Also they will have leverage when negotiating deals with TSMC

I am somewhat skeptical about high-NA EUV. There are reasons why Intel got priority on that: ASML marketing materials themselves imply high-NA EUV is worse than low-NA EUV in terms of cost.

https://www.semianalysis.com/p/asml-dilemma-high-na-euv-is-w...

> Also they will have leverage when negotiating deals with TSMC

Except Intel hasn't talked about price at all...they have a completely different cost structure that isn't competitive at all. Think about AMD gross margin vs. Intel's gross margin. AMD has similar...all while giving TSMC 50%+ gross margin also. Now tell me Intel's cost will be similar to TSMC...

Their foundry business is a tiny fraction of their revenue at this point and I'm not sure their overall gross margin is particularly relevant here (also Intel has lower gross margins than AMD or TSMC these days?)

If they can't make good gross margin when making their own products that are higher priced (than similar AMD products)...how are they going to do it when they have to support an end customer that expects high gross margins?

Intel GM ~46% AMD GM ~46% TSMC GM ~53%

As they always say in investing, "Past performance is no guarantee of future results."

It wasn't that long ago that Intel was dominating and it looked like nobody would ever dislodge them. Once the top management that drove a company to success retire, they are almost guaranteed to be replaced by people whose main skill set was political navigating in a large bureaucracy and then the decline begins.

In this case it does make sense for Intel to catch up - right now NVIDIA is a quasi-monopoly in the AI space due to the stranglehold on CUDA - and all competitors are working day and night to close the gap in the toolchain.

I think it's a plausible conservative prediction that Intel will close the gap somewhat and will have a significant (but not dominant) slice of the pie.

To be fair their stock wasn't necessarily doing that great at all (relative to AMD , let alone Nvidia over the last few years) when they were completely dominated the market.

Even with all dividends included their stock is worth about as much as back in 2000 (which they only managed to surpass in 2019/2020)), and if they are excluded they are yet to surpass their dot.com peak...

Even if they can catch up with the technology...can they high volume manufacture it? Can they manufacture it at a similar price point as TSMC?

All signs point to No. They have the talent to have competing tech...but they have shown time and time again that they can't do high volume manufacturing or do it at a price that also allows their customers to have high gross margins (like AMD, NVIDIA, etc).

>but they have shown time and time again that they can't do high volume manufacturing

What do you mean with that?

There is a big difference in making a technology work and making a lot of high yielding wafers with it. I'm not sure Intel 7nm ever got to high volume manufacturing...there were constant delays and yield issues. Read the 3rd article I have below to show how bad it is. Plans to implement SR in 2021, but yields were crap and mass production was put out to 1H23. Intel and TSMC use very different terminology for when things are ready...manufacturing ready (Intel) is being assumed to be the same as high volume manufacturing (TSMC), but they are not the same at all. TSMC is doing 2nm risk production in Q4 this year, we will see what Intel ends up doing...

Example(s): https://www.theverge.com/22597713/intel-7nm-delay-summer-202... https://www.pcmag.com/news/intel-sorry-but-our-7nm-chips-wil... https://www.eenewseurope.com/en/poor-yield-at-intel-delays-s...

In 2015, why would anyone have believed AMD could catch up to Intel? Let's not forget that the Bulldozer/Excavator days are not that long ago, and those chips were even less competitive in 2011-2016 than Intel's current offerings are in 2024. I think we can confidently say that bigger architectural gaps have been closed before.

Comparing Intel to other fabs, it's uglier, sure. The Krzanich years were characterized by over-promising and under-delivering, and Intel isn't far enough removed from that period for trust to be rebuilt. That said, they have shrunk the process gap over the last couple years, even if the gap isn't entirely closed.

Given the firehose of government money being directed at domestic manufacturing right now, I assume Intel is doing fine on the "shitloads of money" front.

I don't think Intel is a favorite to be the world's process leader by 2025, but it's reasonable to think they could get close. I don't think the doom-and-gloom perspective is well-founded at all.

That's a great username, do you have another hardware investments you've made?

I invested in AMD around 95 levels, but I started selling them on recent tops around 175-180 to buy more INTC when they dropped after their guidance for Q1

It is not that I dont believe in AMD, I just believe that their growth potential is lower than INTCs.

Foundry businsss is crazy and has waaay higher entry level

What is your cost analysis on Intel's foundry arm? If AMD can support high profits using TSMC as a supplier...who also has high profits, why do you think Intel can compete on price and make good money?

Well, it is not like they arent aware of it.

They were pretty open about this on some earning calls and how do I see it?

I see it as room for improvement and maybe some kind of potential edge over duos like amd+tsmc

I see TSMC and Intel cost structures all the time...they run fabs very differently and the costs are no where near close to each other. When running a foundry the main component of being competitive is price and yield. Intel will struggle with both...mark my words.

ARM is about to announce a breakthrough in microarchitecture like you’ve never seen.

How about a hint? A source? A time frame? Anything?

Their name is insider-trade, so I assume it is a joke gimmick account.

Reply on news.ycombinator.com