News Bill Gates says Intel has lost its way, hints that 'brave' Pat Gelsinger exited too soon

[Admin]

Microsoft co-founder Bill Gates wonders whether Intel can survive.

Bill Gates says Intel has lost its way, hints that 'brave' Pat Gelsinger exited too soon : Read more

 

[bit_user]

While Intel rested on its laurels and then missed the entire rise of AI both in terms of hardware and in terms of investment in Open AI and other startups, Microsoft became one of the leaders of the artificial intelligence world.

This is rather unfair. By 2016, Intel definitely saw the potential for AI. That's when they acquired edge AI chip designer Movidius and where the IP for the NPUs in CPUs like Meteor Lake and Lunar Lake came from.

Also, in 2016, they acquired Nervana Systems, who designed custom hardware and software for cloud-based AI training & inference. This would later be superseded by their acquisition of Habana Labs, but it shows when Intel really got moving on AI.

Then, in 2017, they bought Mobileye - a developer of hardware & software for self-driving cars.

Also, in 2017, they launched Knights Mill, a version of their Xeon Phi that had additional instructions targeted specifically at deep learning.

Not only that, but if you look at their iGPUs, the Gen8 architecture included in Broadwell (2014) had support for packed fp16 dot product. I think this was probably added with an eye towards accelerating convolutional neural networks, since it doesn't have much application in 3D graphics. PC games weren't coded to use fp16 and I'm not sure it even supported the full contingent of operations you'd need if they were.

Compared to Nvidia, Intel certainly did get a late start on AI. However, no worse than AMD, and Intel clearly had a lot more resources at its disposal back then. I can't say exactly why they failed to succeed in this market, but I don't think it's entirely due to being blindsided and starting too late. Especially not if you count starting from the founding of the companies they acquired.

I will say that Intel might be in a very different position, had they gone with their iGPU architecture, instead of basing Xeon Phi on x86. Ponte Vecchio definitely came much too late (and probably cost too much?) to achieve its market objectives. Technically impressive, but that only matters to us geeks.

While Intel can hardly go bankrupt considering that it owns some of the best semiconductor production facilities in the industry and sells the lion's share of CPUs for PCs and datacenters ...

Oh, yes it can. Declaring bankruptcy is what you do when you can't get loans to cover the shortfall left by your revenue and cash. Bankruptcy is just a status which allows them to negotiate with their creditors. It doesn't mean they stop operating or that they won't emerge from it as a once-again healthy company, which is often the goal.

 

[Gururu]

This article certainly magnifies the importance of what was a few brief comments in an interview covering a number of topics. Bill Gates is probably the utmost authority on the industry however and his views are extremely important.

 

[QuestionDude]

Hello,

“Intel has missed several consecutive tech industry megatrends.”
Smile
I don't think what Bill Gates says matters. But hey, he's a billionaire, so he's one of the Gods of this planet...

It's still ironic that we listen to a guy who ignored and still ignores four billion potential consumers: smartphone users. And we're not just talking about smartphones, but everything that surrounds them, gadgets, and more importantly, the millions of software developed for them. And not for Windows.

Yes Bill, you should continue to count your billions and pat yourself on the back.

"Oddly enough", Intel followed Microsoft in its delirium, and ignore these four billion potential consumers.

And his successor, Nadella, has his head completely in the "clouds". Yes, Microsoft makes profits, but only because it has a monopoly in the PC field.

Nadella even once talked about selling Xbox. It's to show how "intelligent" he is. But he makes profits so he's a "genius".

And Microsoft and Intel continue to ignore the future, only looking at profits. They throw employees out, and we give them subsidies.

It's well known, the public feeds the "private" (the only thing "private" is profits).

Bill the "visionary". In a year, without lifting a finger, he will have ten billion more in his tax havens...

 

[Pierce2623]

This is rather unfair. By 2016, Intel definitely saw the potential for AI. That's when they acquired edge AI chip designer Movidius and where the IP for the NPUs in CPUs like Meteor Lake and Lunar Lake came from.

Also, in 2016, they acquired Nervana Systems, who designed custom hardware and software for cloud-based AI training & inference. This would later be superseded by their acquisition of Habana Labs, but it shows when Intel really got moving on AI.

Then, in 2017, they bought Mobileye - a developer of hardware & software for self-driving cars.

Also, in 2017, they launched Knights Mill, a version of their Xeon Phi that had additional instructions targeted specifically at deep learning.

Not only that, but if you look at their iGPUs, the Gen8 architecture included in Broadwell (2014) had support for packed fp16 dot product. I think this was probably added with an eye towards accelerating convolutional neural networks, since it doesn't have much application in 3D graphics. PC games weren't coded to use fp16 and I'm not sure it even supported the full contingent of operations you'd need if they were.

Compared to Nvidia, Intel certainly did get a late start on AI. However, no worse than AMD, and Intel clearly had a lot more resources at its disposal back then. I can't say exactly why they failed to succeed in this market, but I don't think it's entirely due to being blindsided and starting too late. Especially not if you count starting from the founding of the companies they acquired.

I will say that Intel might be in a very different position, had they gone with their iGPU architecture, instead of basing Xeon Phi on x86. Ponte Vecchio definitely came much too late (and probably cost too much?) to achieve its market objectives. Technically impressive, but that only matters to us geeks.


Oh, yes it can. Declaring bankruptcy is what you do when you can't get loans to cover the shortfall left by your revenue and cash. Bankruptcy is just a status which allows them to negotiate with their creditors. It doesn't mean they stop operating or that they won't emerge from it as a once-again healthy company, which is often the goal.

Wonderful points today sir. I agree 100%.

 

[Pierce2623]

Hello,

“Intel has missed several consecutive tech industry megatrends.”
Smile
I don't think what Bill Gates says matters. But hey, he's a billionaire, so he's one of the Gods of this planet...

It's still ironic that we listen to a guy who ignored and still ignores four billion potential consumers: smartphone users. And we're not just talking about smartphones, but everything that surrounds them, gadgets, and more importantly, the millions of software developed for them. And not for Windows.

Yes Bill, you should continue to count your billions and pat yourself on the back.

"Oddly enough", Intel followed Microsoft in its delirium, and ignore these four billion potential consumers.

And his successor, Nadella, has his head completely in the "clouds". Yes, Microsoft makes profits, but only because it has a monopoly in the PC field.

Nadella even once talked about selling Xbox. It's to show how "intelligent" he is. But he makes profits so he's a "genius".

And Microsoft and Intel continue to ignore the future, only looking at profits. They throw employees out, and we give them subsidies.

It's well known, the public feeds the "private" (the only thing "private" is profits).

Bill the "visionary". In a year, without lifting a finger, he will have ten billion more in his tax havens...

Linux users really like to wear their heart on their sleeve don’t they.

 

[logainofhades]

Linux users really like to wear their heart on their sleeve don’t they.

Not a Linux user, and I could care less what Bill Gates has to say. Never have liked the man.

 

[TheSecondPower]

One area where Intel is doing great right now is GPU designs. GPUs kept AMD alive during the 2011-2016 years.

 

[phead128]

Intel and Boeing, what happened man.

 

[TerryLaze]

It's still ironic that we listen to a guy who ignored and still ignores four billion potential consumers: smartphone users. And we're not just talking about smartphones, but everything that surrounds them, gadgets, and more importantly, the millions of software developed for them. And not for Windows.

They are not ignoring them, both intel and MS tried real hard to get into the mobile market, there was a bunch of tablets and phones based on wintel.
Windows 8 was designed with touch in mind first and got huge backlash for it.

They couldn't compete against Asian companies that make their products for much cheaper. It's a fools errand to enter a market that is already saturated by big players which is also why nobody starts a new x86 company.

They will try again if anything changes and makes a try reasonably possible to succeed.
Maybe when intel makes all their fabs they will be able to make their e-cores so cheaply that they could make a competing smartphone soc that actually is profitable.

 

[Eximo]

Intel and Boeing, what happened man.

You mean IBM, Xerox, and Kodak right? Titans fade when they stop innovating.

Boeing, and a lot of other heavy engineering industries are suffering a bit from brain drain. Last 5-10 years has seen retirement of a lot of experienced engineers, and in a lot of cases, wholesale outsourcing of engineering overseas.

I don't see that with Intel, just they haven't managed to be as much of a market leader as they were in the past. And when there are only a couple of comparable companies, it is easy to show the fault.

 

[Gururu]

Intel and Boeing, what happened man.

Boeing is killing people, hard to compare.

 

[CelicaGT]

Boeing is killing people, hard to compare.

Beat me to it. Boeing is in a class all by itself.

 

[Captain Awesome]

Microsoft tried with their Windows phone. They even bought Nokia. But their OS was missing a lot of the apps Android and iOS had.

Intel also tried to profit from phones, they had a mobile processor out for a year or two before giving up. I think Asus had it in their original Zenfone.

 

[JamesJones44]

The Microsoft comparison is a bit overstated. Yes, they are in the AI mix. However, let's be honest about Microsoft, if they don't Xerox copy AWS with Azure, Microsoft would be in a similar boat to Intel at this moment. They just happened to have the capital to build a competitor to a highly successful product which allowed them to generate a lot of cash, for which they used that money to buy their way into AI. Outside of that Microsoft's innovations have be less than stellar

 

[2Be_or_Not2Be]

I think he hit the most salient point to Intel's dip - their lack of staying on the forefront of tech, like Gordon Moore was adamant in pursuing all those years. I'd posit that it's specifically setbacks in process tech advances that is the current cause of their woes.

All of the "+" advances to Skylake 14nm rather than successful design & transition to 10nm, 7nm, etc., really caused them to lag in technology. TSMC came up & basically ate their lunch by delivering better process improvements than Intel. Then successive Intel CEOs, instead of focusing even more on getting their more advanced processes improved (and working), became stagnant (and lazy) with their huge current market sales of existing 14nm Xeon & client processors. Those CEOs allowed Intel to fall behind, and now they're seen as trailing TSMC, who supplies their direct competitors (AMD et al).

I do think Intel can come back, and their new stuff shows promise. They still enjoy a large market share, but they are losing money. They need to get more focused on their process improvements. Intel Foundry Services could have been the one to make chips for everyone, just like TSMC. So even if their own AI accelerators weren't the most sought-after, they could have been making the chips for the others & continued to make money either way.

 

[dalek1234]

Did Bill just come out of a long hibernation? Intel has lost its ways many years ago.

 

[bit_user]

It's still ironic that we listen to a guy who ignored and still ignores four billion potential consumers: smartphone users. And we're not just talking about smartphones, but everything that surrounds them, gadgets, and more importantly, the millions of software developed for them. And not for Windows.

Huh? You're missing quite a lot. No, Microsoft doesn't ignore smartphone users. They were trying to break into that market with their own OS & even bought Nokia to help, but to no avail. Ultimately, they killed Windows Phone and instead opted to piggy back on Android and IOS. This enables app portability between Windows and those phone platforms and enables you to use Microsoft's technologies and development toolchain to do it.

Yes Bill, you should continue to count your billions and pat yourself on the back.

Phone stuff was mostly after his time. He stepped down as CEO in 2000. Windows Mobile technically launched that same year:

https://en.wikipedia.org/wiki/Windows_Mobile
​

He withdrew from day-to-day operations at MS in 2008, which was the year after the iPhone launched.

"Oddly enough", Intel followed Microsoft in its delirium, and ignore these four billion potential consumers.

No, Intel actually had its own Phone SoC and platform. It spent years trying to break into that market, with limited success, but finally gave up in 2016:

https://www.anandtech.com/show/10288/intel-broxton-sofia-smartphone-socs-cancelled
​

Interestingly, that was the same year they acquired Movidius and Nervana. Coincidence?

And his successor, Nadella, has his head completely in the "clouds". Yes, Microsoft makes profits, but only because it has a monopoly in the PC field.

If you count both Apple and Chromebooks, it's slowly eroding.

https://gs.statcounter.com/os-market-share/desktop-mobile/worldwide/#monthly-200901-202501
​

Bill the "visionary". In a year, without lifting a finger, he will have ten billion more in his tax havens...

He actually said he's paying more tax than he could, but he doesn't think it's right to use things like tax havens so he can pay less.

This is also a sort of weird claim, because he's committed to giving away basically all of his wealth. So, why he supposedly feel a need to use tax havens is somewhat beyond me.

 

[bit_user]

Not a Linux user, and I could care less what Bill Gates has to say. Never have liked the man.

Cool story: I used to harbor a very strong dislike of Bill Gates, due to his ruthless business practices. This was even before I started using Linux, BTW.

When he quit MS and started his Gates Foundation, I was waaay cynical. I figured it was just a cheap ploy to buy himself a legacy. In time, and the more I've learned about their projects and operations, the more I truly believe he's sincere in his aims.

If you listen to some of the recent interviews with him, he sounds very candid and really not self-aggrandizing. For a man of his wealth and talents, I do think what he's done since leaving MS is pretty much a best-case scenario of trying to help as many people and in as many ways as possible, and give away nearly all of his wealth, in the process. I don't know of billionaire who's done better in their retirement. That's my take.

Did Bill just come out of a long hibernation? Intel has lost its ways many years ago.

No, he just published a memoir, which has him doing the interview circuit to generate publicity for it. With Intel's recent flailing, I guess the interviewer just decided to ask him for his take on it.

 

[why_wolf]

Microsoft tried with their Windows phone. They even bought Nokia. But their OS was missing a lot of the apps Android and iOS had.

Intel also tried to profit from phones, they had a mobile processor out for a year or two before giving up. I think Asus had it in their original Zenfone.

In the phone space Microsoft was somehow both to early and too late. Windows Phone CE during the PDA era and then Windows phone after iPhone/Android took over the market and murdered Blackberry.

Microsoft was just asleep at the wheel and didn't care until it was to late.

 

[sjkpublic]

Intel, Boeing, HP, IBM, etc. To big to fail. All surving more on the public vs private sector. The article only briefly touches the mistakes Intel has made. One analogy would be something like steering a giant oil tanker ship in a river.

 

[bit_user]

Hi Terry. It's been a while!
: )

They are not ignoring them, both intel and MS tried real hard to get into the mobile market,
...
They couldn't compete against Asian companies that make their products for much cheaper. It's a fools errand to enter a market that is already saturated by big players which is also why nobody starts a new x86 company.

Well... you can read Anandtech's epitaph of Intel's phone efforts, here (same link I posted above):

https://www.anandtech.com/show/10288/intel-broxton-sofia-smartphone-socs-cancelled
​

As for Microsoft, they basically failed because they were trying to compete against Android, which was free. There just wasn't enough demand for a Windows Phone that people were willing to pay like $100 extra for it. At the time, MS hadn't embraced the spyware mindset, so they saw basically no way to recoup the development costs and that was that.

As I mentioned above, they devised an alternative and lighter-weight mobile strategy that piggybacked on the established phone platforms, but still served a lesser goal of helping them achieve API ubiquity.

 

[bit_user]

You mean IBM, Xerox, and Kodak right? Titans fade when they stop innovating.

Or, maybe they get too profitable, then Wall St. gets too greedy and starts siphoning off too much (in the form of dividends and stock buybacks), and the company is slowly starved of the means to continually reinvent itself, in order to stay relevant.

 

[bit_user]

Outside of that Microsoft's innovations have be less than stellar

I think what they did with Hololens and VR was very impressive! Hololens was truly ground-breaking. Their VR platform was the first to rely entirely on inside-out tracking. Since then, they've been eclipsed by Apple, but MS got there way sooner.

That was really a case of mispredicting the market. VR and AR just never took off like they thought they would. I wish MS didn't shutdown those programs, but I understand they just couldn't keep shoveling more money into that pit, without any foreseeable way to turn a profit.

 

[rkmcquillen]

"This is rather unfair. By 2016"
-- Bill is conflating a couple of topics.

Intel used to have a tick-tock strategy:
https://en.wikipedia.org/wiki/Tick–tock_model

They had a 3-5 year process technology lead over the the competition, ie smaller transistors. They got stuck on 14nm from 2015 to 2021, and lost there multi-decade lead over AMD, IBM, etc.. to newcomer TSMC and even fell behind; this slow move to 10nm is a legendary disaster; so yes, they rested on their laurels... or inability to shrink due to bugs, cost them years and Many, Many Billions of Dollars.

"I can't say exactly why they failed to succeed in this market, but I don't think it's entirely due to being blindsided and starting too late."
-- I think we both agree, that they have not succeeded, but Bill's stated reason was wrong.