I think one day we’re going to wake up and discover that AWS mostly runs on Graviton (ARM) and not x86. And on that day intel’s troubles will go from future to present.<p>My standing theory is that the m1 will accelerate it. Obviously all the wholly managed AWS services (Dynamo, Kinesis, S3, etc.) can change over silently, but the issue is EC2. I have a MBP, as do all of my engineers. Within a few years all of these machines will age out and be replaced with m1 powered machines. At that point the idea of developing on ARM and deploying on x86 will be unpleasant, especially since Graviton 2 is already cheaper per compute unit than x86 is for some work loads; imagine what Graviton 3 & 4 will offer.
The thing about all of these articles analyzing Intel's problems is that nobody really knows the details of Intel's "problems" because it comes down to just one "problem" that we have no insight into: node size. What failures happened in Intel's engineering/engineering management of its fabs that led to it getting stuck at 14 nm? Only the people in charge of Intel's fabs know exactly what went wrong, and to my knowledge they're not talking. If Intel had kept chugging along and got down to 10 nm years ago when they first said they would, and then 7 nm by now, it wouldn't have any of these other problems. And we don't know exactly why that didn't happen.
I found the geopolitical portion to be the most important aspect here. China has shown a willingness to flex its muscles on enforcing its values beyond their borders. China is smart, and plays a long game. We don't want to wake up one day and find they've flexed their muscles on their regional neighbors similar to their rare earths strong-arming from 2010-2014 and not have fab capabilities to fall back on in the West.<p>(For that matter, I'm astounded that after 2014 the status quo returned on rare earths with very little state-level strategy or subsidy to address the risk there.)
> This is why Intel needs to be split in two. Yes, integrating design and manufacturing was the foundation of Intel’s moat for decades, but that integration has become a straight-jacket for both sides of the business. Intel’s designs are held back by the company’s struggles in manufacturing, while its manufacturing has an incentive problem.<p>The only comparable data point says that this is a terrible idea. AMD spun out GlobalFoundries after a deep slide in their valuation, and the stock (as well as the company's reputation) remained in the doldrums for several years after that. Chipmaking is a big business and there are many advantages to vertical integration when both sides of the company function appropriately. If you own the fabs and there is a surge in demand (as we see now at the less extreme end of the lithography spectrum), your designs get preferential treatment.<p>Intel's problem isn't the structure of the company, it's the execution. Swan was not originally intended as the permanent replacement to Krzanich[0], and it's a bit strange to draw conclusions about whether the company can steer away from the rocks when the new captain isn't even going to take the helm until the middle of next month.<p>People are viewing Intel's suggestion that it may use TSMC's fabs for some products as a negative for Intel, but I just see it as a way to exert pressure on AMD's gross margin by putting some market demand pressure on the extreme end of the lithography spectrum (despite sustained demand in TSMC's HPC segment, TSMC's 7nm+ and 5nm are not the main driver of current semiconductor shortages).<p>[0] <a href="https://www.engadget.com/2019-01-31-intel-gives-interim-ceo-robert-swan-the-full-time-job.html" rel="nofollow">https://www.engadget.com/2019-01-31-intel-gives-interim-ceo-...</a>
The "US manufacturing is actually stronger than ever" camp used to cook their books by over-weighting Intel profits. Hopefully this will be a wakeup call.
How does Moore's law figure into this? I suspect that TSMC runs into the wall that is quantum physics at around 1-2nm. Considering that TSMC has said that they will be in full production of 3nm in 2022, I can't see 1nm being much beyond 2026-2028. What happens then? Does a stall in die shrinks allow other fabs to catch up?<p>It appears to me that Intel stalling at 14nm is what opened the door for TSMC and Samsung to catch up. Does the same thing happen in 2028 and allow China to finally catch up?
Feel this piece ducks one of the most important questions - what is the future and value of x86 to Intel? For a long time x86 was one half of the moat but it feels like that moat is close to crumbling.<p>Once that happens the value of the design part of the business will be much, much lower - especially if they have to compete with an on form AMD. Can they innovate their way out of this? Doesn't look entirely promising at the moment.
I worked at Intel in 2012 and 2013. Back then, we had a swag t-shirt that said "I've got x86 problems but ARM aint one".<p>I went and dug that shirt out of a box and had a good laugh when Apple dropped the M1 macs.<p>Back then, the company was confident that they could make the transition to EUV lithography and had marketing roadmaps out to 5nm...
Perhaps now the "Why do people obsess over manufacturing?" question that many tech workers ask when other US industries were decimated will become a bit less quizzical.
> And in that planning the fact that TSMC’s foundries — and Samsung’s — are within easy reach of Chinese missiles is a major issue.<p>Are processor fabs analagous to auto factories and shipyards in World War II? Is the United States military's plan for a nuclear exchange with China dependent on a steady supply of cutting edge semiconductors? Even if it is, is that strategy really going to help?<p>This article is mostly concerened with Intel's stock price. Why bring this into it? Let's say Intel gets its mojo back and is producing cutting edge silicon at a level to compete with TSMC and supplying the Pentagon with all sorts of goodies... and then China nukes Taiwan? And now we cash in our Intel options just in time to see the flash and be projected as ash particles on a brick wall?<p>"The U.S. needs cutting edge fabs on U.S. soil" is true only if you believe the falied assumptions of the blue team during the Millenium Challenge, that electronic superiority is directly related to battlefield superiority. If semiconductors are the key to winning a war, why hasn't the U.S. won one lately?<p>And what does any of this have to do with Intel? Why are we dreaming up Dr. Strangelove scenarios? Is it just that some people are only comfortable with Keynesian stimulus if it's in the context of war procurement?
This was a well-written article, but I don't think it came from someone with a deep understanding of semiconductor technology and fabrication.<p>Intel hasn't lost to Apple and AMD because they employ idiots, or because of their shitty company culture (in fact, they're doing surprisingly well <i>in spite</i> of their awful company culture). Intel lost because they made the wrong bet on the wrong <i>type</i> of process technology. 10 years ago (or thereabouts), Intel's engineers were certain that they had the correct type of process technology outlined to successfully migrate down from 22nm to 14nm, then down to 10nm and eventually 7, 5, and 3nm. They were betting on future advances in physics, chemistry, and semiconductor processes. Advances that didn't materialize.<p>EUV turned out to be the best way to make a wafer at lower transistor size.<p>So now Intel's playing catch up. Their 10nm process is still error-prone and far from stable. There are no high-performance 10nm desktop or server chips.<p>That's not going to continue forever though. Even on 14nm, Intel chips, while not as fast as Apple's M1 or AMD's Ryzen 5000 series, are still competitive in many areas. Intel's 14nm chips are over 6 years old. The first was Broadwell in October 2014. What do you think will happen when Intel solves the engineering problems on 10nm, and then 7nm? And then 5nm?<p>It took AMD 5 years to become competitive with Intel, and over 5 to actually surpass them.<p>If you think the M1 and 5950X are fast, then wait till we have an i9-14900K on 5nm. It'll make these offers look <i>quaint</i> by comparison.<p>EDIT: I say this as a total AMD fanboy by the way, who bought a 3900X and RX 5700 XT at MicroCenter on 7/7/2019 and stood in line for almost five hours to get them, and as someone who now has a Threadripper 3990X workstation. I love AMD for what they've done... they took us out of the quad-core paradigm and brought us into the octa-core paradigm of x86 computing.<p>But I am under <i>no</i> illusions that they're technically superior to Intel. Their <i>process</i> is what allows them to outperform Intel, not their design. I guarantee you that if Intel could mass produce <i>their</i> CPUs on <i>their</i> 7nm process (which is far, far more transistor dense than TSMC's 7nm), AMD would be 15-25% behind on performance.<p>It isn't so much that AMD is succeeding because they're technically superior... they're succeeding because Zen's design team made the right bet and because Intel's engineering process team made the <i>wrong</i> bet.
A major incentive for the US government to get involved is touched on. Not only is Taiwan 'just off the coast' from China, China is coming for it and intends to assimilate Taiwan back into China just as Hong Kong and Macau have been.<p>At that point, the only sustainable leverage the rest of the world would have in chip technology would be ASML.
Contrary to the article, AMD is not yet shipping 5nm in volume. (Rumors point to Zen 4 later in 2021.)<p>Additionally, Intel works with ASML and other similar suppliers. Intel even owns a chunk of ASML.
I wonder how the move from x86 to arm is going to affect desktop apps. With the move to ARM apple is already pushing it's iOS apps into MacOS. Once it becomes commonplace on Windows, it would be super easy to run Android Apps on Windows via simulation(rather than emulation which is much slower).<p>Given that mobile apps are more lightweight and consume far less resources than their electron counterparts, would people prefer to use those instead? Especially if their UIs were updated to support larger desktop screens.
> Solution One: Breakup<p>> Solution Two: Subsidies<p>Solution Three: lower prices/margins (temporarily) to match the value proposition of AMD on Windows PCs and Linux Cloud servers.
If individual companies developing their own chips is a trend, and it sure seem like it is starting to, intel has a lot more competition they have to contend with. Before is always a buy, now add build into the equation. That’s where intels problems are coming. That’s a lot of head winds, they can capture that by splitting and going the TSMC route, and specialize further on the design and use some form of licensing model like ARM.<p>This is like the Microsoft pivot into cloud to save itself.
The funny thing is, in the time period being addressed first in the article (2013) Intel was better at mobile than it is now. Its Bay Trail and Cherry Trail chips had more performance per dollar than even today's offerings, eight years later. Intel just decided low-margin wasn't a concept in which they were interested.
When I learned that allegedly Intel and nVidia were fixing the laptop market, I just hope that this company goes down or goes through a substantial transformation. Their current management situation is untenable. If I was a shareholder (fortunately I am no longer), I would pressure them to sack everyone involved.
Totally reasonable argument, and I think most would be better off with an independent, US-based foundry.<p>Unfortunately, I doubt that the US government functions well enough at this point to recognize the threat and overcome the influence Intel's money would wield against the effort.
A question for those who've been around in tech longer: was Google really the first and "disruptive" user of x86 commodity hardware in datacenters that everyone else then lagged behind? Or was it just a general wave and shift in the landscape?
> <i>a federal subsidy program should operate as a purchase guarantee: the U.S. will buy A amount of U.S.-produced 5nm processors for B price; C amount of U.S. produced 3nm processors for D price; E amount of U.S. produced 2nm processors for F price; etc.</i><p>I really like this concept, though I’d advocate for a straight subsidy (sales of American-made chips to a U.S.-registered and based buyer get $ credit, paid directly to the supplier and buyer, on proof of sale and proof of purchase) given the logistical issues of the U.S. government having a stockpile of cutting-edge chips it can’t dump on the market.
If intel made a SBC or SOC design for low power applications, I'd consider it if they had long term support. Intel used to power all of our edge needs in POS and Security, now I see that slipping as well.
Slightly more aggressive take: fully automated contract manufacturing is the future, those that resist its march will be trampled and those that ignore it will be left behind.<p>Semiconductor manufacturing is just one example where this is happening, electronics is another. Maybe one day Toyota Auto fabs will be making Teslas.
Companies that use millions of micros will grow tired of paying royalties for ARM & other IP. I'm putting my money on RISC-V. If Intel is smart, they will too and offer design customization and contract manufacturing of RISC-V.
Solution three : Invest to make x86 to become power efficient to reach at par or better than arm counter part, while outsourcing manufacturing to TMSC to fill the gap. Reach to the level Future Apple M chip will achieve. At the same time, start building bare metal cloud hosting solutions that allow other companies to provide their own cloud solutions, ( and using energy efficiency at your advantage ), also use that energy efficiency to create mobile platform that can enable Mozilla and Ubuntu like provider to make operating syatem on.
It's time for more predictions.<p>1. Apple's CPUs will not improve anywhere near as fast as the competition. Computation per watt of (some) competitors' products will outpace Apple's in just a few years.<p>2. Intel will come roaring back on the back of TSMC, but first will need to wait on growth of manufacturing capacity, as certain competitors can get more money per mm^2.<p>3. Intel will fail to address its product-quality problem, but it will not end up hurting them.
Though provoking, to be sure, but the problem with his solution of building up Intel's manufacturing arm through spinoff and subsidy is that we simply don't have the labor force to support it, and with much more controlled immigration in the future, it will take decades to build up the engineering education needed to make the US compete with Taiwan, South Korea, and of course China.
A solution to yesterday's problems shouldn't discount tomorrow's innovations. I don't think iOS and Android are in the best long-term position. There's more things happening in our global infrastructure that should be accounted for.
Internet is priming for a potential reverse/re- evolution of itself. (5G is a large factor for this).
How does RISC factor into all of this? China is working on its own chips after Trump poked the bear. It wouldn't surprise me if, seemingly out of nowhere, they suddenly started promoting and selling RISC chips and compatible motherboards at competitive prices.
If TSMC is going to be a monopolist fab for x86, then they will ultimately suck all the profits out of the server/desktop markets. This isn't just kinda bad news for Intel/AMD, it's really bad news.
Well I got down to the part where the author said that AMD never threatened Intel in the data center market and I closed the tab. AMD won entire generations of data center orders while Intel was flailing with Itanium and NetBurst.
A company that is being cannibalized by companies ripping the rug under them by developing their own chips, yet Intel makes no efforts to return the favor. They need to make an open source OS phone, maybe that will make a dent and serve as a carrier for the chips. They dont need to do all the work they can partner.
Lets speed run a doomsday:<p>2022: share price tanks, ceo booted, they shuffle but dont have a plan, no longer blue chip so finance is hard to come by. delisting. everyone booted. doors close.<p>2023/4: AMD only game in town. profits and volumes up. so are the faults and vulnerabilities. They spend most of their effort in fixes and not innovation.<p>2024: M1 chip available on dells/hps/thinkpads. AWS only use Graviton unless customer specifically buys another chip.<p>2025: Desktop ARM chip available on dells/hps/thinkpads.
2025: AWS makes a 'compile-to-anything' service. decompiler and recompiler on demand.<p>2026: AMD still suffering. Hires Jim Keller for the 20th time. makes a new ZEN generation that beats M1 and Arm. AMD goes into mobile CPUs.