Intel cutting cutting-edge node funds would mean no more Moore's Law Intel chief executive Lip-Bu Tan has warned that he may pull investment from Intel's leading-edge 14A semiconductor process node unless "a meaningful external customer" can guarantee profits – a move which may finally spell the end of the chipmaker's loyal adherence to Moore's Law. "Our external foundry strategy has always …
Moore's Law was only an observation and forecast, later revised downwards.
While we do in get incremental improvements, a decent laptop doesn't need replaced for maybe 5 or 6 years now, unlike annually for a huge boost when the comment was originally made.
It hasn't applied in the original sense for maybe over 20 years.
Agreed. I only recently replaced my X1 Carbon gen 6 (8th gen Intel)...but not for performance reasons, I just wanted something smaller.
Its still fast enough and nothing has come along that has blown my hair back. Everything since 8th gen Intel has felt like a battery life or thermal trade off rather than an actual upgrade.
My desktop gets upgraded at a fairly regular cadence (once every two years or something) but laptops though...not so much.
My server (which I recently upgraded some components in) is from 2017. Its been in service for me for 5 of those years 24/7 and it has been running in less than ideal conditions for a server. Its a Fujitsu Primergy RX and its had zero problems other than the odd disk failure, which is to be expected. I know, I know...for some reason people hate Fujitsu servers...but hear this...in terms of servers, Fujitsu make the best servers if you're comfortable modding them to make them silent...their proprietary stuff (power supplies) are also very easy to repair if you're handy with a soldering iron.
Fujitsu servers are the thinking mans solution for a homelab / freelance dev platform.
Mine runs at just under 90W, its completely silent (Noctua mod) and has PCIE lanes for days.
I also have a couple of PowerEdge R230 boxes, but they suck for various reasons. Those bastards run hot!
If 14Å and smaller is abandoned by Intel and other US entities and the PRC reaches that goal then that would pretty much proclaims to the planet which nation was preeminent.
National prestige and the "soft power" projected by culture and technology have delivered incalculable benefit to the US from well before the outbreak of WWII†; now rapidly and senselessly squandered by a unprecedented convocation of fools.
† for better part of world September, 1939.
Chinese chips will soon be outperforming US chips (soon to be renamed as fries for all the use they'll be).
Come the day when the US wants powerful Chinese chips. Sorry, me old mates, remember 2025? Too many military uses. And that will be that. Game over. Geniuses didn't see that coming.
"National prestige and the "soft power" projected by culture and technology have delivered incalculable benefit to the US"
It was Bretton Woods what did it. They told that has-been, England[1] that they were the big boys now, and to step aside and let the dollar take the strain. No other currency was convertible to gold apart from the dollar (you could, of course, get dollars for your Marks, Francs, Pounds, but not gold). Then they could happily print dollars and the US allowed large imports and their economy bloomed.
They made a secret deal with Saudi Arabia that gave the Saudis military support on the back of only selling their huge amounts of oil in dollars. Opps. That put far too many dollars into the global economy, and after the US decided they wanted to bomb the shit out of Vietnam -- and printed yet more dollars to pay for it -- inflation took off.[2] This had the result that a certain Mr Nixon in 1971 took the US off the gold standard, meaning _no_ currency was backed by anything. And we have been living in that make-believe world ever since. All currencies are now fiat currencies -- just pretend money that we all agree not to call out.[3]
[1] Or is that the UK, is there any difference? Pounds are produced by the Bank of England.
[2] Note that the 'oil price shock' of the 1970s wasn't the cause of the high inflation and low productivity ('stagflation') of that decade, but the consequence of these oil-producing nations getting less and less for their oil. So they put up the price. What would you have done in their shoes, one wonders, if you have nothing else you can sell?
[3] So will crypto finally upset the apple cart? After all, who controls that?
" Pounds are produced by the Bank of England."
And Bank of Scotland, Royal Bank of Scotland, Clydesdale Bank, Bank of Ireland, Northern Bank and National Westminster bank (Ulster Bank). We should probably go with UK, there is a very significant difference.
My local pharmacy has a sign up that says they refuse to accept notes from the Bank of Scotland. Is that legal? That's one reason why I emphasised England over the UK. Of course I didn't mention that England has 85% of the UK's population. And that is not relevant to anything, no siree.
Edit to add: mind you I got some cash out recently -- the first time I've had to use cash in months. It looked wrong -- where's the queen gone?
My local pharmacy has a sign up that says they refuse to accept notes from the Bank of Scotland. Is that legal?
Yes. But it is not why you think.
"Legal tender" as a concept exists in English & Welsh law, but only applies to payment of a debt (there are some restrictions like you have to offer the exact amount: you cannot demand change). And retail sales do not create a debt (but a bill at a restaurant, where you pay at the end of a mail, does). Assuming you have a debt to pay, then the creditor has to accept legal tender. However (1) for small change there are maximum limits (eg. the creditor only has to accept 20p in copper coinage), (2) Bank of England notes are legal tender, (3) All other bank notes issued in the UK (eg. Bank of Scotland) are not legal tender (they are not even legal tender in Scotland: only £1 and £2 coins are for an unlimited amount).
For retail it is up to the vendor... and providing they clearly state limitations on the payment types they accept.
My, have the mighty fallen.
Time for obituary, I guess :
Here lies Intel, what's left of it's reputation bleeding out as fast as the ink fills its books.
Let us not forget that Intel basically created and the PC market and the upgrade treadmill we all so joyously participated in for more than a decade and a half, with Microsoft reaping much of the rewards.
Alas, poor Intel, your very success attracted the competition that ended up eating your lunch.
I, for one, will remember you fondly, always.
If you don't build it, then, obviously, they won't come! The uncertainty-inducing volatility of this Intel announcement reminds me of Trump ... is he now leading Intel in the straight zigzag course of a right drunkhard's random walk?! (or what!?)
Intel needs leadership, it needs a vision, a path through the future that generates inspiration, awe, and confidence, but what does it get: "huh, if it rains, we will make umbrellas, but if the sun shines, huh, we won't". What a plan!!!
I mean, the outfit's run 30,000 High NA EUV wafers on the Twinscan EXE:5000 machines it aggressively invested in and followed through with EXE:5200 devices on its way to 14Å volume production (some billions in investment) as first in this tech (leadership) and now ... "huh, yeah, huh, we'll just surrender our lead, hey, how 'bout that, huh ...". That's worthy of the Most Super Dumb Award right there in my book!
So, yeah, they can do 18Å in Low NA EUV which may be cheaper (and TSMC wants to do A14 with this "EOL"-to-be tech) but abandoning leadership is nothing more than un-American surrender monkeyism imho. UNLESS there's a technical reason why High NA EUV 14Å is troublesome (something like a mysterious Stochastic Resolution Gap, if that exists at all) ... otherwise, NO! Stalled lukewarmness is not leadership; not by a long shot!
The only thing Intel has left right now is inertia. They have no meaningful performance leads anywhere, by any meaningful metric. They're behind AMD on high-end compute, behind ARM on low-power compute, nowhere on AI / GPU compute, and abandoning the other areas (fabrication, networking, FPGAs, etc.) where they at least has a possibility of differentiating in a positive way, even if it's Intel not having to bid for wafer starts at TSMC vs. all of its rivals.
At this point Intel's only hope is that AMD makes a major stumble as it grows, or somehow gets stuck with bad leadership after Lisa Su retires. These things happen to the very best of companies, eventually (see: Intel). But with many of their best engineers retiring or leaving the sinking ship, I think Intel is now, as Douglas Adams so eloquently put it, dead - it just hasn't stopped moving yet.
"At this point Intel's only hope is that AMD makes a major stumble"
If Intel aren't even trying to compete, I don't see how that would help them.
Announcing that you won't try to reach 14Å is, in this industry, basically a winding up order. What's the point in holding Intel shares at this point?
If Intel no longer has a premier fab line to compete with TSMC, it is over. Might as well turn off the lights on your way out tonite Lip. Meanwhile China is aggressively investing in everything from the ASML clones to fabs. All with no guarantee of a return. I watch helplessly as Rome burns. The color orange of the clown is appropriate for the US burning down.
The increase in capital costs at Intel 14A make it clear [however] that we need both Intel products and a meaningful external customer to drive acceptable returns on our deployed capital. And I will only invest when I'm confident those returns exist."
Because taking a risk never did anyone any good?
A veteran tech person told me long ago that companies "can't cut their way to success".
You may be able to cut tactically, e.g. to get over a hopefully-temporary economic slump or an end-of-life product or something along those lines.
But just cutting (people, projects, buildings, budgets etc.) because your business is bad, and maybe has been for a while, is basically a slow painful war of attrition which you'll probably lose.
Once you start cutting, there needs to be a plan for when to stop, and what to do afterwards. "Back to business as usual" is not a plan, neither is hoping for a buyout/merger. Though the latter often turns out okay for those with golden parachutes.
Not clear yet if the people with say-so at Intel have a plan for "after". LBT has apparently been talking about selling off parts of the business, but it may be wishful thinking there will be a rush of takers, or that the moves will be enough to turn the company around.
Selloffs and layoffs seem like trading people for time, but how to use that time and what to do after remains unseen outside the executive suite. For Intel employees' sake hopefully there is a plan.
And software should be able to run on any system, regardless of the processor. GAFA haven't innovated for a very long time. We got lazy and let them get away with it.
But as others have said, we simply don't need more speed. Just the ability to turn crap like AI and Recall off. Bandwidth took over from speed a while back, and we may not need more bandwidth now either. Most of us could still be using 15 year old machines, if it wasn't for MS and Apple forcing upgrades to maintain their revenues. The resulting e-waste is on them.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
