[32] Ian Interviews: Pat Gelsinger, Intel CEO
Вставка
- Опубліковано 25 чер 2024
- In February 2024, Intel held its first Intel Foundry Direct Connect event - an opportunity to bring together partners, customers, potential customers, and the media and align everyone with the singular message and vision of Intel Foundry. This means progress reports, updates, and the future. The tone of the event was one of success - Intel Foundry (IF) is set to be a semi-autonomous body of Intel that aggressively fights for business, whether Internal or External, without playing favorites. Similarly, Intel Product Groups (IPG) are being detached from Intel Foundry, with the remit to use the best process nodes available. The whole goal, according to CEO Pat Gelsinger, is to make both sides better for business - built the best process nodes regardless of product, and build the best products regardless of process node. Intel, for the first time, is extracting out the margins between product and manufacturing, in the goal to return to a world leading entity in both categories.
[00:00] Is High-NA economical?
[01:57] Cheaper transistors and more advanced products?
[03:07] Being Number 2 Foundry?
[04:33] Does that include packaging?
[05:32] Intel vs CoWoS
[08:18] Driving Foveros pitch
[09:20] Keeping Intel and Foundry together
[13:20] Why bring UMC in?
[15:58] Does Intel bet the company every node still?
[16:57] Is regaining leadership: foundry, product, or both?
[18:50] Story Time with Pat
Written transcription: open.substack.com/pub/moretha...
-----------------------
Need POTATO merch? There's a chip for that!
merch.techtechpotato.com
more-moore.com : Sign up to the More Than Moore Newsletter
/ techtechpotato : Patreon gets you access to the TTP Discord server!
Follow Ian on Twitter at / iancutress
Follow TechTechPotato on Twitter at / techtechpotato
If you're in the market for something from Amazon, please use the following links. TTP may receive a commission if you purchase anything through these links.
Amazon USA : geni.us/AmazonUS-TTP
Amazon UK : geni.us/AmazonUK-TTP
Amazon CAN : geni.us/AmazonCAN-TTP
Amazon GER : geni.us/AmazonDE-TTP
Amazon Other : geni.us/TTPAmazonOther
Ending music: • An Jone - Night Run Away
-----------------------
Welcome to the TechTechPotato (c) Dr. Ian Cutress
Ramblings about things related to Technology from an analyst for More Than Moore
#semiconductors #intel #patgelsinger
------------
More Than Moore, as with other research and analyst firms, provides or has provided paid research, analysis, advising, or consulting to many high-tech companies in the industry, which may include advertising on TTP. The companies that fall under this banner include AMD, Applied Materials, Armari, Baidu, Facebook, IBM, Infineon, Intel, Lattice Semi, Linode, MediaTek, NordPass, ProteanTecs, Qualcomm, SiFive, Supermicro, Tenstorrent, TSMC. - Наука та технологія
![[33] Ian Interviews: Renee James, Ampere Computing CEO](http://i.ytimg.com/vi/oVYRYMgPgio/mqdefault.jpg)
![[33] Ian Interviews: Renee James, Ampere Computing CEO](/img/tr.png)
Written version: open.substack.com/pub/morethanmoore/p/an-interview-with-pat-gelsinger?r=1ilwc6&showWelcomeOnShare=true
will companies like nvidia move to intel fabs .. seeing how intel and nvidia are direct competitors ..wouldnt it better idea for intel to spin off its fab business into an independent entity ..so all american chip design companies can make use of it ..
As someone who's been working for Intel for a while now, I was so happy when Pat came on board. Having worked with VMware in the past for my last job, I knew that he knew the technology. Tech companies need to stop throwing business people in at CEO. Just listening to Pat talk about what we are doing means the world to me as an employee.
it's very very true, as an Engineer Pat Gelsinger sees the long term, innovation, and future technologies. MBA's see short term profits, layoffs, and good numbers on financial spreadsheets. They might not even know the technologies being used but call the shots. Not something people should approve of in the hardware tech industry.
Yeah he seems to be a great CEO
out of curiosity, do you think your coworkers share the sentiment?
A lot of interviewees understandably give non-answers to Ian’s questions but Pat gave actual answers. Great interview!
It seriously throws me off seeing a tech CEO give real answers. Wow
Gelsinger has pretty technical mind... which is not surprising as he is a former 486 CPU architect. No doubt Intel will strike back in 2025 in CPU and iGPU front (development of new CPU takes 5 years, Gelsinger become CEO in 2021, so 2026 should be the main strike). I think Intel's new CPU generation will beat AMD Ryzen's and GPU ARC Battlemage will be the best iGPU on PC market (beating AMD Radeons with problematic RDNA3 and 3.5). We can see that ARC GPU with bug-fixed Alchemist in Meteor Lake can beat sometimes AMD Radeon in Phoenix. And Battlemage will be even better.
Pat even admitted that TSMC will stay #1 which sound he stands firmly on the ground of reality. This means we will see the best Intel CPUs on TSMC 3nm while the low and mid-range will be produced by cheap in-house Intel 3. Pat is smart enough to repeat Core2 Duo shock again. And this time not only on CPU front but also on GPU front. Very exciting times ahead.
In large part because Ian asks the right questions.
Well Pat's background is as an intel designer/engineer ... not some marketing schmuck that is all about product segmentation and the technology just "develops itself" .. It sounds like Pat is empowering the manufacturing side again and mediating between manufacturing and design/development ... sounds really exciting ...now we need to see a new compute architecture ... not a fully realized one, but one that we developers and engineers can start hanging our hats on to develop towards
You can always tell when a CEO actually understands the technology that the company uses like Gelsinger or Gates or just has an MBA and fakes it like Balmer. Tech companies always do better when the executives understand the tech.
Wow, Pat Gelsinger interview. Congrats for the interview!
Great talk! Best of luck to Mr. Gelsinger and the folks from Intel.
Great interview and congratulations on landing an interview with Mr Gelsinger. Really great to see Pat talking just straight up technology (no marketing).👍
Wow a tech CEO that is actually answering the questions.
Wonderful and insightful interview! This is why engineer should run manufacturing companies and NOT MBAs. Because engineer knows their stuff and not "increase shareholder value 101". The more foundry capacity that exist and diversified away from Taiwan the better it is for the world cost wise and strategic security wise.
Just imagine what the $152 billion (since 1990) that Intel spent on stock buybacks could have instead funded...
@egalanos well, they own those stocks. They could just sell them again to get the money back. Stock buybacks are basically a tax advantaged savings account
@@moist_ointment I'm sure it was the most tax effective short term course of action for increasing shareholder value by rewarding executives with remuneration packages based on the stock price.
Though perhaps with more reinvestment, Intel could have remained the leader as the market's computation demand changed in areas such as low power cores & SoC, GPU, highly threaded (i.e. workloads run GPGPU), AI and perhaps also never have lost their process leadership.
It's a shame we can't peer into an alternate timeline to see what would have happened...
@ointment I think the point he's making is that it's a massive shame Intel basically sat on that money instead of investing in the future. It's easy to have money and let it sit in a bank/stock account. They didn't add any value to the world by doing that though.
Intel is now clearly adding value; new nodes, being built all over the world instead of just Taiwan, and new computing architectures - instead of churning out 100 variants of almost identical nodes and architectures every year, or buying startups and stripping them for parts.
@liam.weight R&D isn't just some black box where the more money you put in, the better the result is. Intel was the best fab in the world for most of their existence. The market we live in now, where TSMC is better, is a relatively new caused my Intel stumbling 10nm.
But think about 10nm for a second: Intel was trying to release 6 core 10nm Icelake desktop CPUs back in 2015. They wouldve had Alder Lake (or at the very least, Golden Cove) in 2017 if all went according to plan. They wouldve had Meteorlake in 2019, and Arrow Lake 3 years ago.
These were extremely ambitious plans. They failed primarily BECAUSE of their ambitions. Just throwing more money at the problem wouldn't have avoided this.
What would have avoided this is risk mitigation. A portable architecture that could be manufactured on TSMC if their fab is stuck. Less ambitious node improvements. Multiple nodes developing in parallel
Pat qualifies as Jim Keller-class of a technology geek. It would have been great to see both of them do their magic at Intel. Great achievement that Ian could get him on his show!
It is great for Ian to achieve to have him on the show yes.
But:
No, he does not qualify as that. He's a professional bullsh*tter talking major corporate speech while stifling internal Innovation as well as external competition. And he holds questionable personal beliefs that should not factor in on a technological context, but they do tell something about the culture his leadership will and does spread in his company
Pat once said a stupid thing about Apple, that a consumer electronic company has no chance. Here we go, Apple Silicon is so much superior to what Intel has ever achieved. What a clown.
@@1samm1 You know that Gelsinger was the architect of the 80486, right? One of the most successful CPUs in Intel's history. He's not just a talker, he's actually put in his years as a front-line engineer, just like Keller.
@@warren_r I am aware, as well as that is an impressive level of technical achievement I will never reach. Still doesn't make the x86 ISA a good thing, nor does it make him a good leader, nor does it help him spread a good corporate culture moving Intel on towards something better than they have been
@@1samm1 I don't know what you expect of him. But let's keep it realistic. Intel as a large company is a beast for every CEO to tackle with a lot of infighting between each division for ressources. Thanks for Intel, Pat has it what it takes to steer them into the right direction. He made his own mistakes in the past (Larrabee) and he needs to be more cautious about his PR statements ("unquestioned leadership"), but I admire that he doesn't shy away from hard decisions and mighty investments to tackle some of Intel's fundamental problems. Their previous three CEOs were not as capable nor ambitious technologically (while milking customers as they could, so financials were great as long as they kept their design/manufacturing advantage). But due to a chain of poor choices and underinvestments into their core products and technology, they eventually stand where they are today. I hope they got that message to keep up on the technological level or the competition will surpass them sooner or later.
You know that feeling when you walk into the room where two theoretical physicists talk about the bundles, the gauge theory, the Calabi-Yau manifolds? This talk mentioning all the technologies and you have no idea, because you paid no attention to the whole discussion? Yeah, that's me right now. :D
Great interview. Way better than most interviews we have heard so far. We get to hear the economic side of high NA and their plan of advanced packaging.
Most interviewers fail to address these questions.
Great interview! You can tell Pat is happy to talk about the technologies and future he’s developing for Intel and a great accolade of any interviewer is being able to give the interviewed party enough structure to intelligently answer long form questions in a natural manner without needing course correction or constant interruption!
Congrats for the interview! Nice!
Well done Ian this is a great interview! You've managed to thread the needle well, in convering some of the critical enabling technologies Intel is betting on, while still being able to consider the bigger picture and key proving steps Intel as a business have ahead. Too many tech ceo inteviews from mainstream outlets - riding the wave of the ai boom - feel a mile wide but an inch deep in terms of substantive coverage. Are there bigger guests left in the semiconductor space that you haven't interviewed ...C. C. Wei?
Lots of tech CEOs I haven't interviewed 😂 it only comes about when the company has stuff to say, and usually aligned with a particular announcement. The interviews with the engineerijg focused execs are always the most enlightening
Pat is a wealth of knowledge and seems very confident. Would’ve been nice to see him and Jim Keller cook up some ideas together. Excellent interview Ian I like that pat is frank and answers questions. Like some ceos probably wouldn’t say “I bet the company on 18A” but it’s good to see insights into how pat is thinking.
I wish Intel well. I hope they come back and execute well and on time. Full steam ahead onto the angstrom express.
16:42 Pat Gelsinger: I want the customer to become more multi-threaded
Out of context quote ;) It continues as “… in terms of revenue and options…”. Here is the Foundry Customers, not PC
Still better and "Silly-conomy"
Pat Gelsinger is really impressive, with his in-depth technical knowledge. He's literally an expert in his field, and not just another MBA CEO like before at Intel.
One of the rare moments where I give it a thumbs up before watching the video. Nice.
Wow that was an amazing interview! And interesting he was talking about the 386 at the end and not the 486. Thanks Ian for getting this together and sharing with us!
Very good, fairly frank. Nicely done!
He is the visionary leader that Intel needs, I am hoping the best for him and his plans at Intel.
Fantastic interview. Thank you to Mr Gelsigner and Dr Cutress!
Unprecedented, intel ceo sitting around with youtubers, it's a great initiative good job Pat!
I'm mostly a consultant, UA-cam is a side gig :)
Love both these dudes
Considering how much he costs an hour, they were very gracious in providing time for the interview. And as far as I can tell it was a good choice. Lovely chat and great to sit in on it from UA-cam.
Getting the technical message out is also important, from both an investor and recruitment perspective. Ian talks to big money people, but also has an engineer audience.
Excellent interview! Love Pat's frank and through answers!
I really like Pat, I hope he and Intel can pull this all off.
This is massive
Having the mighty Pat ❤❤of Intel is worthy of celebration 🎉
I love Intel and their competitors, their race made me love tech more.
Thanks for your content, you're the best
Great interview, I understood all questions and answers and I feel like Pat Gelsinger is really pushing Intel right now. Let's hope it works out, because we do need more competition in the FAB market.
Pat joined Intel in 1979 at the age of 18 and he became the company's Chief Technology Officer in 2001.
Unlike some of Intel's previous CEO's in the last decade, there should be little doubt that he knows the
critically important nuts and bolts of the semiconductor industry.
This is potentially huge. Best of luck to Intel here!
The potential has always been huge, same can be said for all INTC's competitors. What matter is the actual execution of the roadmaps and plans, this is where intel has been lagging badly over the past decade. One slip after another followed by cancellations. To a point where both investors and customers has lost a lot of faith in the company.
It's pretty obvious, unless intels actually start proving that they can be a leader again other than their market share lead which is ever shrinking, they need to put actual product on the floor. A product that is clearly better than the competition in key performance metrics.
Outstanding Ian, I really don't think you could have done better in the time allowed, truly excellent interview to get the answers that the industry needs to hear. For me the vital points were that he absolutely emphasized the management of risk - frankly this is what the Chips Acts in USA & EU are actually about. Another, was the range of packaging options for all sectors - I recall days when Intel was hostile to automotive. Now of course, the guy that some other comments refer to - Jim Keller - is focusing on auto. Your previous opinion of Dr Ann Kelleher's ability, to deliver the nodes we are seeing, is vindicated. Far more worthwhile my time than the FT & chaplets.... Thank you and looking forward to lots more high quality discussions about semiconductor advances and impact.
It's good to see Pat talk tech, rather than marketing. Seeing really intelligent people talk marketing really dumbs them down.
Taking marketing in a severely dumb way is what Gelsinger does in public all the time. You see the outcome of that in corporate culture already. Chips glued together, the naming slides ignoring the own inane naming choices, etc.
Yeah, he should have stuck to the technical side within Intel.
It's incredible how much this man singlehandedly has done! 🤯
This interview is extraordinary. Credit to both Ian and Pat. Perhaps you do it from time to time.
Great talk, wish it was longer though 😂. Thanks for uploading!
Great interview! I hope Intel pull off 18A.
Great Interview!!
The mics are a good investment! Much better to listen to interviews now.
To be fair, the room and noise floor were better than the average!
Great to hear some details. Nice interview Ian!
Lots of good information!
Great questions and interesting insight into Intel's perspective. Thanks Dr. Cutress.
a very good interview and how deep Pat is still into the technology
Pat Gelsinger is literally the man that wrote the book on the 386. He's also a great human being.
This was fantastic and I hope you can get another one perhaps with more time available down the road. Also hoping for something this year with Dr Kelleher as this year is going to be huge for foundry.
Great interview. Good job. Thank you
Great work as always!
Great interview as always!
This was an excellent interview!
damn how'd you nab this interview? nice job and nice interview. this the kind of work that gets you a big subscriber bump. keep it up mr. potato. P.S. it took me about 5 minutes before i realized there's a huge man standing in the background.
Fantastic interview, Ian. Great stuff
Intel needed Pat about 12 years ago. No doubt he's smart, it's more a question of was Intel too late to make the change.
Amazing content, thanks Ian
The combo of High-NA and advanced packaging is looking really interesting. It might help Intel to pass TSMC for chiplet designs even though TSMC might maintain an edge for monolithic designs. It would be funny if AMD switched to Intel as their primary foundry considering their longtime rivalry.
As an engineer myself I appreciate, that Intel installed so many great engineers, which also can communicate their visions, in high management position. Pat Gelsinger of course, but also Tom Petersen für the grafics-division does a great job bringing trust back in Intels future... at least for us geeks understanding some of the tech.
Overcoming the current hurdles (reviving Moor's law, entering the dedicated GPU market) these guys seem to be the absolute right decision, cause you need to invest a lot of money first, to reach this visions, before you get the returns making the shareholders happy again. You need to live a long term tech-strategy, not a short term investment-strategy for some years.
But as I'm also half an MBA I want to counter this among geeks widespread impulse to blame everything on MBAs - best "let all companies (except maybe banks) be run just by engineers". This also will not work - companies need both. RnD costs a lot of money... engineers tend to aim for perfection no matter of the costs for the last 1%... and also this initial RnD investment must be covered somehow... Most of us are not only engineers, but also share-holders of one or the other company... and at least I only give my money to companies, where I expect some positive return at least mid-term... that's the job of the MBAs ensuring investors and shareholders have trust in the companies future and keeping the cash burn rate controlled.
Finally there are also these types of MBAs, who don't understand or even care how the company works, think they know everything better, than even the most experiences employees and just decide on some totals at the bottom of a spreadsheet... get rid of these guys asap!
This was a great interview, Pat is a really good CEO.
Nice interview! Keep up with the great work
Awesome work!
Love getting an interview with a head honcho that isn't just fluff
Great respect for Intel and Pat agreeing to sit down with you knowing you won't pull punches. Fabulous
What an interview, fantastic!
Wow amazing interview man! Soo interesting!! 😺😺
Can't say I've been the biggest fan of Intel lately, but Pat seems like he has them headed in the right direction. The US needs more semiconductor manufacturing independence with Taiwan's.. peculiar geographical position.
Great interview and a weird background as a bonus ;-)
Great interview, thx a lot Ian!
"I' bet the whole company on 18A."
Woow excellent interview. Very informative and detailed.
What a fantastic interview. Thank you, Ian!
Great interview 👍 Thank you for bringing it to viewers.
Boy, it feels like a high wire act over at Santa Clara. Why is that?
Kindest regards, friends and neighbours.
Would be great if you could get a whole interview on what pat has done before like the last segment instead of the investor keynote kinda thing.
You can pitch it to him as giving stock holders confidence in Intel's CEO
Well done Ian!
what a scoop! looking forward to seeing this tonight
Lisa Su next?
What a great interview
Did you hear anything about Falcon Shores? Will that be on 18A?
Great interview. Pat did minimal spin, also funny he caved and just calls them chiplets. You know he's talking to industry more for that switch.
Pat is amazing with body language.
Very interesting interview. I hope they achieve their goals!
5 Nodes in 4 Years.....🤐
How to get in touch with you? I need some career guidance
Time for a vodka break, I think! 😉 Great interview, Ian. Thanks! 👍
Wawww ,i can't believe that you got this interview.
It’s great to see an engineer back at the helm of Intel
This background makes Pat look like he has some serious muscle guarding him lol.
Is that Ryan George in the Intel employee badge?
I miss his 386 and 486 design.😊 I talked someone on another UA-cam channel and found out that Roger Van Brunt was his colleague next to its cube in the fab of intel.
Wow, hum... that I did not expect to see :)
The innovation in products and services is what serves Intel's bottom line, not just volume, not just financial trickery. Get the innovation right, everything else follows. Pat definitely gets that.
Lol my first pc was a Topis 386 SX 25 Mhz in 90's. I have that Pc until today and Still Works.❤
Why not #1 as an external foundry? Push for even more bold tech.
Pat seems like a real person and not a CEO. He actually answers questions and is technical enough to help you understand things in simple terms. I'm glad AMD is doing well but Pat makes me want to root for Intel which feels weird.
Something wierd here,foundry means to support and share the success of customers instead of saying or proving I (Intel) am the best tech leader or supplier in the field,right ?
Great interview btw
I'm glad Intel is still fighting their best
“Kids, Papa’s Here,”
Bravo!
Did Papa bring you all the chips?
*Summary* generated with gemini advanced 1.0 on 2024-02-25
* Intel is on a path to regain leadership in both process technology and product offerings by the end of 2025.
* The company is making significant investments in its foundry business, including a new partnership with UMC to manufacture 12nm chips.
* Intel is also developing its own advanced packaging technologies, such as EMIB and Foveros.
* These new technologies will enable Intel to offer customers more competitive and differentiated products.
*Summary*
*High-NA Economics*
* *0:23* Intel is confident that the economics of high-NA EUV will work out.
* *0:43* High-NA EUV can achieve economic benefits compared to double patterning and self-alignment techniques.
* *1:08* Intel is working with ASML and internal teams to achieve bigger mask sizes, enabling better economics across EUV processes.
* *1:34* Intel's focus is on driving down transistor costs to maintain the economic benefits of Moore's Law.
*Intel's Foundry Ambitions*
* *3:08* Intel aims to be the world's #2 foundry by 2030 (by revenue, excluding internal wafers).
* *3:43* Their focus is on external foundry customers, aiming to surpass GlobalFoundries, UMC, and Samsung.
* *4:30* Advanced packaging will be a significant, though not dominant, part of foundry revenue.
*Competitive Advantages*
* *5:32* Intel sees Foveros (and its future iterations like Foveros Direct) as comparable to TSMC's CoWoS.
* *5:59* Some customers are switching from CoWoS to Foveros due to supply constraints.
* *6:42* Intel believes EMIB offers cost advantages for high-performance bridging in certain designs compared to CoWoS.
* *7:09* Advanced testing capabilities for individual die/chiplets are another key differentiator.
* *7:38* Foveros Direct enables full die-on-die 3D assembly for flexibility in node choices and design optimization.
*Intel Foundry and Product Synergy*
* *9:21* Intel's IDM 2.0 strategy necessitates close collaboration between product and foundry teams.
* *9:53* Formalizing the interface between product and foundry groups will improve cost structures and engineering discipline.
* *10:26* Product teams will have flexibility to use external foundries if needed, while the foundry will seek external customers.
* *10:37* Intel's factories require internal product volume to be economically viable. This interdependence is key to success.
*Intel's Foundry and the Value of Partnership*
* *11:42* Intel's foundry customers benefit from Intel committing its own revenue to new process technologies, reducing risk for external clients.
* *12:26* Intel's product teams gain access to a wider range of mature technologies developed by the broader industry in partnership with the foundry, improving choice and flexibility.
*Intel and UMC: Why the Collaboration?*
* *13:43* Partnering with UMC allows Intel to learn from a company experienced in creating and supporting multiple process design kits (PDKs) for diverse purposes (high-voltage, RF, etc.).
* *14:58* UMC gains a well-capitalized factory and supply chain expansion, making them more resilient for their customers.
* *15:24* UMC's existing customers are interested in the Intel 12 node, creating new business opportunities.
*Intel's Continued Commitment*
* *15:58* Intel has historically 'bet the company' on new process nodes, and 18A is no different.
* *16:24* Intel's focus is on multi-faceted growth: process technology improvements, packaging innovations, and customer diversification.
*Regaining Leadership*
* *16:57* Intel's aim is to regain leadership in both foundry offerings and product quality by the end of 2025.
* *17:09* Intel 18A offers leading-edge transistors. Next-generation products like Clearwater Forest (server) and Panther Lake (client) are already in fabrication, pushing that technology forward.
Disclaimer: I used gemini ultra 1.0 (2024.02.21) to summarize the
video transcript. This method may make mistakes in recognizing words
and it can't distinguish between speakers.
What I would ask Pat: How would you convince clients to move to IFS when Intel is moving to TSMC to make their next leading edge CPUs?
What does he say in 1:18? Is it mass or mask size? What's he talking about here, how would it allow bigger field sizes?
So clearly not sillyconomy!