Intel teams with Google Cloud to develop new class of data center chip
Intel Corp and Alphabet Inc's Google Cloud on Wednesday said they have worked together to create a new category of chip that Intel hopes will become a major seller in the booming cloud computing market.
The new chip, which is called Mount Evans and will be sold to others beyond Google, reflects the way that cloud computing providers operate. They build huge data centers full of powerful physical computers and sell virtual slices of those machines to other businesses, who in turn get better bang for the buck than building the machines themselves.
For cloud providers, tasks like setting up the virtual machines and getting customer data to the right place are essentially overhead costs. The Mount Evans chip, which Google and Intel have dubbed an "infrastructure processing unit" (IPU), separates those tasks out from the main computing tasks and speeds them up. Doing so also helps ensure the safety of those functions against hackers and adds flexibility to the data center.
- Reuters
L
Remember Apple M1 performance, which to be honest is impressive (from efficiency point of view) and how much (deserved) hype it generated?
Keeping it at very high level, the main reasons were a very wide decode unit (=8, compared to max 5 in x86 usually) and a DEEP Re-Order Buffer (ROB, which keeps a buffer of requests to be executed by the CPU. Please google Apple M1 Anandtech if you really want to get the nitty gritties). Of course a 5nm node fab also helped Apple in M1 which BTW is an impressive product.
Why can't Intel emulate that was a question asked? The main reason is Apple owns the whole ecosystem, including software (read OS) which Intel DOES not so unless software supports these, there is no point in implementing those in hardware (CPU) by Intel.
Enter Alder Lake (with BIG.LITTLE arch) and MAJOR Windows release (WIndows Sun Valley), due late fall 2021. The timing is not a coincidence. In order to compete against Apple, Wintel has to work in tandem which they seem to have done. Moore's law is dead channel has mentioned explicitly that MSFT is coming with major updates to scheduler and it is no coincidence Intel's Alder Lake will be able to exploit those to perform much better. Check that out. If you connect the dots, there is a good likelihood Alder Lake will be revolutionary.
I also just checked randomly on Intel openings, they have several opening for CPU micro-architects with experience in RTL design and expertise in Out-of-Order execution in CPUs, check out yourself. While OOO is not something entirely new, it fits aligns with this narrative.
And yes, it may give a big boost to Intel and its stock price. GLTA longs!
a
Much of the negative reasoning against Intel i see in here revolves around the argument "since Intel has lost a certain amount of market share recently this must therefore be an irreversible process. In spite of the fact that Intel has changed management recently which vows to go in a new direction. In spite of the fact that Intel has declared that it has chosen to forgo stock buybacks for the intermediate term and plow the money back into CAPEX to fund growth. To satisfy market driven chip shortages created by the coronavirus situation. While mighty Intel may be a bit late to the party one cannot assume this proves they missed the party entirely or that they will not end up, in the end, becoming the life of the party itself.
h
Everyone says Intel is dead. Fabs increased production by 100% in last 3 years, running at full capacity and then some. $80billion in revenue! Let me spell that out:
$80,000,000,000
Around $18,000,000,000 in net profit!
Apple and Microsoft hurt, but total they make up about 5% of revenue. The deals that Apple and Microsoft had with Intel, based on them being Apple and Microsoft, were low profit margin areas that will quickly be replaced by higher margin customers. The demand is tremendous.
Intel’s moving to AI chips, MobileEye, cloud and that will keep revenue growing.
I believe this will be one of the gems that people,in 3-5 years, are going to be very happy they purchased.
Gartner estimated Surface sales of about 600K per qtr - say 2.5M a year.
1. Say Surface ARM is wildly successful (NOT) and reach 1M units a year in 3 years at $500 each (Surface RT $500) ie Surface ARM revenue of $500M a year for a company doing more than $150B sales by that time! Say it's very profitable at 20% net margin ie net profits of $100M for a company that's today doing $45B of net income from operations. Does it make sense for them to embark on a MAJOR Chip R&D effort for a measly additional annual revenue of $500M and a measly 0.2% contribution to the bottom line IF IT'S VERY SUCCESSFUL - which is very doubtful.
2. Surface ARM will have great battery life. At the time Surface ARM RT was introduced it had about 2X the typical X86 battery life BUT IT FLOPPED .... "Microsoft's original Surface model was a $400 ARM-based Windows RT product, flanked by a more conventional, low-voltage x86 Surface Pro model that the company compared to Apple's MacBooks. Surface RT flopped dramatically across four years of trying because it didn't do what Windows customers expected of a Windows device: run most Windows software. That left Microsoft to focus on Surface Pro as its mobile, x86 PC tablet designed to be used with a keyboard and stylus. " Today a typical Ryzen 4000 notebook runs 8 hrs routinely if the manufacturer doesn't skim on the battery size. OK Surface ARP Pro X runs upto 15 hrs - isn't this similar scenario to RT days - at 8hrs it's less compelling to need 15hrs! for me.
3. Microsoft ARM today addresses the performance and compatibility issue or at least the compatibility issues..by creating a 64Bit emulation and an ARM variant Microsoft ARM SQ1 which is "more suited for computer based applications than CP" - with it comes a larger power draw than CP ARM.
"THE KEY HERE IS MICROSOFT'S NEW SQ1 CHIP BASED ON THE "ARM" ARCHITECTURE. These chips are often made for mobile devices like smartphones, but the SQ1 chip has been engineered to provide computer-like performance rather that smartphone performance. " - "As a result, an ARM-based laptop designed for computer performance isn't going to get the same battery life as a smartphone."
"While the SQ1 chip in the Surface Pro X is designed for computer performance, IT WON'T BE POWERFUL ENOUGH FOR HEAVY DUTY TASKS LIKE GAMING OR VIDEO EDITING. But for everything else, like web browsers and productivity apps like Microsoft Office, it's likely all anyone would ever need."
4. Here's the TKO punch .... "The key here is Microsoft's so-called "SQ1" chip, a processor made by Qualcomm that's based on the "ARM" architecture. " SO MICROSOFT IS NOT GOING INTO THE CHIP BUSINESS! AMD OR ANYONE ELSE CAN LICENSE THE SQ1 MOST LIKELY FOR FREE AND DO IT!!
5. NOW - GO BACK AND READ #4 - NOT FOR VIDEO GAMING OR EDITING OR ANY POWERFUL APPS! IF THERE'S ONLY ONE COMPANY IN THE WORLD THAT CAN ENABLE THIS IN ARM - IT'S AMD!!! So in time we'll see an AMD version with all its glory - BIG LOSER $INTC AGAIN!
A
I may not understand new wave investment styles, but I do know $5.44 eps beats $0.51 eps. Others can fantasize about what their dream company is going to do in the future, and that is fine, it is good to dream, but I will stick with proven financials and forgo the pie in the sky expectations. When intel’s earnings start to decline and other companies start to pony up with solid returns, I will gladly look at those other companies. Do not get me wrong. AMD may be a great company with a great future, but for now, they are way overvalued, and Intel is way undervalued. Too much downside potential for AMD and some good upside potential for Intel.
H
11.28 PE and Intel will probably have blowout earnings as always. Intel is most consistent company, but never gets respect it deserves. Good stock for young long term investor. My average cost is $22.44
D
Of all the boards I monitor and post on this one is by far the best. I learn from and enjoy the information shared on this board! Someone mentioned that Lisa Su stated that there is room in the market for both Intel and AMD. At present AMD is winning the performance battle for GPU's for data centers. I believe the competition is good to drive the technology. With all of that said I believe that we are currently in a golden age of technolgy and semiconductors are the DNA. There is a much bigger picture than competition for the data center. AI, 5G, self driving cars, gene sequencing are just a few of the emerging technologies all requiring semiconductors. I saw an article where semiconductor growth is projected to be $90.8 billion between 2020 and 2024. Intel (and TSMC and most likely Samsung) are building capacity to support and it will be in high demand. Intel has a strong balance sheet and PE under 12. It is well positioned to add capacity and I believe utilize some of TSMC's high tech capacity in future. Looking to the future Intel is well positioned to survive and prosper even if they don't win the performance battle with AMD for the data center. Intel and AMD are national resources!! And, given Intel's renewed focus on technology coupled with their financial strength their stock price should be up with the likes of NVDA, AMD et al!!!
r
While INTC is very slow to reward shareholders with capital gains, there is no doubt that those who are patient will be rewarded if they wait long enough. INTC is very cheap on all metrics. 10.84 PE trailing twelve months, 9.22% earnings yield, EV/EBITDA 6.81, EV/free cash flow 12.53, free cash flow yield 7.90%. No way to argue, this stock is a bargain. Only question how long will you be waiting for that return, and is the dividend adequate compensation enough to pay you to wait? These are the things an investor needs to take into account with stocks like INTC. The future itself appears quite bright to me so ultimately I am willing to be patient. I like AMD and NVDA too though. I think all have something to offer. You just have to decide if you prefer quick returns or slow ones, as well as how much risk you're willing to take to get those returns. Compared to AMD and NVDA I see INTC as being low risk. All things to consider when investing. The risk/reward scenario is very important.
e
With recent updates from Intel about oneAPI and Intel Server GPU, it's time to recall why Raja Koduri left AMD for Intel: "As I was thinking through that and the elements that are required, the core pieces of technology required, and which company has these assets, these people, these resources, the only company that checked my list...was Intel," Koduri added.
Of the above 20,000 folk, Koduri revealed that he has a 4,500 strong team working on Intel's future integrated and discrete graphics projects. Koduri emphasised the importance of balance in this design team – between hardware and software expertise. Excitedly, the Intel graphics chief explained "What I'm doing is helping them figure out how to build products that scale up from the low power, mobile domain up to petaflops—the big data center GPUs. Both internally and externally, there's a lot of excitement from our customers, from enthusiasts, from the market about our entry into discrete graphics in 2020."
J
The market is realizing that Pat Gelsinger cannot turn the INTC ship overnight. The latest delay in Sapphire Rapids will allow $AMD to continue taking share of the data center market in 2022. I expect analysts with revise INTC PT down and AMD PT up.
Also, AMD is seeing lots of positive news on XLNX acquisition and the NEW GPU for cell phones. INTC is a value trap. Sell ASAP and redeploy into $AMD.
$NVDA has 608M shares outstanding. $INTC has 4.5B shares outstanding. $AMD has about 1B shares. This is also why $nvda could hit $300 while $intc will struggle to hit $60. This is why when 2H19 AMD takes the performance crown across majority platforms and solutions in the market, $AMD can also claim $100+. Ryzen is already the top 3 Amazon best selling CPUs in USA, Canada and Europe.
2019 Best HPC solution will be AMD Epyc Rome. Best desktop processor will be Ryzen 3rd gen, that much is already a given but it's only when the products are released will the world accept the fact and tides will turn. We're already seeing Vulkan API on Android, Linux and Google Stadia. Vulkan is royalty free and Vega 64 challenges the RTX2080Ti in Vulkan optimized titles already. Without a doubt Radeon will do extremely well in the GPU and Compute market.
Assuming market conditions stays positive, AMD has huge upside imo, worst case scenario it will 2X by next year if Epyc Rome simply delivers on it's performance promise. Best case scenario could be 4X+ in 2 years if Milan can continue the adoption streak together with AMD leadership demonstrating ongoing success in their vision and 5 year execution roadmap.
H
Prices don't make any sense. It is at march 2020 levels now... with all the upcoming enormous plans and new successful products. I have listened to the call - I kinda don't understand what analysts refer to when they say that cloud business has fallen 18% - revenue in datacenter segment is flat for 2019 and 2020, which isn't spectacular - but they can't say that it has fallen and it is all known already and had to be priced in by now. I thought that DCG would be down slightly 1-2% but no - it is even up compared to 2020, so what is this reaction? I am lost.
F
In a note put out by KeyBanc this morning regarding Global Foundries
A GlobalFoundries acquisition would "significantly accelerate plans for (Intel's) IDM 2.0, vs. waiting for the construction of its Arizona fabs, which are expected to ramp meaningful production in 2023," KeyBanc writes in a note today. Intel (NASDAQ:INTC) is reportedly in talks to buy GlobalFoundries for about $30B. That dovetails with Intel's integrated device manufacturer expansion plan (IDM 2.0), a $20B initiative that boosts capacity for Intel to make its own chips and chips for other companies. KeyBanc gives two more reasons why a deal would be positive: It "would provide the Company with significant experience and leadership in operating a third-party service foundry, something the Company lacks today." It "would provide INTC with a broad and established customer base that could provide revenue synergies for potentially upselling to other services, i.e., x86."
R
Some of the talk in the Yahoo video with the Wedbush analyst touched on Intel losing market share to firms now "making" their own chips, i.e. Apple and Amazon. I don't hear that they are actually "making" them. They are sending them to Taiwan Semi to have them made because only TSM (and maybe Samsung) has the fab processes capable of making the chips Apple, Amazon, Qualcomm, AMD, Tesla, etc are designing.
So although Intel Server and Desktop CPU's designs are now being pressured by AMD technologically, I suspect that the ability of TSM to service AMD and all of these other customers will be the bottleneck that saves Intel from meaningful AMD competition.
Also, Intel's strategy of providing fab services to companies like QCOM, AMZN and others is absolutely the right approach, since fab capacity seems to be the issue here.
Now the main issue that INTC investors have to face is how quickly can Intel execute on bringing their new fab processes into production at decent yields as well as simultaneously building new fab capacity. Thank goodness they now have an engineer in charge, and not a CPA, lawyer or salesman running the firm! Pat just has to execute (not a small task).
A
For Intel to intentionally put out negative news in order to scare investors away thereby causing the price to fall so Intel could buy back more shares at a cheaper price would probably result in Intel getting sued for fraud by pension funds and large investors alike in a class action lawsuit. For that reason, it's doubtful Intel is doing anything to intentionally lower price. However, it may be doing what everyone does, study the market and wait for a lower price before buying. Just my two cents.
e
Bob's purchasing INTC shares before leaving INTC says SOMETHING. I believe INTC will announce a breakthrough in 7 nm manufacturing in the next CC. Keep it in mind that Intel keeps Investing in 5nm, 3nm chip manufacturing technologies. The NEW CEO, a technical guy, will help INTC gain chip manufacturing process superiority. It is a matter of time!!!
T
Intel says it’s delaying production of Sapphire Rapids, the 10-nanometer Xeon Scalable successor to the recently launched Ice Lake server CPUs, because of extra time needed to validate the CPU.
L
Party has just started. From the press release:
With today’s CEO announcement, Intel also said that fourth-quarter revenue and profits will exceed prior guidance, and the company said it has “made strong progress” on its 7 nm process technology. The company will report fourth-quarter results on Jan. 21
M
Put the news of Apple's new chip and the NVIDIA-IBM project aside, for a moment. One has to evaluate any stock by its price relative to its expected growth. Even with competition, Intel is still the dominant CPU producer in the world and is liable to remain that way for a long time (at least a decade).
Further, relative to its expected revenue, the stock's price will grow substantially. Currently, its forward P/E is 9.65. That is unbelievable low. It assumes only a 10% growth in earnings for this coming year. Last year, Intel's earnings grew by 28%.
In short, Intel is a huge bargain. It would be one thing if it was overvalued. But, it is isn't.
Intel Corp and Alphabet Inc's Google Cloud on Wednesday said they have worked together to create a new category of chip that Intel hopes will become a major seller in the booming cloud computing market.
The new chip, which is called Mount Evans and will be sold to others beyond Google, reflects the way that cloud computing providers operate. They build huge data centers full of powerful physical computers and sell virtual slices of those machines to other businesses, who in turn get better bang for the buck than building the machines themselves.
For cloud providers, tasks like setting up the virtual machines and getting customer data to the right place are essentially overhead costs. The Mount Evans chip, which Google and Intel have dubbed an "infrastructure processing unit" (IPU), separates those tasks out from the main computing tasks and speeds them up. Doing so also helps ensure the safety of those functions against hackers and adds flexibility to the data center.
- Reuters
Keeping it at very high level, the main reasons were a very wide decode unit (=8, compared to max 5 in x86 usually) and a DEEP Re-Order Buffer (ROB, which keeps a buffer of requests to be executed by the CPU. Please google Apple M1 Anandtech if you really want to get the nitty gritties). Of course a 5nm node fab also helped Apple in M1 which BTW is an impressive product.
Why can't Intel emulate that was a question asked?
The main reason is Apple owns the whole ecosystem, including software (read OS) which Intel DOES not so unless software supports these, there is no point in implementing those in hardware (CPU) by Intel.
Enter Alder Lake (with BIG.LITTLE arch) and MAJOR Windows release (WIndows Sun Valley), due late fall 2021. The timing is not a coincidence. In order to compete against Apple, Wintel has to work in tandem which they seem to have done. Moore's law is dead channel has mentioned explicitly that MSFT is coming with major updates to scheduler and it is no coincidence Intel's Alder Lake will be able to exploit those to perform much better. Check that out. If you connect the dots, there is a good likelihood Alder Lake will be revolutionary.
I also just checked randomly on Intel openings, they have several opening for CPU micro-architects with experience in RTL design and expertise in Out-of-Order execution in CPUs, check out yourself. While OOO is not something entirely new, it fits aligns with this narrative.
And yes, it may give a big boost to Intel and its stock price. GLTA longs!
$80,000,000,000
Around $18,000,000,000 in net profit!
Apple and Microsoft hurt, but total they make up about 5% of revenue. The deals that Apple and Microsoft had with Intel, based on them being Apple and Microsoft, were low profit margin areas that will quickly be replaced by higher margin customers. The demand is tremendous.
Intel’s moving to AI chips, MobileEye, cloud and that will keep revenue growing.
I believe this will be one of the gems that people,in 3-5 years, are going to be very happy they purchased.
Gartner estimated Surface sales of about 600K per qtr - say 2.5M a year.
1. Say Surface ARM is wildly successful (NOT) and reach 1M units a year in 3 years at $500 each (Surface RT $500) ie Surface ARM revenue of $500M a year for a company doing more than $150B sales by that time! Say it's very profitable at 20% net margin ie net profits of $100M for a company that's today doing $45B of net income from operations. Does it make sense for them to embark on a MAJOR Chip R&D effort for a measly additional annual revenue of $500M and a measly 0.2% contribution to the bottom line IF IT'S VERY SUCCESSFUL - which is very doubtful.
2. Surface ARM will have great battery life. At the time Surface ARM RT was introduced it had about 2X the typical X86 battery life BUT IT FLOPPED .... "Microsoft's original Surface model was a $400 ARM-based Windows RT product, flanked by a more conventional, low-voltage x86 Surface Pro model that the company compared to Apple's MacBooks. Surface RT flopped dramatically across four years of trying because it didn't do what Windows customers expected of a Windows device: run most Windows software. That left Microsoft to focus on Surface Pro as its mobile, x86 PC tablet designed to be used with a keyboard and stylus. " Today a typical Ryzen 4000 notebook runs 8 hrs routinely if the manufacturer doesn't skim on the battery size. OK Surface ARP Pro X runs upto 15 hrs - isn't this similar scenario to RT days - at 8hrs it's less compelling to need 15hrs! for me.
3. Microsoft ARM today addresses the performance and compatibility issue or at least the compatibility issues..by creating a 64Bit emulation and an ARM variant Microsoft ARM SQ1 which is "more suited for computer based applications than CP" - with it comes a larger power draw than CP ARM.
"THE KEY HERE IS MICROSOFT'S NEW SQ1 CHIP BASED ON THE "ARM" ARCHITECTURE. These chips are often made for mobile devices like smartphones, but the SQ1 chip has been engineered to provide computer-like performance rather that smartphone performance. " - "As a result, an ARM-based laptop designed for computer performance isn't going to get the same battery life as a smartphone."
"While the SQ1 chip in the Surface Pro X is designed for computer performance, IT WON'T BE POWERFUL ENOUGH FOR HEAVY DUTY TASKS LIKE GAMING OR VIDEO EDITING. But for everything else, like web browsers and productivity apps like Microsoft Office, it's likely all anyone would ever need."
4. Here's the TKO punch .... "The key here is Microsoft's so-called "SQ1" chip, a processor made by Qualcomm that's based on the "ARM" architecture. " SO MICROSOFT IS NOT GOING INTO THE CHIP BUSINESS! AMD OR ANYONE ELSE CAN LICENSE THE SQ1 MOST LIKELY FOR FREE AND DO IT!!
5. NOW - GO BACK AND READ #4 - NOT FOR VIDEO GAMING OR EDITING OR ANY POWERFUL APPS! IF THERE'S ONLY ONE COMPANY IN THE WORLD THAT CAN ENABLE THIS IN ARM - IT'S AMD!!! So in time we'll see an AMD version with all its glory - BIG LOSER $INTC AGAIN!
Of the above 20,000 folk, Koduri revealed that he has a 4,500 strong team working on Intel's future integrated and discrete graphics projects. Koduri emphasised the importance of balance in this design team – between hardware and software expertise. Excitedly, the Intel graphics chief explained "What I'm doing is helping them figure out how to build products that scale up from the low power, mobile domain up to petaflops—the big data center GPUs. Both internally and externally, there's a lot of excitement from our customers, from enthusiasts, from the market about our entry into discrete graphics in 2020."
Also, AMD is seeing lots of positive news on XLNX acquisition and the NEW GPU for cell phones. INTC is a value trap. Sell ASAP and redeploy into $AMD.
2019 Best HPC solution will be AMD Epyc Rome. Best desktop processor will be Ryzen 3rd gen, that much is already a given but it's only when the products are released will the world accept the fact and tides will turn. We're already seeing Vulkan API on Android, Linux and Google Stadia. Vulkan is royalty free and Vega 64 challenges the RTX2080Ti in Vulkan optimized titles already. Without a doubt Radeon will do extremely well in the GPU and Compute market.
Assuming market conditions stays positive, AMD has huge upside imo, worst case scenario it will 2X by next year if Epyc Rome simply delivers on it's performance promise. Best case scenario could be 4X+ in 2 years if Milan can continue the adoption streak together with AMD leadership demonstrating ongoing success in their vision and 5 year execution roadmap.
A GlobalFoundries acquisition would "significantly accelerate plans for (Intel's) IDM 2.0, vs. waiting for the construction of its Arizona fabs, which are expected to ramp meaningful production in 2023," KeyBanc writes in a note today.
Intel (NASDAQ:INTC) is reportedly in talks to buy GlobalFoundries for about $30B.
That dovetails with Intel's integrated device manufacturer expansion plan (IDM 2.0), a $20B initiative that boosts capacity for Intel to make its own chips and chips for other companies.
KeyBanc gives two more reasons why a deal would be positive:
It "would provide the Company with significant experience and leadership in operating a third-party service foundry, something the Company lacks today."
It "would provide INTC with a broad and established customer base that could provide revenue synergies for potentially upselling to other services, i.e., x86."
So although Intel Server and Desktop CPU's designs are now being pressured by AMD technologically, I suspect that the ability of TSM to service AMD and all of these other customers will be the bottleneck that saves Intel from meaningful AMD competition.
Also, Intel's strategy of providing fab services to companies like QCOM, AMZN and others is absolutely the right approach, since fab capacity seems to be the issue here.
Now the main issue that INTC investors have to face is how quickly can Intel execute on bringing their new fab processes into production at decent yields as well as simultaneously building new fab capacity. Thank goodness they now have an engineer in charge, and not a CPA, lawyer or salesman running the firm! Pat just has to execute (not a small task).
With today’s CEO announcement, Intel also said that fourth-quarter revenue and profits will exceed prior guidance, and the company said it has “made strong progress” on its 7 nm process technology. The company will report fourth-quarter results on Jan. 21
Further, relative to its expected revenue, the stock's price will grow substantially. Currently, its forward P/E is 9.65. That is unbelievable low. It assumes only a 10% growth in earnings for this coming year. Last year, Intel's earnings grew by 28%.
In short, Intel is a huge bargain. It would be one thing if it was overvalued. But, it is isn't.