Intel ChipShortages To Continue Through 2020 - Are intel just up to shenanigans at this stage?

Multiple times.

This being one of his worst offenses

To be fair, this is quite hard to wrap your head arround:

Or this one (note the “TBD” on the 10n, stuff):

1 Like

Maybe they did and resurrected it. Or maybe he’s referring to above, shows later roadmap with “tbd” for 10nm. Rocket lake changed from 10nm to 14nm on a later roadmap. i.e., abandoned on 10nm…

Either way, 10nm parts STILL aren’t out in quantity, and its 2020. Sounds very much to me like it is an abandoned node, shitting out just enough low-end mostly broken parts to claim to shareholders “See, 10nm worked!” while they skip to 7nm instead. Because 10 is so badly broken.

Lets remember that these were supposed to be shipping in 2015-2016 originally. If 10nm isn’t dead, it’s pretty damn close to it. Unless intel plan on getting into the overpriced keychain business?

Basically intel are caught in a lie to their shareholders. They’ve been claiming 10nm is on track for so long they need to ship “something”, even if it is killed internally.

3 Likes

Intel has been the Apple of the processor industry since the pentium 3.

Yes, they’re idiots.

1 Like

If they go on a fire sale I will happily make keyrings out of them. I have lots of laptop ram but really want some more CPUs. Will have to get on eBay or something I don’t have an AMD keyring yet.

1 Like

Posted this in the wrong thread yesterday and just found it to move here :smiley:

Rumour mill says going fabless is maybe an option.

Not sure how i feel about that.

On the one hand, they will need to compete on equal design footing. That will be… interesting. Previously intel’s trump card was their fab process superiority. This would be a bitter pill to swallow for intel.

On the other hand…this could be bad. They have (much) deeper pockets than AMD (especially if they sell their foundries) and high end fab capacity is limited. They just out-bid AMD for capacity.

Maybe this is how they buy their way out of trouble.

That’s interesting. What would happen their current fabs? Spun off private? Those should still maintain the superior process tech, and I imagine it would be like AMD and GolalFoundries where Intel would have a contract with their previous fans to have Intel use them for their supply while also augmenting it with others.

On the other side that might mean the superior process tech would be available to others. So if they manage to geth their “Intel 7nm” working which they compare to 5nm TSMC I believe then other people besides Intel would have access to that.

Or would hey be seen as only useful for 14nm+++(+++++++) and just collapse under their own inability to keep up while Intel sheds the weight.

I imagine too with the money they could bring whoever they do go with (probably TSMC) could then expand to meet the demand while acomodating everyone. It is times like these I can see AMD racing to sign contracts with TSMC to assure Fab time.

What are Samsung up to these days, I don’t hear them doing much outside their own stuff and apples. And Samsung are also killing off their own Arm designs so.might have a little more space on their lines.

Well.

They USED TO have superior process tech. Now they don’t. If they did, they’d be using it :smiley:

intel 7nm doesn’t exist yet, 10nm is broken. 14nm+++ is good for what it is. But outside of CPUs, not sure what one would use it for. Chipsets? RAM? Other general ICs?

IoT will want ultra low power (7, 5, 3nm as available) and are small/cheap to make anyway.

Maybe FPGA… i think that’s a growing market.

Either way it’s all rumour, but one of the big finance companies is considering that it’s one of their realistic options.

1 Like

Well it all on its head now. They are still doing better than last year and FPGAs are down.

I mean this is no surprise to the people who pays attention to the industry instead of the “intel is going fabless” rumor mill that has zero bearing on reality.

Will see how that works out this year. They’ve been unable to increase volume and have dropped their pants on pricing (their most profitable parts are 50% off right now - whilst AMD has improved performance and will have more capacity available mid year).

And they are no longer holding any sort of performance advantage.

So why is Intel 10nm failing. Did someone just absolutely fuck up the design? Sidenote I know enough about CPU fab to be mildly dangerous.

It’s not failing, the clocks are just relatively low, and ramp up is taking longer than expected.

The first 14nm iteration also capped out on about 4.1GHz, and in fact, TSMC 7nm isn’t clocking much higher.

In April 2018, Intel delayed mass production to 2019
Now it is 2020, so something is clearly very wrong with the process.

Edit:
This source says, 10nm was meant to go into production in 2015

Intel’s 10nm technology had already been substantially delayed (it was supposed to go into mass production by the end of 2015, but that schedule kept getting pushed out), so further delays only make the situation worse.

Delayed no doubt.
Btw it’s already in mass production.

In the smallest possible format, so mobile-only chips.

Indeed, 14nm started out the same way.

14nm did not have a 5 year delay before it started shipping though.

14nm started shipping just over 5 years ago, at that time 10nm was not supposed to ship simultaneously.

While it’s been more delayed than any other process, the process itself is not a failure.

@sanfordvdev Like you, I am interested in the technical reasons for delay of the Intel 10nm process. But Intel has managed to hold this info very closely, and I don’t think we outsiders really know. Of course, making semiconductors with such tiny features presents pretty daunting challenges. I don’t believe that Intel has been “lazy” or “didn’t try hard enough”.

There have been suggestions that they took some decisions on new materials and process steps that were high-ish risk but would have been high reward if it had worked out. But that partly comes from a CEO’s statement which may have been influenced by PR.

The only thing semi-solid I have heard is that they began to make use of cobalt in the electrical connections, and encountered unexpected problems. That was an attempt to work around the issues that copper presents at such small dimensions. However, that isn’t definite, and even if true it doesn’t seem like a complete explanation.

It does seem like they are gradually getting it sorted out; it will be interesting to see when 10nm desktop CPUs ship.