This is the text from a LinkedIn post by a Google engineering lead for CPU design positions they're opening in Portland. The merchant chip market continues to have a weaker outlook by the month, especially in high margin server CPUs.
Intel on the Brink of Death
- Thread starter XYang2023
- Start date
Google is part of x86 advisory group.
Intel and AMD Form x86 Ecosystem Advisory Group to Accelerate...
Leading tech companies to collaborate on architectural interoperability and simplify software development across the ecosystem.
www.intel.com
Lose billions, they subsequently lost billions trying to re-enter years later! If they had accepted they’d have had to put initiatives to become more efficient, Intel would be far better positioned and that fateful dollar based decision changed the industry
MKWVentures
Moderator
Intel is not able to succeed in the that business. Its pretty obvious.
Patrick Mahomes cant play pro basketball, Intel cannot do mobile, Toyota cannot build Processors. I cannot play pro tennis. Accept it.
Intel is not able to succeed in the that business. Its pretty obvious.
Patrick Mahomes cant play pro basketball, Intel cannot do mobile, Toyota cannot build Processors. I cannot play pro tennis. Accept it.
Businesses can pivot. I have seen it many times. In fact, the most successful businesses pivot. The larger the company the harder it is. I can assure you Intel will pivot or they will die a slow death. That is what happens. If IBM can pivot Intel can too.
Business can and do pivot, the key is the leader and the talent and expectations the new leader set around it.Intel is not able to succeed in the that business. Its pretty obvious.
Patrick Mahomes cant play pro basketball, Intel cannot do mobile, Toyota cannot build Processors. I cannot play pro tennis. Accept it.
Apple and Microsoft are two companies that come to mind that did reinvent themselves a bit.
Of course in Intels case the competitors are very big, formidable and have significant advantages. The new leader had better have a clear angel and execution plan.
5N4y and all the other chest beating by Oat about leadership and building shells isn’t the angle that will being Intel back and thus why he is likely gone.
I think Intel's current path is reasonable. Improving its financial condition is important because once it is stabilized, discussions about its strategy and future direction will become less frequent and contentious.
1. Modern PC Utilization: In modern PCs, many tasks are carried out by dedicated units, such as watching YouTube videos, low-power AI inferences, and graphic applications. ARM's primary advantage lies in its lower power consumption for CPUs. However, as demonstrated by Lunar Lake, Intel's x86 architecture has shown it can achieve comparable efficiency.
2. Server Sector Focus: Intel could focus on developing custom server CPUs that integrate CSP (Cloud Service Provider) clients' intellectual property (IPs). Additionally, it should aim to improve its enterprise services to better meet customer needs and strengthen its market position.
3. Quantum Computing Commercialization: Intel could begin commercializing its quantum computing assets by integrating them into its Tiber cloud infrastructure, offering commercial quantum computing services and establishing itself as a leader in this cutting-edge field.
4. AI Accelerator and GPU Market: Intel should continue expanding its presence in the AI accelerator and GPU market. Given the substantial growth opportunities in this sector, Intel could also commercialize its neuromorphic computing research to explore innovative applications and gain a competitive edge.
5. Robotic Applications: By combining low-power E-cores with Xe GPUs on a single chip, Intel could create efficient solutions tailored for robotic applications, positioning itself as a strong competitor to Nvidia in this space.
6. Open-Source Contributions: Intel could actively contribute to key open-source projects that parallel Nvidia's software stack. This would help create a stronger ecosystem around its products, ensuring better adoption and developer support.
7. Strategic Foundry Investments: Intel should pursue appropriate, non-aggressive investments in its foundry business. These investments would not only support its internal manufacturing needs but also serve as a strategic "put option," offering resilience in the face of geopolitical conflicts or other external risks.
8. Cloud Computing Expansion: It might be worth exploring whether Intel could acquire a small cloud provider to further develop its cloud computing business. This would allow Intel to integrate its technologies more deeply into cloud services and potentially create new revenue streams.
1. Modern PC Utilization: In modern PCs, many tasks are carried out by dedicated units, such as watching YouTube videos, low-power AI inferences, and graphic applications. ARM's primary advantage lies in its lower power consumption for CPUs. However, as demonstrated by Lunar Lake, Intel's x86 architecture has shown it can achieve comparable efficiency.
2. Server Sector Focus: Intel could focus on developing custom server CPUs that integrate CSP (Cloud Service Provider) clients' intellectual property (IPs). Additionally, it should aim to improve its enterprise services to better meet customer needs and strengthen its market position.
3. Quantum Computing Commercialization: Intel could begin commercializing its quantum computing assets by integrating them into its Tiber cloud infrastructure, offering commercial quantum computing services and establishing itself as a leader in this cutting-edge field.
4. AI Accelerator and GPU Market: Intel should continue expanding its presence in the AI accelerator and GPU market. Given the substantial growth opportunities in this sector, Intel could also commercialize its neuromorphic computing research to explore innovative applications and gain a competitive edge.
5. Robotic Applications: By combining low-power E-cores with Xe GPUs on a single chip, Intel could create efficient solutions tailored for robotic applications, positioning itself as a strong competitor to Nvidia in this space.
6. Open-Source Contributions: Intel could actively contribute to key open-source projects that parallel Nvidia's software stack. This would help create a stronger ecosystem around its products, ensuring better adoption and developer support.
7. Strategic Foundry Investments: Intel should pursue appropriate, non-aggressive investments in its foundry business. These investments would not only support its internal manufacturing needs but also serve as a strategic "put option," offering resilience in the face of geopolitical conflicts or other external risks.
8. Cloud Computing Expansion: It might be worth exploring whether Intel could acquire a small cloud provider to further develop its cloud computing business. This would allow Intel to integrate its technologies more deeply into cloud services and potentially create new revenue streams.
The Apple pivot was amazing. QCOM also did some impressive pivots. It could happen for Intel, absolutely.
Why not?
bashee_wang
New member
I think x86 is not suitable for mobile like arm does? And Intel need to use new architecture than x86 if it really want to mobile market,but what the cost?
Or just licence ARM.
siliconbruh999
Well-known member
Or if you have empty fabs just fill them with Cheap x86 silicon by signing a deal with Google for smart devices/TVs we already have Raspberry Pi with x86 they can just fill their fabs this way google is merging Chrome OS and Android anyway and Intel has AOSP x86 build with them so there is that
LattePanda - x86 Windows/Linux Single Board Computers
Pocket-sized x86 Windows/Linux Single board Computers. Ultra Low power mini PC with integrated Arduino.
The Motorola/Freescale/Fujitsu cell phone team that Intel purchased in 2013 was an amazing team. Intel then doubled the number of people on it and took twice as long, then threw it away. Intel can try this again and perhaps not screw it up. Limit the team to 100 people maximum at $25M/year budget and get a great old school engineer to run it. Keep the team isolated and allow them to use ARM, RISC-V, whatever, just as long as they use their internal fab this time. Call it a skunk works project. Avoid the "Intel way". It is easier to design a system now than it was in 2013. Do this with 1/2 the employee count of Fujitsu (1/4 the post-Fujitsu employee count).
We are currently working on a RISC-V based WiFi/BT/5g SiP (SiGe + finfet-duv) to prove in our EDA tools/PDKs/environment with our tiny US based team of less than 50 people, but we are trying to do this without a DEI officer, middle management, personnel department, nor bean counters. As we mature and bring on meeting goers, regurgitators, and empire builders, we can layoff our fellows/architects and compete with Intel using foundries that allow us access to their DRM.
We are currently working on a RISC-V based WiFi/BT/5g SiP (SiGe + finfet-duv) to prove in our EDA tools/PDKs/environment with our tiny US based team of less than 50 people, but we are trying to do this without a DEI officer, middle management, personnel department, nor bean counters. As we mature and bring on meeting goers, regurgitators, and empire builders, we can layoff our fellows/architects and compete with Intel using foundries that allow us access to their DRM.
Chuckle ... most people forget that Apple was on its death bed once. Ironically, the same vertically integrated philosophy that nearly killed Apple seems to be working great this time around. Color me surprised.
I guess the difference was that Apple wasn't the majority of the market in PC's (by a long shot) and they are in mobile phones. Perhaps the "buy everything from me in my own ecosystem" business model works only when you are dominant in the industry?
Still, this same philosophy killed IBM in the PC market (remember microchannel?).... and when things in PC started to slide, they simply felt they could drive the market back into mainframes .... wow.
I'm not surprised. Assuming you have leading or competitive technologies, and the R&D funds to maintain and enhance them, IMO vertical integration works better every time. Horizontal technologies and industry specification groups are a PITA for innovation.
It seems to work great for Macs, yet Macs are less than 15% of the laptop and desktop markets.
I don't agree that IBM was trying to drive everything back to mainframes, but the relatively low margins of PCs definitely hurt their corporate margin number. Outsourcing most of the system value to Intel and Microsoft was probably not, in hindsight, a great idea. IBM and their investors clearly liked proprietary vertical integration, and very expensive products sold by the high-tech industry's most costly and premium sales force. PCs needed a retail channel. They couldn't stomach x86-based servers either, and dumped that business too. But IBM still sells about $65B of mainframes (z-systems) per year. Not a bad business, breeding dinosaurs.
This is really the point I’ve been trying to make, that disruption is as much about business models as it is about technology change, and it’s much harder for a company to change its business model than it is to develop technology.I'm not surprised. Assuming you have leading or competitive technologies, and the R&D funds to maintain and enhance them, IMO vertical integration works better every time. Horizontal technologies and industry specification groups are a PITA for innovation.
It seems to work great for Macs, yet Macs are less than 15% of the laptop and desktop markets.
I don't agree that IBM was trying to drive everything back to mainframes, but the relatively low margins of PCs definitely hurt their corporate margin number. Outsourcing most of the system value to Intel and Microsoft was probably not, in hindsight, a great idea. IBM and their investors clearly liked proprietary vertical integration, and very expensive products sold by the high-tech industry's most costly and premium sales force. PCs needed a retail channel. They couldn't stomach x86-based servers either, and dumped that business too. But IBM still sells about $65B of mainframes (z-systems) per year. Not a bad business, breeding dinosaurs.
Intel is still great at developing technology but the underlying business model of the semiconductor industry has changed.
MKWVentures
Moderator
Intel moves too slow. All of their processes and cultures are meant to move slow and steady. They also assume that new products take more time so they make decisions on that.
We know this two ways:
1) People have worked at other companies and Intel, the document the difference
2) Intel has tried mobile use multiple architectures, with multiple organizations, with multiple products. People who know, realize that the reason for failure was not that it wasnt tried or prioritized, its that it was such a financial problem, it had to be stopped.
MKWVentures
Moderator
There are reason why Intel behaviors and skill set and processes doe not lend itself to some other markets. Intel has pivoted many times... it hasnt worked well.
IBM cannot do PCs or Semiconductor manufacturing or hard drives despite being a leader at one point. They pivoted to what they are good at and away from what they are not good at.
There is nothing wrong IMO with changes to who the leaders are and companies refocusing and downsizing and even going away (most great tech companies disappear eventually) It is why Tech is so great.
Intel Product group will be a solidly profitable business for the next 20 years with higher revenue than AMD IMO.
Intel is telling you how good they are at pivoting. Its not the CEOs fault IMO . but let see how their next pivot goes.... we will get the answer.