AMD fans rejoice, Intel has their own Hector Ruiz

[CHADBOGA]

Intel has really been directionless and drifting ever since they appointed Brian Krzanich as their CEO(maybe even a bit before that too) and things look like they are only getting worse as time goes by.

To be fair to Krzanich, Intel did make one of the most idiotic large acquisitions I have ever seen a semi-conductor company make when they purchased McAfee for $7.7B, before Krzanich became CEO, and sure enough that proved to be a financial disaster.

But now Krzanich is making his mark by spending extraordinary amounts on acquisitions(with very uncertain Returns on Investment), first there was the Altera purchase for $16.7B and the subsequent slow, slow role out of Altera products and now Intel look poised to spend perhaps as much as $16B to acquire Mobileye, a Jerusalem-based developer of advanced vision and driver assistance systems.

Now in the context of the above, consider that during the period between the McAfee purchase and the Altera purchase, Nvidia was valued at less than $20B, maybe even as low as $10B.

So by what process did Intel go through to reach the decision that Nvidia wasn't worth buying, but all these other companies are? There seems to be something fundamentally broken within Intel management.

All this also comes after Intel's spectacular failure in mobile for phones and tablets, with billions squandered and who yet knows what price they may ultimately pay for not being a player here.

Intel's much vaunted manufacturing process is stumbling and coming back to the field, and whilst Ryzen probably won't do enough damage to inflict a major wound on Intel in the next 6 months, if I was an Intel investor, I would be very concerned about what Ryzen II delivers, as I suspect it will fix up Ryzen's slightly erratic performance on certain games, come with a much more stable platform and make it a compelling choice over just about all of Intel's offerings.

On so many fronts, Intel is finding ways to fail or at best, not succeed.

I would say that people should not be looking at Intel's financials right now, as Intel has been able to delay the day of reckoning so far, but where they are likely to be in 2 years time, 5 years time, 10 years time.

There is a quote often attributed to Mark Twain which goes something like "History never repeats itself but it rhymes", well Intel has the stench of 2003 about it for me and I am expecting some major shake ups in the semi-conductor landscape over the next 5 years and none of it is going to be in Intel's favour.

 

[Batmeat]

Interesting read. My opinion is Intel will be fine though for years all because they dominate the cpu marketplace. Bad decisions by them? Yes, of course, but what company isn't plagued by bad decisions made along their path of growth and development?

 

[DaveSimmons]

Pre-Trump, Intel would not have been allowed to buy nvidia because of antitrust concerns.

Now that the EPA, FCC, and all other corporate oversight is being "rightsized" intel should seize their chance.

 

[raghu78]

I have a certain extreme view. I think the Intel we see today is probably the strongest ever we will see in terms of revenue or profit. Slowly and surely Intel will see its revenues and profits erode. I see multiple reasons.

1. Loss of Process Manufacturing leadership. Intel has ruled the semiconductor industry for the past 3 decades with a process node leadership of anywhere from 1 to 2 generations or 2-4 years. But thats slowly and surely coming to an end. Just to highlight what has happened between 2012 and whats happening between now and 2018.
TSMC 28nm high k- Jan 2012
Intel 22nm FINFET- Apr 2012
TSMC 20nm high k - Sep 2014
Intel 14nm FINFET - Dec 2014
TSMC 16FF+ FINFET - Sep 2015
TSMC 10nm FINFET - Q2 2017
Intel 10nm FINFET - late 2017 / early 2018.
TSMC 7nm FINFET- H1 2018.

TSMC 7nm and Intel 10nm are pretty much similar on metal pitch M1 and contacted gate pitch. TSMC even has a high performance flavour of their 7nm process for CPUs optimized for 4 Ghz operation. Surely TSMC and the ARM ecosystem are going for Intel's cash cow - the lucrative server market. How did Intel lose their multi generational process node leadership. Such things do not happen overnight. Something is fundamentally broken at Intel to have let this happen. This is not Andy Grove's " Only the paranoid survive" Intel.

2. ARM - ARM will do to Intel what Intel did to RISC/Unix. Intel used the massive volumes from the PC business to fund their CPU R&D and their process node R&D and building a state of the art fab network . They used that to build competitive server CPUs/platforms which were much cheaper than the RISC/UNIX counterparts and in partnership with MS went about dismantling the UNIX/RISC domination of server markets in the 1990s. Today thats exactly what ARM and the ecosystem including Apple is doing. The massive volumes from the growing smartphone and shrinking tablet business keeps funneling ever more money into the foundries like TSMC and the direct result is we see TSMC catching up to Intel in bleeding edge process nodes. This leads to the ARM ecosystem benefiting as they are moving beyond phones into servers with the help of TSMC. Qualcomm is directly going after Intel's markets and with TSMC process they have a very good opportunity to make significant inroads in few segments of the server market. Microsoft are throwing their support behind ARM and thats the tipping point imo for the march of ARM into servers.

3. Resurgent AMD. Intel has not seen significant competition for well over a decade in the CPU market. AMD has made a very good comeback to start off a multi year recovery in servers and high performance computing. AMD is going to take market share in all areas - desktop, notebook and server. But I think its the server market which will give AMD the much needed profits it needs to fund future generations of CPUs/GPUs/APUs.

I think over the next 5 years we will see how Intel copes with the multiple threats its faced with. I believe Intel has lost its process leadership once and for all. Apple and TSMC are going to keep driving process nodes at a furious pace which Intel is unlikely to keep up with. I do believe eventually ARM will become the dominant ISA in the computing world across all segments, but it will probably take a decade atleast. Intel will have to evolve to avoid becoming a DEC.

 

[frozentundra123456]

Intel has really been directionless and drifting ever since they appointed Brian Krzanich as their CEO(maybe even a bit before that too) and things look like they are only getting worse as time goes by.

To be fair to Krzanich, Intel did make one of the most idiotic large acquisitions I have ever seen a semi-conductor company make when they purchased McAfee for $7.7B, before Krzanich became CEO, and sure enough that proved to be a financial disaster.

But now Krzanich is making his mark by spending extraordinary amounts on acquisitions(with very uncertain Returns on Investment), first there was the Altera purchase for $16.7B and the subsequent slow, slow role out of Altera products and now Intel look poised to spend perhaps as much as $16B to acquire Mobileye, a Jerusalem-based developer of advanced vision and driver assistance systems.

Now in the context of the above, consider that during the period between the McAfee purchase and the Altera purchase, Nvidia was valued at less than $20B, maybe even as low as $10B.

So by what process did Intel go through to reach the decision that Nvidia wasn't worth buying, but all these other companies are? There seems to be something fundamentally broken within Intel management.

All this also comes after Intel's spectacular failure in mobile for phones and tablets, with billions squandered and who yet knows what price they may ultimately pay for not being a player here.

Intel's much vaunted manufacturing process is stumbling and coming back to the field, and whilst Ryzen probably won't do enough damage to inflict a major wound on Intel in the next 6 months, if I was an Intel investor, I would be very concerned about what Ryzen II delivers, as I suspect it will fix up Ryzen's slightly erratic performance on certain games, come with a much more stable platform and make it a compelling choice over just about all of Intel's offerings.

On so many fronts, Intel is finding ways to fail or at best, not succeed.

I would say that people should not be looking at Intel's financials right now, as Intel has been able to delay the day of reckoning so far, but where they are likely to be in 2 years time, 5 years time, 10 years time.

There is a quote often attributed to Mark Twain which goes something like "History never repeats itself but it rhymes", well Intel has the stench of 2003 about it for me and I am expecting some major shake ups in the semi-conductor landscape over the next 5 years and none of it is going to be in Intel's favour.

Strange how "AMD is doomed" has suddenly turned to "intel is doomed" and we havent even seen a market impact of Zen yet. Everybody is on the AMD bandwagon now, as evidenced by the amazing run up in their stock, but long term, I see both in danger. Both companies are pretty much dependent on X86 continuing to be dominant, and have almost zero presence in mobile, i.e. phones, smartwatches, etc. AMD has the consoles and dgpus to turn to, but dgpus are a mature if not declining market, and AMD has not executed well there either, as evidenced by nearly a year without top tier competitors to nVidia. I also see a still distinct possibility of the consoles eventually turning to ARM, thus being on the same platform as the mobile gaming market. So AMD fans (not you) better not take too much glee in intel's problems, because long term I think both still face a serious threat from Apple/ARM.

That said, I do agree with you regarding intel's acquisitions and attempts to break into new markets. I dont really know how a company so good at designing, building, and marketing x86 cpus can make such poor business decisions and fail so badly from a technical standpoint in other areas. Mobile was clearly a disaster, and I dont have a clue what they were thinking when they acquired McAfee. I also think the chances are very slim they are going to be a success in IoT or the automobile/self driving cars either. Just look at the landscape now. We have tons of smart watches, fitness trackers, smart TVs, etc, and not a sign of intel. We also have Android auto and apple car play, and again intel is late/poorly equipped for the game. Eerily similar to mobile actually.

I do think "their own Hector Ruiz" is a bit melodramatic though. At least they havent yet sold off thier fabs and negotiated a crushing WSA to buy products from them.

 

[VirtualLarry]

Well, Intel may or may not be "doomed", but I feel that their golden years are behind them, the writing is on the wall. Without an Intel-leading / exclusive breakthrough in device physics R&D, to invent an entirely new form of transistor / semi-conductor (metal hydrogen, if superconducting?), then their most pressing competitors seem to be catching up to their process manufacturing lead, or at the very least, making it far less relevant as a decisive factor in the semiconductor market, and the competing architecture(s) seem to have a licensing advantage, compared to Intel. Maybe if they weren't so stingy about giving out x86/x64 licenses, then maybe the overall x86 market would be healthier, and in more devices. ARM is at the gate, and they're slowly beating it down. It seems, by unit volume, x86 is outnumbered by ARM, if not sales / revenue volume.

 

[CHADBOGA]

Strange how "AMD is doomed" has suddenly turned to "intel is doomed" and we havent even seen a market impact of Zen yet.

I'm not saying Intel is doomed, but I think they are poised to take a big hit on multiple fronts.

And there is no quick fix, as even by the time the Intel Board of Directors is forced to act, much will have been baked in and if the new guy is any good, it will take maybe 4 or 5 years to right the ship.

Raghu78 may well be proven right when he say "I have a certain extreme view. I think the Intel we see today is probably the strongest ever we will see in terms of revenue or profit. Slowly and surely Intel will see its revenues and profits erode."

 

[frozentundra123456]

I'm not saying Intel is doomed, but I think they are poised to take a big hit on multiple fronts.

And there is no quick fix, as even by the time the Intel Board of Directors is forced to act, much will have been baked in and if the new guy is any good, it will take maybe 4 or 5 years to right the ship.

Raghu78 may well be proven right when he say "I have a certain extreme view. I think the Intel we see today is probably the strongest ever we will see in terms of revenue or profit. Slowly and surely Intel will see its revenues and profits erode."

I know you did not say that. It is already starting to show up in other threads however, espoused by the usual suspects who were the most outraged when the same thing was said about their favorite company. Doesnt happen often, but I probably would even agree with Raghu that intel will be on a generally downward trend, especially in margins. And like I said, AMD seems to be on a roll now, but long term they are facing the same dangers as intel. They just seem to be on a roll now because they are coming from a very low place and improvement (both technical and financial) are easy.

 

[CHADBOGA]

I know you did not say that. It is already starting to show up in other threads however, espoused by the usual suspects who were the most outraged when the same thing was said about their favorite company. Doesnt happen often, but I probably would even agree with Raghu that intel will be on a generally downward trend, especially in margins. And like I said, AMD seems to be on a roll now, but long term they are facing the same dangers as intel. They just seem to be on a roll now because they are coming from a very low place and improvement (both technical and financial) are easy.

Yes I forgot to address AMD in my previous post.

I see AMD as a large thorn in Intel's side in the short and medium term, which will make it harder for Intel to do what it needs to, for the longer term, but yeah, ARM are the big danger in the long term.

 

[beginner99]

McAfee for sure was idiotic and pointless. However you can only buy a company if it is for sale and NV is not for sale at least not at the figures you mentioned.

I can the the reasoning behind the purchase as self-driving cars will need lots of processing power and potentially also cloud/server infrastructure for example for intercommunication between cars like information about traffic density so that a different route is faster than the default route. Intel doesn't want to miss the next big boat like they did with smartphone market. Why could ARM (Apple) and TSMC catch up to intel? Because of the huge profits from smartphone market. Intel doesn't want to miss another such boat. They can't. They need to sell huge amounts of chips or building new fabs becomes a losing business.

 

[tamz_msc]

We don't even need to look at this and previous acquisitions to see where Intel is headed under Brian Krzanich. Instead we should have a look at a recent development, which has seemed to gone unnoticed:

Intel has dropped sponsorship of Science Fairs for High School students.

The Intel decision provoked a sharp difference of opinion between Brian Krzanich, Intel’s current chief executive, and Craig R. Barrett, a former Intel chairman and chief executive.

Mr. Krzanich has told colleagues privately that the science fairs were the fairs of the past and had become tilted to life sciences and biotechnology, not primary fields for Intel, according to two people who are not authorized to speak publicly for the company.

Mr. Barrett disagreed. In an email, he said, “you might instead conclude that Intel is a company of the past, just like Westinghouse when they dropped” sponsorship of the national science fair in 1998.

Mr. Barrett, who is on the board of the Society for Science, also said that all of science has become data-driven and computational, so Intel has a stake in nurturing youthful innovators in all scientific disciplines, including the life sciences.

 

[VirtualLarry]

Interesting about the Science Fair stuff, I did not know that.

"Company of the past" indeed.

 

[coercitiv]

I know you did not say that. It is already starting to show up in other threads however, espoused by the usual suspects who were the most outraged when the same thing was said about their favorite company. Doesnt happen often, but I probably would even agree with Raghu that intel will be on a generally downward trend, especially in margins. And like I said, AMD seems to be on a roll now, but long term they are facing the same dangers as intel. They just seem to be on a roll now because they are coming from a very low place and improvement (both technical and financial) are easy.

We still have voices from both barricades spamming the forums with information that resembles propaganda. That should not stop the rest of the forum from having a decent discussion on hot topics, even if we have to ignore or occasionally confront skewed points of view.

 

[JimmiG]

I think over the next 5 years we will see how Intel copes with the multiple threats its faced with. I believe Intel has lost its process leadership once and for all. Apple and TSMC are going to keep driving process nodes at a furious pace which Intel is unlikely to keep up with. I do believe eventually ARM will become the dominant ISA in the computing world across all segments, but it will probably take a decade atleast. Intel will have to evolve to avoid becoming a DEC.

How long are TSMC and Apple going to be able to keep shrinking the process node though? From what I understand, we're very close to hitting a brick wall in terms of process size. Sounds to me like this would serve as a great equalizer. Whoever reaches 5nm (or whatever the limit turns out to be) first gets a small advantage, but with no further progress, everyone eventually catches up.

 

[VirtualLarry]

Whoever reaches 5nm (or whatever the limit turns out to be) first gets a small advantage, but with no further progress, everyone eventually catches up.

That's how I see it too, and what I alluded to by suggesting that Intel needs a fundamental semi-conductor R&D breakthrough to maintain their industry lead.

 

[railroadmaster]

http://www.businessinsider.com/intel-is-probably-the-worst-aquirer-in-tech-history-2010-9

 

[Valantar]

Intel used the massive volumes from the PC business to fund their CPU R&D and their process node R&D and building a state of the art fab network . They used that to build competitive server CPUs/platforms which were much cheaper than the RISC/UNIX counterparts and in partnership with MS went about dismantling the UNIX/RISC domination of server markets in the 1990s. Today thats exactly what ARM and the ecosystem including Apple is doing. The massive volumes from the growing smartphone and shrinking tablet business keeps funneling ever more money into the foundries like TSMC and the direct result is we see TSMC catching up to Intel in bleeding edge process nodes.

I think you're partially right, but there is one fundamental difference: Intel timed their push into the fab business perfectly - on a wave of increasing PC and server sales while building fabs was still "cheap." Today, fab and process development costs have risen massively, while the smartphone market is already largely leveling out/morphing into a commodity/luxury two-segment market with no real growth (just see how mid-range phone sales are taking off, while flagship models see ever-increasing prices). Tablet sales have plummeted. And while ARM is making huge money on the IoT push, i doubt that'll last for long either (after all, nobody is going to replace their "smart" stove or coffee maker every two years).

While ARM is seeing massive growth, they've had several false starts in the server market (with cores simply not suited for the task, mind you, which may not be true any longer), and it's reasonable to expect their income to level off or even dip in coming years. Which makes the time to truly invest in the server market now, as it's likely won't ever have more money to invest at once than right now. If the current push fails to gain traction, it's reasonable to think that ARM servers will stay dead until we see some significant breakthrough. Still, the server and datacenter market is more than big enough for X86 and ARM to coexist. This will, of course, entail big losses for Intel, especially if AMD returns to this market in a noticeable way.

Still, I think what we're seeing is just an extension of what we've seen in the PC market for years. Intel has grown complacent, with a huge process and architecture lead, owning pretty much the entire server and datacenter market during its biggest growth period ever. This has caused them to ... perhaps not take this for granted, but definitely "go the safe route" in terms of process and arch development. Slow and steady wins the race-mentality. "Suddenly", we're faced with process improvement difficulties (and increased costs), which understandably hits the most advanced players first, which means competitors start catching up. To boot, AMD comes pretty much out of nowhere with an architecture that's competitive in IPC while arguably beating Intel for efficiency - for the first time in a decade or more. While I don't think server CPU sales are dropping off any time soon, I don't think the total amount of chips sold will increase dramatically either (unless those chips are smaller and more efficient in large multi-socket setups), I think we'll be seeing something close to the "commoditization" of the comsumer PC market enter the server market, with cheaper options gaining market share and competition tightening margins.

 

[hansolo69]

Funny thread, but what about http://www.anandtech.com/show/10797/qualcomm-to-acquire-nxp ?
Qualcomm and NXP Semiconductors on Thursday announced that they had signed an agreement, under which Qualcomm will acquire NXP. The boards of both companies have already unanimously approved the all-cash deal representing a total enterprise value of about $47 billion.
What is interesting is that Qualcomm intends to buy NXP in an all-cash deal, which means that the company will have to borrow money, as it has around $17 billion in cash and other short-term assets its pockets today (according to U.S. regulators).

Both Intel and Qualcomm are trying to diversify its businesses, both aquisitions will not have a ROI for several years.
Or maybe is some creative accounting involved in which both companies will have lower taxes to pay. It was done before by many big companies.

 

[NTMBK]

Intel's current business is built on two main pillars- the PC market, and the server market. The PC market is collapsing (and has been for years), and Intel have already put it into "managed decline" mode by reducing investment in it:

They're desperately trying to keep their edge in the datacenter, and flailing around looking for a new market with fab-filling volumes to replace the evaporating PC market. They're jumping on every bandwagon going. They blew billions on a failed mobile push, they wasted money and effort on the tablet bubble, they're pissing around with x86 in embedded in an attempt to get into IoT, and now they are spending billions on the latest fad (automotive).

Meanwhile their long run of being unchallenged in the datacenter is over. AMD finally has a chip which can compete in servers again, and ARM vendors like Qualcomm are coming for them big time. If rivals start chipping away at this second pillar, things are going to start looking very ugly for them, fast.

 

[Glo.]

This is biggest problem in long term, that Intel may not be investing in key markets(which still is, and will be for foreseeable future - x86). Just right now, we have again competition back to this market. Lets hope, that Krzanich either will get enlightenment, or will be phased out from Intel, and they will get their wind back with technology.

 

[USER8000]

I have a certain extreme view. I think the Intel we see today is probably the strongest ever we will see in terms of revenue or profit. Slowly and surely Intel will see its revenues and profits erode. I see multiple reasons.

1. Loss of Process Manufacturing leadership. Intel has ruled the semiconductor industry for the past 3 decades with a process node leadership of anywhere from 1 to 2 generations or 2-4 years. But thats slowly and surely coming to an end. Just to highlight what has happened between 2012 and whats happening between now and 2018.

I might be remembering this wrong,but when AMD still had their own fab,I thought they actually beat Intel to one node??

 

[hansolo69]

Intel's current business is built on two main pillars- the PC market, and the server market. The PC market is collapsing (and has been for years), and Intel have already put it into "managed decline" mode by reducing investment in it:

They're desperately trying to keep their edge in the datacenter, and flailing around looking for a new market with fab-filling volumes to replace the evaporating PC market. They're jumping on every bandwagon going. They blew billions on a failed mobile push, they wasted money and effort on the tablet bubble, they're pissing around with x86 in embedded in an attempt to get into IoT, and now they are spending billions on the latest fad (automotive).

Meanwhile their long run of being unchallenged in the datacenter is over. AMD finally has a chip which can compete in servers again, and ARM vendors like Qualcomm are coming for them big time. If rivals start chipping away at this second pillar, things are going to start looking very ugly for them, fast.

I saw this repeating for years at sites like SA:
- PC are doomed
- Nvidia is doomed, has broken GPUs and soon will be financially bankrupt
- ARM servers are cheaper and better than x86, so Intel is doomed
- ultrabooks and x86 tablets are stupid and will not sell well

If I look at Intel's financial reports none of this is real.
Now Intels has a good AMD competitor and AMD has the technical knowledge to compete Intel in servers segment (and I hope they will do well). ARM... not yet, just unwarranted hopes from fans and some coercitive tactics from Microsoft to lower CPU prices (the same did with WARM tablet).
Automotive and selfdriving cars are the next boom in technology, Nvidia already has a good track record and Intel wants to be a player in that market. However from their estimation ,,the vehicle systems, data and services market opportunity to be up to $70 billion by 2030", so it's a little to far and a little to small market opportunity regarding current $15.3 billion aquisition.
But Qualcomm aquisition is even more senseless, same the $31 billion for ARM.
Again, mergers and aquisitions may have some creative accounting behind them.

 

[NTMBK]

I saw this repeating for years at sites like SA:
- PC are doomed
- Nvidia is doomed, has broken GPUs and soon will be financially bankrupt
- ARM servers are cheaper and better than x86, so Intel is doomed
- ultrabooks and x86 tablets are stupid and will not sell well

If I look at Intel's financial reports none of this is real.
Now Intels has a good AMD competitor and AMD has the technical knowledge to compete Intel in servers segment (and I hope they will do well). ARM... not yet, just unwarranted hopes from fans and some coercitive tactics from Microsoft to lower CPU prices (the same did with WARM tablet).
Automotive and selfdriving cars are the next boom in technology, Nvidia already has a good track record and Intel wants to be a player in that market. However from their estimation ,,the vehicle systems, data and services market opportunity to be up to $70 billion by 2030", so it's a little to far and a little to small market opportunity regarding current $15.3 billion aquisition.
But Qualcomm aquisition is even more senseless, same the $31 billion for ARM.
Again, mergers and aquisitions may have some creative accounting behind them.

And yet, PC sales continue to fall:

Guess what? Ultrabooks and x86 tablets/2-in-1s have completely failed to reverse that trend. (2016, missing from that chart, was yet another decline.)

While the foundries have now caught up with Intel in process technology. Why do you think Intel has announced that they will go "server first" with future nodes? Because their competition has caught up. ARM server chips are finally appearing on nodes competitive with Intel (10nm Qualcomm server parts), with high performance custom cores capable of competing instead of just repurposed 32-bit phone CPUs.

Intel's financials have continued to look good because, as I said earlier, they haven't had any serious competition in the datacentre for years and years. They blew AMD out of the market with Nehalem and then AMD ran over their own foot with a Bulldozer, the RISC players (SPARC, POWER) were relegated to high end niches which were getting slowly eaten up by cheaper x86 machines... They had a free reign to crank out healthy profits, and slowly push up the prices. Well the fun times are over.

 

[TemjinGold]

I saw this repeating for years at sites like SA:
- PC are doomed
- Nvidia is doomed, has broken GPUs and soon will be financially bankrupt
- ARM servers are cheaper and better than x86, so Intel is doomed
- ultrabooks and x86 tablets are stupid and will not sell well

If I look at Intel's financial reports none of this is real.
Now Intels has a good AMD competitor and AMD has the technical knowledge to compete Intel in servers segment (and I hope they will do well). ARM... not yet, just unwarranted hopes from fans and some coercitive tactics from Microsoft to lower CPU prices (the same did with WARM tablet).
Automotive and selfdriving cars are the next boom in technology, Nvidia already has a good track record and Intel wants to be a player in that market. However from their estimation ,,the vehicle systems, data and services market opportunity to be up to $70 billion by 2030", so it's a little to far and a little to small market opportunity regarding current $15.3 billion aquisition.
But Qualcomm aquisition is even more senseless, same the $31 billion for ARM.
Again, mergers and aquisitions may have some creative accounting behind them.

The problem is financial statements aren't forward looking. They can tell you how dominant Intel was in the past (which no one is arguing against) but they can't tell you what Intel faces going forward.

 

[Valantar]

Guess what? Ultrabooks and x86 tablets/2-in-1s have completely failed to reverse that trend. (2016, missing from that chart, was yet another decline.)

There is no way this wouldn't have happened. Why? Because the growth hinged on one factor: people who didn't yet have PCs. The only market now (outside of kids growing old enough to want/need their own PC) is current PC owners replacing old ones. Which doesn't happen annually or even biannualy. Sales are bound to drop, and significantly - especially when there's little to no performance incentive to upgrade. That smartphones are becoming powerful enough to fill this role for some people spells even more trouble for Intel, although saying that phones will replace PCs wholesale is naive at best.

The same thing is also starting to happen with smartphones, the main difference being that they have some market potential in "emerging markets" due to relatively low cost. Everyone and their dog in western (and many non-western) countries has one. The drop-off in replacement/upgrade rates is already happening (from 1y2m on average here in Norway at its high point a few years ago to nearly 3y now). Again, most phones for the last two years or so are good enough that the average user sees little reason to shell out $600+ annually for a replacement for something that still works.

I believe both smartphones and PCs will level off at lower sales/production levels than today, but they haven't yet. As such, chipmakers are scrambling to fill unused fabs and trying to recoup their glory days of five years ago. I don't think that's going to happen, but this again requires serious restructuring of their entire business model - which they're not exactly happy about. Thus, flailing attempts at diversification.

Chances are the silicon fab business will look very different from now in 10-15 years.