28 ms·

Has Intel lost its mojo?

29 points by pieterr 6y ago

_ph_ 6y ago
The success of Intel very largely depended on one single thing: volume. By owning most of the PC market, Intel could spend a lot of money into R&D. Which affected both chip design resources as of course manufacturing. For many years, Intel used to have at least one step advantage in the manufacturing process, often more. This is how Intel basically killed all non x86 based processor designs. Most of which were considered technically better, but on inferior processes lost the competition with Intel (I am thinking about all the famous RISC-architectures). AMD could leapfrog Intel from time to time with a really clever design, but struggled to keep up in manufacturing. This is where the Zen2 gets so dangerous to Intel, a very good design produced on the better manufacturing process. Where did that process come from? That leads to the decisive point to where Intel is today: the emergence of the smartphone. The impact on Intel is twofold: first of all, the PC market is shrinking as the smartphones are the computing device many people use, PCs have lost their importance for many, even if people still have a PC, they don't get updated as often any more. And secondly: there is a new huge market for semiconductors. Which Intel managed to not be a part of. The volume of the mobile marked helped TSMC to get to the point, where their processes eclipsed Intel. This benefits all the competitors of Intel to compete not only on even ground, but on higher ground. Suddenly, non-x86 architectures, mainly ARM become economically feasible. When Apple designs the Apple Silicon, they can afford to throw enormeous R&D resources at any problem. Not only because Apple is a rich company, but also, because the spending is spread across over 200 million iPhones per year. Not counting iPads and soon Macs. This is, why I am so excited about Apple Silicon. For the first time since the passing of the classical high-performance RISC vendors, a non-x86 architecture enters the desktop market, this time backed by huge R&D spending and a superior process. Is all lost for Intel? Certainly not, Intel still owns the largest part of the PC market. While the problems with their 10nm process were huge, they are not too far behind TSMC, as the 10nm Intel process isn't that much behind the TSMC 7nm as the numbers would suggest. But their moat consisting of volume and process lead has been breached, so success is no longer automatic and Intel needs to work hard going forward.
- MangoCoffee 6y ago
  It might be better for Intel to split into two. One part focus on design and the other on pure-play manufacturing. AMD and TSMC shown us that it work.
  - minipci1321 6y ago
    > It might be better for Intel to split into two. How exactly would it be better -- again, given the volumes? -- The "design" split-off would be instantly lacking modern-node fab capacities (TSMC already said they wont scale up for Intel), -- the manufacturing part would find itself in competition with those already having better node tech ==> not enough orders to fill the pipe (customers with comparable volumes are already all set WRT fabbing).
    - LargoLasskhyfv 6y ago
      Smaller orgs. Leaner, meaner, faster.
artemonster 6y ago
„... in which he blames a management culture that favors MBAs over technical expertise for making bad technical decisions over the past few years.“ Why do we keep hearing this over and over again for major technology driven companies and why they keep doing this?
- rapsey 6y ago
  Because MBAs run the show.
- chadcmulligan 6y ago
  I've always thought its because MBA's are experts at inserting themselves in the process, and grabbing all the cash. As an example, most charities have multiple levels of management now, where they used to be run by unpaid volunteers. Why? because MBA's convinced everyone they needed professional management.
- DrBazza 6y ago
  "Investor value". It's a symptom of western economies, and the US in particular. It's all about the share price and doing whatever you can to boost it, and then pay dividends (if the share price isn't doing well). Companies are now seen as financial instruments. If you buy shares now is your first thought about the company, or about share price and how much money you can make? That's it really. MBAs know how to inflate share price and 'add value' in the modern world. And when the CEO is no longer an engineer, he chases the money, because he's usually paid in shares and options.
  - jmnicolas 6y ago
    > "Investor value". It's a symptom of western economies, and the US in particular. I absolutely agree. And pushed to extremes it makes for sociopathic reasoning where nothing else matters. We humans have a hard time finding the middle way : no I don't want communism but I don't want full blown predatory capitalism either.
    - DrBazza 6y ago
p1necone 6y ago
Intel has just ridden on marketing and anti competitive maneuvering for so long that their actual engineering side has gotten complacent/suffocated. I'm sure they could be (and probably are on their way to being) back as a serious player with management that ceases the bullshit.
jokoon 6y ago
Spectre and meltdown are key events that explain why intel is in this situation right now.
- Jweb_Guru 6y ago
  The key thing that caused Intel to be in this situation right now was that their EUV play totally failed. If it had succeeded when they hoped it would, they would be so far ahead of the competition that none of this would matter.
  - throwaway2048 6y ago
    intel 10nm was never an EUV play.
    - Jweb_Guru 6y ago
      I don't recall saying a single thing about 10nm.
      - throwaway2048 6y ago
        There was never an EUV play they were making that would have catapulted them far ahead at this point.
        Jweb_Guru 6y ago
        Prior to 2015, Intel believed they would have production ready, 7nm chips by 2018 using EUV, which would indeed have catapulted them far ahead of the competition. Obviously, many people were skeptical about this, but if it had worked out I don't think we'd be having this conversation, nor would people have cared about them missing the 10nm target.

29 points by pieterr 6y ago

_ph_ 6y ago

The success of Intel very largely depended on one single thing: volume. By owning most of the PC market, Intel could spend a lot of money into R&D. Which affected both chip design resources as of course manufacturing. For many years, Intel used to have at least one step advantage in the manufacturing process, often more. This is how Intel basically killed all non x86 based processor designs. Most of which were considered technically better, but on inferior processes lost the competition with Intel (I am thinking about all the famous RISC-architectures). AMD could leapfrog Intel from time to time with a really clever design, but struggled to keep up in manufacturing. This is where the Zen2 gets so dangerous to Intel, a very good design produced on the better manufacturing process. Where did that process come from? That leads to the decisive point to where Intel is today: the emergence of the smartphone. The impact on Intel is twofold: first of all, the PC market is shrinking as the smartphones are the computing device many people use, PCs have lost their importance for many, even if people still have a PC, they don't get updated as often any more. And secondly: there is a new huge market for semiconductors. Which Intel managed to not be a part of. The volume of the mobile marked helped TSMC to get to the point, where their processes eclipsed Intel. This benefits all the competitors of Intel to compete not only on even ground, but on higher ground. Suddenly, non-x86 architectures, mainly ARM become economically feasible. When Apple designs the Apple Silicon, they can afford to throw enormeous R&D resources at any problem. Not only because Apple is a rich company, but also, because the spending is spread across over 200 million iPhones per year. Not counting iPads and soon Macs. This is, why I am so excited about Apple Silicon. For the first time since the passing of the classical high-performance RISC vendors, a non-x86 architecture enters the desktop market, this time backed by huge R&D spending and a superior process. Is all lost for Intel? Certainly not, Intel still owns the largest part of the PC market. While the problems with their 10nm process were huge, they are not too far behind TSMC, as the 10nm Intel process isn't that much behind the TSMC 7nm as the numbers would suggest. But their moat consisting of volume and process lead has been breached, so success is no longer automatic and Intel needs to work hard going forward.

MangoCoffee 6y ago
It might be better for Intel to split into two. One part focus on design and the other on pure-play manufacturing. AMD and TSMC shown us that it work.
- minipci1321 6y ago
  > It might be better for Intel to split into two. How exactly would it be better -- again, given the volumes? -- The "design" split-off would be instantly lacking modern-node fab capacities (TSMC already said they wont scale up for Intel), -- the manufacturing part would find itself in competition with those already having better node tech ==> not enough orders to fill the pipe (customers with comparable volumes are already all set WRT fabbing).
  - LargoLasskhyfv 6y ago
    Smaller orgs. Leaner, meaner, faster.

artemonster 6y ago

„... in which he blames a management culture that favors MBAs over technical expertise for making bad technical decisions over the past few years.“ Why do we keep hearing this over and over again for major technology driven companies and why they keep doing this?

rapsey 6y ago
Because MBAs run the show.
chadcmulligan 6y ago
I've always thought its because MBA's are experts at inserting themselves in the process, and grabbing all the cash. As an example, most charities have multiple levels of management now, where they used to be run by unpaid volunteers. Why? because MBA's convinced everyone they needed professional management.
DrBazza 6y ago
"Investor value". It's a symptom of western economies, and the US in particular. It's all about the share price and doing whatever you can to boost it, and then pay dividends (if the share price isn't doing well). Companies are now seen as financial instruments. If you buy shares now is your first thought about the company, or about share price and how much money you can make? That's it really. MBAs know how to inflate share price and 'add value' in the modern world. And when the CEO is no longer an engineer, he chases the money, because he's usually paid in shares and options.
- jmnicolas 6y ago
  > "Investor value". It's a symptom of western economies, and the US in particular. I absolutely agree. And pushed to extremes it makes for sociopathic reasoning where nothing else matters. We humans have a hard time finding the middle way : no I don't want communism but I don't want full blown predatory capitalism either.
  - DrBazza 6y ago

p1necone 6y ago

Intel has just ridden on marketing and anti competitive maneuvering for so long that their actual engineering side has gotten complacent/suffocated. I'm sure they could be (and probably are on their way to being) back as a serious player with management that ceases the bullshit.

jokoon 6y ago

Spectre and meltdown are key events that explain why intel is in this situation right now.

Jweb_Guru 6y ago
The key thing that caused Intel to be in this situation right now was that their EUV play totally failed. If it had succeeded when they hoped it would, they would be so far ahead of the competition that none of this would matter.
- throwaway2048 6y ago
  intel 10nm was never an EUV play.
  - Jweb_Guru 6y ago
    I don't recall saying a single thing about 10nm.
    - throwaway2048 6y ago
      There was never an EUV play they were making that would have catapulted them far ahead at this point.
      - Jweb_Guru 6y ago
        Prior to 2015, Intel believed they would have production ready, 7nm chips by 2018 using EUV, which would indeed have catapulted them far ahead of the competition. Obviously, many people were skeptical about this, but if it had worked out I don't think we'd be having this conversation, nor would people have cared about them missing the 10nm target.