The Intel Split
stratechery.com
stratechery.com
The risk is obvious: The company's internal manufacturing operations can survive only if they figure out how to outperform competitors. The upside is that all other parts of Intel stand a good chance of surviving and thriving -- but that remains to be seen.
Only one thing is certain: Gelsinger and his team have a lot of cojones.
This seems to be sound long-term though it may cause significant shrink in the business in next couple of years to cut lines of business that are not above water.
From the point of view of infinite game it makes a lot of sense to diversify and then cut any branch of business that is not profitable, before it drags entire company.
While I think it is difficult to predict results, one thing is sure -- this will force a lot of change in Intel and we would likely not recognise the company ten years from now.
This is putting lots of high-level employee's future earnings on the line in a far more direct manner. It will be interesting to see if they accept this challenge, or fight it in order to accept the slow decay that will still ensure at least a longer-term personal financial gain (i.e., instead of failing in 2 years, failing slowly over 10).
I think this might be an interesting retrospective application of The Innovator's Solution's 'be willing to cannibalize revenue streams in-order to avoid disruption.'
I can see those general outcomes:
* Intel manufacturing is soon again better than other manufacturers in all aspects as it has been a long time. All Intel business goes to Intel manufacturing and things return to what they were.
* Intel manufacturing is better in some aspects and continues to outcompete other manufacturers in some areas. Most Intel business goes to Intel manufacturing but some special business goes to other manufacturers. Some other companies use Intel where they have an advantage. Everybody wins.
* Intel manufacturing cannot keep up with the market leader. Most Intel business goes to the leading Fab. Intel manufacturing cannot catch up as they don't earn enough money to do so. They start cutting costs and compete downmarket with lower tier fabs. In the long run Intel becomes fabless.
Sophie Wilson has said that cost-per-transistor is rising since we shrank below 22nm, and that a 7nm die cannot run more than half of its transistors at the same time without melting.
Sophie addresses the cost rise at 22:00.
https://m.youtube.com/watch?v=zX4ZNfvw1cw
Heat is more of a problem because finfet can't dissipate as well as planar, and reliability is declining at smaller nodes.
"With a planar device, you do not have to bother about self-heating. There are a lot of ways to dissipate heat with a planar device, but with finFETs that is not the case. The heat gets trapped and there are few chances for that heat to get dissipated."
https://semiengineering.com/chip-aging-becomes-design-proble...
https://news.ycombinator.com/item?id=29889951
If I needed a 20-year service life, I would not choose 7nm.
Edit: I did not realize that ARM1 had fewer transistors (~25k) than the 8086 (29k), and over ten times less than the 80386 (275k). Intel should have bought ARM in the 80s; instead Olivetti got them.
The data has been saying otherwise. 5nm is the only node that increased $/transistor beyond it's previous node (7nm), and that's at a time when Apple payed out the ass for a monopoly on the node except testing shuttle runs from competitors but isn't a sign of a fundamental increase in cost.
https://cdn.mos.cms.futurecdn.net/Unwdy4CoCC6A6Gn4JE38Hc-970...
> and that a 7nm die cannot run more than half of its transistors at the same time without melting.
That's true for trying to run multiGhz designs in classic plastic package BGAs like most cell phone chips, but that's been true for a while, hence why flip chip packages are a thing. Actually having a heat spreader connected to the die goes a long way.
Wilson's comments aren't incorrect from a certain point of view, but tend to get extrapolated out of her niche to the greater industry in a way that pushes the statements into inaccuracy.
That sort of coincides with TSMC having competitive, close to leading edge nodes (so the 28nm timeframe) which would line up with the rumor. The information simply hasn't been updated over the timeframe of the node. Previous to that the cost of the node was pretty fixed as long as someone like ARM cared about, now there's a lot more economic effects from the increased buyer competition that heavily changes final cost over time.
At the same time, AFAIK Intel was doing quite well at 14nm finfet even then (likely better than any other foundry?), but that production capacity was not available to ARM, so I guess it didn't count.
> At the same time, AFAIK Intel was doing quite well at 14nm finfet even then (likely better than any other foundry?), but that production capacity was not available to ARM, so I guess it didn't count.
Yeah, and Intel was right in the middle of their biggest misstep. Intel 10nm was in risk production for Cannon Lake with awful yields and therefore a huge $/transistor. It got shipped anyway as one SKU in May of 2018 that didn't make financial sense (there's rumors that management bonuses were tied to that release), before being relegated to process improvements for years until just recently.
It would have been fair for her to extrapolate a trend there that actually ended up being more complex in hindsight.
One simple reason that similarly to aeronautics chip manufactures are critical industries for US.
US government will not let either one die.
Splitting out the fab business and letting it run into the red however could incentivize the government to save the fabs and the jobs with billions of bailout money while the other Intel continues to rake in money with TSMC chips.
> We don't need small nodes for military/aero chips. It would suffice to keep some old fabs running to secure supply.
What? How do you think the military works, via smoke signals? They need computer, a LOT of computers to operate day to day. And that PCs for just boring clerk work. Not to mention the idea that old chips would somehow be good enough to keep up technologically, both military and economically.
> The US needs fabs but they don't need Intel as a company.
Oh yeah? They will just a walk a engineer corps to a fab and have them press big red manufacture button few million times to make enough chips right? Can't be more difficult than that...
https://en.wikipedia.org/wiki/International_Traffic_in_Arms_...
So there won't be any electronics manufacturing to ban at all, but it leaves the US in the same place.
The level of computerization and networking is going up in the military so the needs will only increase. Intel's CONUS fabs are a national security concern.
This isn’t Intel’s shareholders’ problem. Split the businesses. If the U.S. wants to bail out Intel’s fabs, let them. (Though a better strategy would be a general subsidy for anyone building fabs in America, which should be a thing.)
Hell I just replaced thousands of older desktops with slower laptop chips — nobody noticed. Frankly, I only bought new devices because of Microsoft’s bullshit Windows 11 requirements.
The guaranteed cash flow train is slowing down, which makes setting billions of dollars on fire with high risk fab processes a big deal.
Intel had to adapt like this in the early 90s when memory became a commodity. We’ll see if they make it.
Big enterprises think like that yeah, because they can't get their crappy Dell Inspirons to new employees anymore.
Personally I just had to get a Ryzen laptop replacement temporarily because Lenovo took too long to repair my thinkpad. And the result was delightful. Better sound, fantastic compile times, beautiful screen for half the price of my thinkpad. Next time I'll get a cheap Ryzen and if it breaks just buy a replacement device instead of relying on a pro intel enterprise device with crappy overpriced enterprise service.
NYC Public Schools buy like 200,000 computers a year. A decade ago, it was probably 50% higher due to faster refresh cycles. There are a lot more big dumb enterprise endpoints than you might think. When I sold computers in college to commercial customers, the average refresh was 30-40 months. Now it’s closer to 60.
This is certainly plausible if these were mostly Office users but how confident are you that you'd know if they did notice? Most of the large IT departments I've interacted with would post glowing self-evaluations while the users had plenty of complaints and/or were doing their real work on personal devices. Depending on how the business is structured and the relationships, this can be hard to measure without a fair amount of effort — I've heard people say they weren't going to report something because they didn't think it was worth the effort or they were sure that the IT people would notice and fix it.
Google Docs runs well on a Chromebook. As does MS Office online. The modern desktop has really turned into a terminal - it's jut using HTML & JS and works... pretty well.
For me personally, web apps only got fast when I got a computer with M1 processor.
If #1, then everything is rosy and Intel will regain its dominance from the days of Sandy Bridge. Its chips will be fabricated internally and they get to enjoy the profits from vertical integration.
If #3, then Intel will certainly spin off or sell off its manufacturing/fabrication business very similar to AMD more than a decade ago.
The caveat is that I don't think Intel's problem right now is simply a manufacturing issue. Sure their Alder Lake is competitive but it's not superior to AMD's offerings or even measured against Apple's SOCs. Remember that unlike the last iteration (Rocket Lake), Alder Lake doesn't suffer from a mismatch in design vs. manufacturing cadence - it's arrived as expected.
The problem is that by switching to another fab they lose these advantages.
That is one of the reasons Intel was able to get away with their garbage in-house EDA tooling and processes. Intel was also able to get away with delays to fix their bugs (due to awful validation methodology).
Even if Intel process catches up TSMC, they would have fix their design tooling, hire talented engineers and get rid of the entrenched hacks who built their careers on mediocrity.
Trained and mentored by Andy Grove. Disciple of the old school "Only the paranoid survive". I expect nothing less.
I hope someday he could do the next unthinkable, Once their foundry could at least compete against Samsung, open up the x86 ISA. Or clean up the ISA and call it something like AE86. In the very long term, x86 is pretty much dead. By 2030 you would expect all hyperscaler to be on ARM with x86 serving some minor legacy clients. You have PC manufacture eager to compete with Apple and ChromeBook. So x86 could loss up to 50% of its current volume by 2030. Then a long tail of 10 - 15 years decline.
Is a new way I like to be
Is a new way to set me free
The naming is intended to show old things can be made good, and fast :)
( Nice to see somebody gets it )
Is outperforming competitors necessary for Intel's survival? There's plenty of fabs in the world doing quite alright well behind TSMC, the vast majority of which can't even come close to Intel's current capabilities [1]. Even if Intel never succeeds with their existing process roadmap - which is not on pace to beat TSMC - they still possess some of the most advanced chip manufacturing processes in a world that's all the more dependant on chip manufacturing.
GlobalFoundries got out of the race on 14nm - a 14nm not as good as Intel's 14nm and far behind Intel's 10nm - and is still posting YoY revenue growth, despite losing much of the volume they were producing for AMD over the last few years.
In addition to that, I suppose that even if Intel merely succeeds in roughly keeping up with TSMC and Samsung (their current 10nm being on par with Samsung's and TSMC's 7nm would classify as "roughly keeping up", I think) there's American national security interests at play. Especially so if Intel's manufacturing capabilities are accessible to (American) third parties. No way the powers that be would let Intel's manufacturing plants go under.
It's a pretty bold strategy at face value, but I think it's actually a pretty straightforward choice and the risks aren't all that existential.
[1]: https://en.wikipedia.org/wiki/List_of_semiconductor_fabricat...
Perhaps that's part of the calculus too; if they know that Uncle Sam will backstop any failures, are they not also making the same play as their competitors were in the 80s?
The problem is that by the mid-1980s Japanese competitors were producing
more reliable memory at lower costs (allegedly) backed by unlimited funding
from the Japanese government, and Intel was struggling to compete…One needs quite a crystal ball to predict such slow deaths, especially when it's going to be decade where chip production improvements are expected to slow down further and further. Finally, a slow death looks even more likely anyway if Intel doesn't change their ways.
And then Global Foundries couldn't keep up, cancelled their 7nm work, and AMD has been sending more and more business to TSMC.
This is a monumentally stupid idea and, even worse, we have seen it before. Every company that has done this in the past is gone.
I would say that every VLSI designer in Intel should be printing their resumes, but I'm pretty sure that the decent ones left long ago.
In addition, this completely throws away the advantage that having a fab brings at a time when it's finally important.
Fabless works great when you can push a synthesis button every 18 months and your chip doubles in speed. Smart design can't keep up with that.
However, once Moore's Law stops and you can't run every transistor without melting your chip, suddenly design matters again. You want to be able to do creative things with a bit of tweak from your fab.
The approach Intel is taking - outsource cutting edge products to TSMC while continuing to invest in their fabs making their lower end stuff and other people's stuff like automotive chips in-house is the best strategy to buy some time to advance their fabs while letting them earn money to support R&D investments.
It's a huge problem and nobody except TSMC has succeeded at it. Besides there is years of lack of focus, incentives and interest in specialized education and manufacturing processes that'll take time for Intel to fix. Meanwhile they will be competitive in consumer markets by going TSMC 3nm and continue to improve on the side by taking on outside fab orders. Seems reasonable to me.
The holistic efficiency comparison definitely favours Intel on desktop, as Ryzen has heavy idle power usage, due to its 12nm I/O die from GloFo.
In laptops idle power draw and heat are far more important because they effect system endurance. In fact most gains in mobile device power management are with lowering idle consumption not load consumption.
It certainly is if you're a major corporate entity deploying tens of thousands of desktops idling 24/7. The 30-40W IO die idle cost adds up.
If you're doing distributed rendering you're better off buying Ampere Arm CPUs anyways as they're more power efficient than any x86 offering.
That money can dry up very quickly.
Intel's in a much worse position now. Their process woes aren't issues that the rest of the industry will hit in a year or so like the end of Dennard scaling was.
>One of the big takeaways from our initial Core i7-11700K review was the power consumption under AVX-512 modes, as well as the high temperatures. Even with the latest microcode updates, both of our Core i9 parts draw lots of power.
The Core i9-11900K in our test peaks up to 296 W, showing temperatures of 104ºC, before coming back down to ~230 W and dropping to 4.5 GHz. The Core i7-11700K is still showing 278 W in our ASUS board, tempeartures of 103ºC, and after the initial spike we see 4.4 GHz at the same ~230 W.
There are a number of ways to report CPU temperature. We can either take the instantaneous value of a singular spot of the silicon while it’s currently going through a high-current density event, like compute, or we can consider the CPU as a whole with all of its thermal sensors. While the overall CPU might accept operating temperatures of 105ºC, individual elements of the core might actually reach 125ºC instantaneously. So what is the correct value, and what is safe?
https://www.anandtech.com/show/16495/intel-rocket-lake-14nm-...
Perhaps it could downlock more to address the wattage while still coming well ahead in terms of performance, although some older CPUs doing this gave AVX512 a bad reputation.
Few workloads are as well optimized to take advantage of AVX512 as the 3d particle movement benchmark, so both the increases in performance and wattage seen in that benchmark are atypical. If they were typical, then AVX512 would be much more popular.
FWIW, I'm a big fan of wide CPU vector instructions. The real reason it was disabled is probably to push people like me to buy Saphire Rapids, which I would've favored anyway for the extra FMA unit. Although I'll also be waiting to see if Zen4 brings AVX512, which some rumors have claimed (and others have contradicted).
>Electromigration is the movement of atoms based on the flow of current through a material. If the current density is high enough, the heat dissipated within the material will repeatedly break atoms from the structure and move them. This will create both ‘vacancies’ and ‘deposits’. The vacancies can grow and eventually break circuit connections resulting in open-circuits, while the deposits can grow and eventually close circuit connections resulting in short-circuit.
Chips that get very hot are expected to be the first to show this sort of failure.
>In Black’s equation, which is used to compute the mean time to failure of metal lines, the temperature of the conductor appears in the exponent
https://www.synopsys.com/glossary/what-is-electromigration.h...
To invest Intel's money in a spinoff fab company while losing strategic control over its direction and have little to gain from its success doesn't feel all that attractive to me.
Specifically if leadership gets ossified and there's no realistic path for ambitious young people to rise into positions of power and responsibility.
When this happens, the ambitious young people are more likely to vote with the feet and build a competing organization that they will actually have power running.
This is especially problematic when it happens in government. What are young people in America supposed to do when the political conversation can't break out of asking which of near geriatric octogenarian who had a childhood with rotary phones should be our next president?
Our culture seems to have lost the very important feature that there comes a time when you've had your turn at the helm, and now you need to step down and support new leadership.
Intel has had a number of embarrassing technology misses that have put them behind: the first I remember was through-silicon vias for 3D stacking and it was only a matter of time before they missed something necessary for an advanced fab, and they missed on EUV.
Their foundry engineers at the same time were recruiting more engineers out of school on the basis that they were the "best in the world." They thought their position was unassailable because they had high yield at 22 nm, so they rested on their laurels and missed technology after technology before it bit them.
Of course all companies go through cycles of explosive growth, ossification and renewal. Small companies often die because too much of the company goes sour at the same time. In big companies this can be more spread out over time and space and there's always enough left to rebuild the leadership from the inside and push out new profitable products in time.
That being said I don't have any indication that Intel right now is in a particularly good position. Nobody in the engineering area seems to feel too honored for having worked at Intel. Yet, they got 10nm finally working and continue to rake in money. Of great concern to me is that much of that money is paid out to investors instead of reinvesting it into the company. There doesn't seem to be a convincing plan of growing the business in the future. Also they did not admit that they have a serious problem at their hands for a long time. You just can't trust anything that Intel releases; it's just a rosy projection of the future when everything goes as planned.
Unless a group of senior Ex-Intel get together and tell their war-stories we won't know what actually was going on inside.
This is somewhat true of every industry. Industrial companies in particular are always playing the game of conglomerization vs. specialization. But I agree it will definitely be interesting to see this play out in Tech in the near future
This is highly anecdotal. Apple got to where they are with M1 by vertically interesting which wouldn't be possible if they broke up like you said.
Rather, in a pipeline the bottleneck determines throughout - Intel's bottleneck is manufacturing and that's what they need to fix or cut out. We have no reason to believe their design with TSMC chips won't best AMD it even apple.
They beat AMD's benchmarks at ~double the power. I get it's winning, but it's not a fair fight. I certainly don't want my computer to be 10% faster at double the wattage.
For home use, it's going to depend on the consumer. Some people want a chip that's easy to power and cool, and still gets 80/90/95% of the performance. Some people want the absolute best performer (for their specific uses) with less regard to cooling and power.
But yes, it depends on the user.
If you care about power efficiency you can drop the power limit and still be very competitive with AMD. The only reason you don't see Ryzen CPUs drawing the same power as a maxed out 12900K is because if you pumped that much power into a Ryzen CPU it would explode (at least without LN2 being involved).
AMD hadn't released there new chips when some of that stuff was written. but its actually a little bit exciting that competition has returned and I suspect it will be good for consumers as long as both chip makers have competitive products.
If we've learned nothing else, the big winners are at the cutting edge and vertically integrated. Splitting design and fab seems like a good idea: force fab to compete for design business. But middle management will take over and create perverse incentives where neither fab nor design radically improves.
I've heard tales of Intel being rife with politics, fiefdoms and empire-building. This is classic middle management run amok in the absence of real leadership. I honestly think the best thing Intel could do is fire everyone above the director level and start over.
Intel dominated when their fab was cutting edge and no one else had access to it. Splitting this means if their fab improves then everyone has access to it.
There's clearly an organizational problem here but this split isn't going to solve it.
Judging from that it seems like their fab business doesn't have any long term future, the outside world just doesn't know it yet. Now they put on a show until they are ready to fully convert to other fabs. After all, the current fabs are still cranking out chips that make them billions each month.
It has a fine future, just not in a 60%+ margin business. Intel should bite the bullet and spin out its fabs. Let it compete and grow as a separate company from Intel’s design units. Stratchery is spot on, as he’s been about Intel from day one.
- Intel lost the mobile platform
- Apple is moving to ARM, PC laptops to follow soon.
- AMD is eating intel lunch in the high performance x86
- Servers are starting to move to ARM
Last piece of the puzzle - Intel is not doing great at the fab level.
It doesn’t look good for Intel, they need to do a radical transformation fast.
Innovate now or die.
> - Apple is moving to ARM, PC laptops to follow soon.
> - AMD is eating intel lunch in the high performance x86
That first one really needs... a source? It's a guess, but far from a certainty!
The second one is partially true? AMD has greatly increased their market share and is much more competitive with Intel than they were 5 years earlier. They are not beating Intel in market share, but they are gaining on them. They recently had leads in performance, but it's not a one-sided race.
So sure, Intel buying TSMC time helps TSMC. However, TSMC would have sold that fab time anyway, and gotten the same scale fabbing (say) AMD chips. It's much more useful for Intel to buy production and hurt the actual Intel competitors* that would have contracted for that TSMC time while still making profits for Intel.
Intel will still have enough scale to still grow its own processes, and it can probably rely on some help from Uncle Sam to further tilt the scales.
I'm frankly surprised that regulators haven't interfered - in most other markets people would have noticed the anti-competitive move (a manufacturer 'starving' their competitors' supply).
* Apple excepted, since they have more than enough cash to outbid Intel. Then again, Apple doesn't bother with servers, so Intel can live with this. Also, the Apple ecosystem is different enough that switching has its own complex set of upsides and downsides.
Intel beating TSMC in node size isn't going to happen in near term even in Intel's best case scenario, and even if that were to happen - fabs like TSMC will always have enough scale to compete. So Intel might as well buy from TSMC and take over its rival's production. Every chip Intel buys from TSMC is a chip taken from AMD and a spare chip Intel fabs could make.
Of course, AMD won't have zero capacity - even Intel doesn't want that - but I doubt AMD paid for all the time they'll ever need. For one thing this is a seller's market, if AMD had more chips people would buy. Also, AMD has a well-publicized chip shortage to the point they contracted with GlobalFoundries again (for the less important chips) about a month ago.
https://www.anandtech.com/show/15578/cloud-clash-amazon-grav...
"Cost Analysis - An x86 Massacre"
This incredible growth rate could not be achieved by hiring an exponentially-growing number of design engineers. It was fulfilled by adopting new design methodologies and by introducing innovative design automation software at every processor generation. These methodologies and tools always applied principles of raising design abstraction, becoming increasingly precise in terms of circuit and parasitic modeling while simultaneously using ever-increasing levels of hierarchy, regularity, and automatic synthesis.
As a rule, whenever a task became too painful to perform using the old methods, a new method and associated tool were conceived for solving the problem. This way, tools and design practices were evolving, always addressing the most labor-intensive task at hand. Naturally, the evolution of tools occurred bottom-up, from layout tools to circuit, logic, and architecture. Typically, at each abstraction level the verification problem was most painful, hence it was addressed first. The synthesis problem at that level was addressed much later.
Intel's own tools/process used to give them an edge over everyone else during the early years in the semis industry. Overtime, tool suppliers have catch up with Intel and Intel is falling behind due to their own tools delay/insufficient.
Intel's own tools is losing out to commodity and standardized tools
On the one hand, if this happened, it could be good for putting an end to the routine obsolescence of electronics. Then maybe we could get to the future that bunnie predicted roughly 10 years ago [1], complete with heirloom laptops.
On the other hand, ongoing advances in semiconductor technology let us solve real problems -- not just in automation that makes the rich richer, enables mass surveillance, and possibly takes away jobs, but in areas that actually make people's lives better, such as accessibility. If Moore's Law had stopped before the SoC in the iPhone 3GS had been introduced, would we have smartphones with built-in screen readers for blind people? If it had stopped before the Apple A12 in 2018, would the iOS VoiceOver screen reader be able to use on-device machine learning to provide access to apps that weren't designed to be accessible? (Edit: A9-based devices could run iOS 14, but not with this feature.) What new problems will be solved by further advances in semiconductors? I don't know, but I know I shouldn't wish for an end to those advances. I just wish we could have them without ever-increasing software bloat that leads to obsolete hardware.
Let the higher margin chip design business utilize whatever manufacturing process is most effective for its latest and greatest chip designs. Chip design and development won't be hamstrung by setbacks in manufacturing. Intel's scale will ensure most optimal pricing and deal terms with TSMC (if INTC doesn't use its own fab).
Along the way, there is additional upside optionality from government subsidies and support for a critical piece of infrastructure in the 21st century. Domestic fabs will be (if they aren't already) as high priority and critical to US national defense as the military hardware contractors (Boeing, Lockheed Martin, etc). That also ensures a floor to the price of the fab business should Gelsinger decide it should be sold.
United States domestic national security concerns make it much more financially prudent for Intel to pursue the same split. AMD took a gamble, Intel is likely to have more of a "sure thing", they have the capital backing them to make a competitive foundry business, if not the federal government will make that capital happen.
This feels like a defensive / short term value maximisation strategy - and might be the right one in terms of total value. It doesn’t feel like a long term vote of confidence.
that's the problem the article mentioned. Intel is an IDM. they developed their own tools/process and it gave Intel an edge during the early years of the semis industry.
however, it look like the industry catch up to Intel. Intel's own tools/process is no longer enough and in fact it became an technical debt.
Intel split into two allowed Intel's foundry to use standard equipment/software thus no needs to waste money on building its own tools.
Seems to me that planning to place more business with TSMC is essentially an admission that they don’t expect to be competitive with TSMC in the near future. And Gelsinger has more visibility on this than anyone.
foundry is about trust. lets say AMD want to use Intel foundry, how can you be sure the Intel's design side won't take a peek at AMD's x86 design. your fabless customers spend million on design and their design is their life blood. how they going to trust you if you also sell semis.
example: Samsung used to make chip for iphone. Apple ditched them as soon as Samsung compete with Apple w/its own smartphone.
TSMC have over 50% of the market share because they only do one thing and one thing only.
I think Apple stopping using Samsung is more related to Apple's higher-level issues: why do business with someone you accuse of violating your design patents. Not because you think they'll copy your IP, but out of principle. There's no IP-related reason Apple needed to stop buying screens from Samsung; Apple has never manufactured screens.
Apple was dual sourcing as recently as the A9 on 14nm. The reason they stopped was because Samsung fell behind the leading edge in performance and efficiency.
https://www.theregister.com/2021/11/12/apple_arm_m1_intel_x8...
> Arm's market share in PC chips was about eight per cent during Q3 this year, climbing steadily from seven per cent in Q2, and up from only two per cent in Q3 2020, before Arm-compatible M1 Macs went on sale.
ARM is very much on the rise, but also still in the single digits, leaving plenty for x86.
https://futurecio.tech/intel-losing-share-to-amd-and-arm/
> 5% of the servers shipped in the third quarter of 2021 had an Arm CPU
1. Lobby the US Government and public to increase public spending on manufacturing semiconductors in the US
2. Corner most of those subsidies/funding
3. Done
Only partially joking; Why has Intel not made the over reliance of overseas fabs a natsec issue?
What in the world is there to be "bearish" about?
But at this point when there putting chips in everything its hard not to see the market expanding to meet increased capacity. (I think some greating cards have chips in them to control music/lights....)
This article is rambling and over-complicated.
For a much more insightful and compelling view into Intel at its greatest I recommend “The Pentium Chronicles”.
The risk with this is that Intel will become a fab for older, less expensive chips, while TSMC and others gobble up the top tier stuff.
But the world needs more, not less, fab capacity, and there are plenty of people would gladly use Intel for their needs if the opportunity was there.
[0] https://twitter.com/IanCutress/status/1481581119740989440/ph...
It’s not like there is an excess of capacity right now, and there are only a small number of players that even offer what intel has available.
We've got a very serious shortage of fab capacity right now that will probably last for at least 2 or 3 more years. The answer would be anyone who needs fab capacity of the sort Intel can provide (not likely AMD of course, though I could see Apple giving it a try).
there is no loyalty. Apple will ditch you if they can get cheaper price or you are competing with them like Samsung.
If you want an actual ultra portable, you are getting a Mac.
The Max has barely any improvement core-wise than the first M1. It’s going to be interesting to see what the real next generation looks like.
Isn't that what you would expect in two chips that use the same core design?
There is more memory bandwidth to the Max, and the system level cache is larger, so there are differences outside of the core, but the core itself didn't change.
That's not what you would expect if you look at the graph of six years of year-over-year SPEC performance gains on iPhone cores. Their history shows a pretty reliable 20% gain per year.
>Whilst in the past 5 years Intel has managed to increase their best single-thread performance by about 28%, Apple has managed to improve their designs by 198%, or 2.98x (let’s call it 3x) the performance of the Apple A9 of late 2015.
https://www.anandtech.com/show/16226/apple-silicon-m1-a14-de...
I'm sure people said the same thing about Intel when AMD introduced 64 bit procs or Sun introduced multicore processors. Intel has adapted and lead the field many times after being overtaken. No reason to expect them not to do the same here, or at least compete.
[1] https://armkeil.blob.core.windows.net/developer/Files/pdf/wh...
[2] https://www2.eecs.berkeley.edu/Pubs/TechRpts/2006/EECS-2006-...
Alder Lake is about dealing with power and thermal constraints. Intel has finally pushed performance so far that they need to do this. The chips are benchmarking so well because of this move, not in spite of it.
Of course Alder Lake is benchmarking well because of it.
My point was ‘already’ makes it sound like Intel is rapidly adopting this technology - not 11 years after Arm.
That's probably an exaggeration, and this exaltation of Apple is a Bad Thing. The M1 had good designers - two of which have moved on already, one to Microsoft [0], and the lead designer to Intel [1] only days later.
[0] https://9to5mac.com/2022/01/12/apple-silicon-microsoft-poach...
[1] https://9to5mac.com/2022/01/06/apple-engineering-director-m1...
Still the real blow will be if someone successfully enters the server business. Graviton* are interesting but that’s not a broad threat yet.