People say intel isn't innovative, but the truth is that they are capable of creating chips using the most cutting edge and advanced technologies of 2007
I mean, nowadays 14th gen is the best CPU on the market for consumers, and Intel is beating AMD across the board from a pure performance standpoint Edit: Man, I can't have shit. Intel can eat my ass. 15th gen better not have all these issues again
@@moomie1634 LOL best? Not sure what rock you're under but I guess you haven't heard about all 13/14th gen degradation, overheating and crashing? Regardless AMD is cheaper, generally better gaming performance, more efficient less power hungry and better upgrade path. There's hardly any reason to buy Intel.
The frightening thing is, even as Boeing shows so many potentially catastrophic technical problems, customers can do nothing about it. It's not like I, a third-world islands country citizen can choose what plane I want to fly with, more than 90% of our aviation industry uses Boeing. It's either Boeing or you take the ship (travel times between 3 days to weeks compared to hours), so when people suggest you to pray before you board the plane, you better pray hard.
@@MrWillypanda88 Firms from Canada, Brazil and more recently China do make aircraft. It's just that Boeing's effective lobbying on both sides of the Atlantic will ensure those firms never get the required licenses to provide full-fledged competition.
@@MrWillypanda88 your fears are a result of the idiots in the media overblowing anything to do with Boeing, whether it was their fault or not. The 737 MAX and a few of the latest 787s are Boeings ONLY aircraft affected, and will likely end up super safe as a result of so much scrutiny. The reality is, most Boeing aircraft you fly on are at least a decade old, built before Boeing truly fell to the low it is in today. The 737 is the most popular narrowbody aircraft on the market and has been for decades, issues like landing gear collapsing or fires are NOT boeings fault they are the airline's. As someone deeply interested in aviation and the Boeing scandals I dislike what Boeing has done to themselves, but I fucking DESPISE the media and how they use the company name for clicks. It's fearmongering and affects people like you when you really shouldn't worry
"When Apple was founded, we were 10 years ahead of the competition. What happened in 10 years? The competition caught up and Apple stood still. That is why Apple is failing." Steve Jobs, 1997
Apple could be innovative, if they put the proper cooling, repairability ram, storage etc into their products, and come on, they use the same low cost labor ad anyone else, lower those prices
They had "business" leaders instead of actual engineers and people who understand the industry. Unfortunately a problem with a lot of companies nowadays.
BK was a legit engineer. Hell, he used to brag about his overclocking exploits internally at Intel. Bob Swan, on the other hand, I was glad to not be around for. Many folks I talked to inside Intel thought Pat was a natural choice to be CEO even when I was there a decade ago, and were extremely happy to see him come back to take on the reigns. The company obviously has its challenges in front of it, but I think they understand that they need to balance out their investment on the semiconductor and software sides to get back in the good graces of consumers.
A “tool” maker??! TSMC wouldn’t exist without ASML’s Lithography Systems. Go check out what other Fabs are making while limited to Steppers and scanners from other legacy lithography and wafer production equipment manufacturing specialists. Nobody is even close. I mean technically if you want to get right down to it ASML is absolutely a chip manufacturer because they do indeed produce wafers and chips in the design and production’s processes of theirs machinery, but that’s not production for the outside of sales but it’s chip manufacturing nonetheless
I used to work as a Compiler Engineer in Intel for 3 years, and I was laid off last year, along with a bunch of other people. I can say that it was good while it lasted. I don't know anything about financial concerns, I just think that Intel is a decent place for employees to work.
I am sorry about your job, I hope you are doing alright. Even after they laid you off, you say it was a good place to work? No judgement, I wanted to ask what the culture was like and why you enjoyed it.
@@sagejpc1175 It was alright. I was paid well, I worked remotely, although sometimes I couldn't handle the assignments that they gave me, and I had to ask for extensions multiple times. My job was a rather difficult one. I was also provided very sparse documentation. A hands off mentality, figure it out on your own, can be both a good thing and a bad thing depending how you look at it. It was my first job out of uni too. They had no formal program for onboarding new hires, just explained how their development environment works, pick out a jira issue, and start working. Nothing bad happened there, and I was treated well.
I guess that's what 7 years of Sky Lake did to Intel. Just imagine - they were making CPUs with the same archtecture for 7 years straight and only increased frequency of their chips. Moreover, they could not move from 14 nm to 10 nm in 6 years until they made very questionable Ice Lake low power mobile chips. Bob Swan was a wrong CEO to choose, for sure.
Yea. That whole series of chips then were almost the same every gen. Then AMD swung hard at Intel. And they panicked lol. Slapped a couple more cores on the same junk and called it brilliant. It wasn't lol. Intel is still stuck on the same thought really. Super hot running chips. That's why I like AMD. They seem to be trying year after year to improve and the gains they have made since 2018 have been amazing! Can't say the same for Intel lol. Unless you count them also doubling as space heaters lmao. Won't miss using Intel in my pc when I build a new one. Plus I remember Intel never being able to move from 14nm to 10nm. And they basically just gave up on 10nm and went for 7nm. They clearly got way too comfortable to actually compete until they were forced to.
@braixeninfection6312 Intels stagnation was horrible. I cannot blame fully the EV node. It was both issues, failing to move past 14nm and getting 5% improvement year over year. They still had room for improvement. Yet, I own a 13900K and it's amazing! AMD launching Zen 5 was a horrible mess and the pricing was too high. I went for Intel because the pricing was a ton better. But my original choice was AMD. AMD sometimes it's obnoxiously expensive in some skus, that I don't find them worth it nor competitive vs Intel. And I am not talking about high end either. I5s are much better the price on Intel. 3D chips they launch them later to milk more the consumer. And AMD failed after Zen 3 to be more price sensitive. Except in server tho. Anyhow. when AMD raise the count of cores, to 24 I would buy them.
Intel fell behind because of one primary reason: They bet against ASML's EUV technology when it was first introduced. Intel believed that EUV systems are extremely complicated and expensive (which they are right about), but they also thought the yield on EUV systems would be too low for it to be worth investing in (They are wrong about this one however). Intel instead spent massive dollars on different patterning techniques such as multiple lithography to squeeze out smaller feature sizes in nanoelectronics with legacy DUV lithography systems. Eventual TSMC, and Samsung caught up as they were the first adopters of EUV technology. Intel actually has spent much more money in R&D than any semiconductor industry company including Nvidia, TSMC, AMD, Samsung, etc. Intel actually spends more than Nvidia, AMD, and TSMC combined, but they are constantly researching crazy technologies that do not pan out well like Intel Optane. They also were stubborn and continued to manufacture older 14 nm process ICs using their own foundries instead of using TSMC. Intel Chip designs are arguable much more innovative than AMD's, but they are stuck on an older process node. Only recently has Intel begun to use TSMC under Pat Gelsinger. Intel is on a come back. They will be the first to utilize ASML's new High NA EUV machines in their upcoming 20A, and 18A process nodes in late 2024. In addition, they will be introducing back power routing in those same nodes - something TSMC does not plan to introduce until 2026. These two technologies are gambles much like TSMC, and Samsung gambled in adopting EUV, but it could pan out well for Intel.
You are delusional, they have not spent much in investment. They pay highly experienced scientists lower than fresh graduates in the market with notions saying the engineers are getting to work with such a big brand.
The biggest mistake of Intel, by far, was it's failure to recognize the potential in smartphone SOC in the mid 2000. Apple was turned down by Intel, wich led the start of the development of their own SOC, wich became the foundation of the M cpu of today. This decision alone open the door for the ARM SOC era of today instead of X86 SOC in smartphone, to wich i'm pretty happy.
I don't think intel fail to recognize the potential but around that time intel also have issues with their primary bussiness. Netburst happen. So putting a resource towards what apple want back then can be big gamble and risk for them. Just look what happen with AMD around that time. They were very successful with their CPU and then they spend a fortune to acquite ATI for themselves. That move end up putting AMD in a world of hurt financinally a decade to come. Plus the mobile soc world also have it's own juggernaut.
I'm ex-Qualcomm with 12Y experience - Intel tried - TWICE - to purchase smartphone chip makers & startups. They ended up shutting BOTH purchases down! It takes lots of wireless system design skill to build smartphones and Intel has NONE. The real issue is the cost of new foundries was going up faster than any of their markets! They were trying to expand into new business areas so they could afford their next foundry. The many attempts at expansion from 2000 to 2020 they all failed. Intel was too selfish and too stubborn to accept outside chip designs for fabrication as a foundry. They felt it was beneath them. This arrogance was their doom ...
One of their biggest mistakes was not backing up from Pentium 4 fast enough, and then not learning and doing the same with Skylake. Apple M1 showed them the door, and they still didn't get it.
Dunno, their decision to not buy into openAI 7 years ago, thinking AI was kot the future might also be their biggest mistake? Or their choice to avoid EUV? It’s funny that there are so many bad mistakes to argue which was the worst.
Intels biggest problem wasn't the CEO it was Murthy Renduchintala who Krzanich poached from Qualcomm in 2015 with the express purpose of upgrading Intels design efforts and foundries. Pre 2011 Intel typically had 2 chips on each node before switching to the next, these changes happened every 2 years roughly. e.g. 32nm in 2009 for Westmere & Sandybridge, whilst AMD was still using 45nm. 22nm in 2011 for Ivybridge and Haswell, whilst AMD was now on 32nm. This was when the complacency started 3 years on a single node. AMD was on it's knees with sheare prices hovering between $2-5 per share and hitting an all time low of $1.81 by 2014. Intel was on 14nm in 2014 whilst AMD was now using 28nm, this was the start of the rot in their foundries, instead of just 2 CPU's per node we got this lot. Broadwell, Skylake, Kaby Lake, Coffee Lake, Cascade Lake, Comet Lake, 6 CPU models on a single node in 5 years, 2 years into this timeline they dropped their "Tick Tock" path of advancement, new architecture 1st year with a refinement the next (hence the 2 CPU lines per process node. we all remember the 14nm++++++++++ memes don't we? With 5-10% generational performance increases along with prices going through the roof. This really only happened as Intel took 5 years to get their 10nm node to produce high enough yields to be profitable, this is where Murthy comes into play.. He constantly was telling Krzanich "10nm is coming along just fine" whilst hiding the outrageously poor yields and even worse performance of the entire node. Then we get to 2016.. Intel launches the 6900K 8 core 16 thread CPU STILL on 14nm launched for $1089US, followed less than a year later by Zen 1, Both on 14nm but intel were on their own foundries whilst AMD were using Global Foundries or TSMC Intel had squandered their process node lead thanks to the 10nm delays and so, Zen arrived @$499 and beat Intel in non gaming benchmarks at less than half the price, yeah Intel still lead by a large margin in gaming in AOTS but a mere 3-20% in fps terms? For double the cost but no where near double the gaming performance? Intel compounded the issue by then launching the 10 core 20 thread 7900X still at a whopping $999US and AMD released the 2700X @ just $329US. Intel were relying on customer awareness to inflate costs (Not to mention dirty tactics in their past like forcing OEM's to not advertise AMD CPU's or risk losing access to Intel CPU's despite those AMD CPU's actually performing better).. As AMD closed the gap Intel stood still until Murthy was ousted in favour of Bob Swan, Intel wasn't helped by Murthy encouraging inter departmental infighting, Bob Swan ended this by basically telling dept head to stop pissing in other departments ponds or face the axe. Swan wasn't an engineer but he DID solve the internal strife at Intel before Gelsinger took over in 2021. Intel is STILL paying the price of Murthy's game playing, this is made worse by the current Intel fiasco over what is Intel "in spec" which is now seeing Intel HAVE to push out much lower wattage BIOS versions due to poor stability because they're THAT desperate for the win they'll do practically anything until they get found out... AMD have constantly innovated since Jim Keller rejoined them in 2012 to design Zen the infamously "glued together chips" which Intel derided so vociferously then desperately followed suit??? AMD also adopted Intel's previous tick tock regimen with 2 entire teams leapfrogging for each new major architecture, a policy reintroduced at Intel in 2021-22 after being abandoned in 2016 right in the midst of the 10nm debacle. Intel only has themselves to blame for their complacency arrogance and short-sightedness right when AMD were upping their game. Put it this way... In 2015 AMD's shares were at an all time low of just $1.82 whilst Intel were at $34.94, how are things now? Intel @ $30.51 down from a high of $68.26 over the last 3 years, whilst AMD? $150.56 and an absolute high of $227.3 in March this year. I gets worse when yo look at it this way... Intel's lead has fallen from 15 times the price per share in Q3 2015 and swung completely in the opposite to now being worth just 20% of AMD's current share prices.... This'll only become worse due to these new to the public stability issues which remained hidden for several years, iirc 11th 12th 13th and 14th gen are ALL affected and I wouldn't be too surprised to hear 10th gen are in there as well so that's 4-5 years of CPU's with instabilities cause by Intel chasing the gaming crown by pushing excessive watts and amps through their CPU's.
Holding on to foundries was a mistake. Intel was reluctant to spend billions to upgrade foundries every few years, it was not a one person decision. It was a decision made to save money by the leadership. Upgrading to smaller nodes keeps getting expensive and for producing chips economically they have to produce larger volumes. Intel had to produce 14nm chips more for them to get a decent ROI. AMD can choose the process based on the price they wanted to sell and not consider about the capital investment.
@@solo10041 Sorry but Intel only ever debated dumping the foundry business AFTER Murthy was fired, beforehand there was no point as they HAD the process advantage until Murthy screwed it all up and created an atmosphere of infighting at Intel. Bob Swan was brought in to lay down the law. Raja Koduri was brought in for the GPU division and saw the furore and thought he stood a chance of getting Murthy's job but left after Gelsinger got the job and actually started returning Intel to the top. With the news of Intel pushing their CPU's to unstable levels and claiming that's "in spec" becoming public in April Intels share prices have now tanked. The exact same thing happened back in the 2000's when Intel was caught out forcing OEM's to not use AMD chips despite them being faster. Intels share prices fell from an all time high of ~$70 to just $21 just after AMD filed the suit, now we have a similar situation where the instability news broke about a aonth ago and intels shares fell off a cliff losing 25% in the last 6 weeks.. The foundry losses of late are due to Murthy and his actions which allowed TSMC to catch and surpass intel's process advantage in 2015 which then forced intel to look to TSMC in order to match or beat AMD..
@@SRS13Rastus As per insider information Murthy had to step down due to delays in production of Intel's 7nm chips. There was a rumor that Intel was going to skip 10nm and directly for 7nm at which point maybe Murthy had to clarify as Intel's Chief Engineering Officer that 10 nm was coming along really well. But some fanboys think Intel should have gone straight to 7nm by ditching 10nm. If Intel had gone that route Incel fanboys could enjoy their 14nm chips and memes for much longer. Instead Murthy delivered 10nm process which has similar transistor density as compared TSMC 7nm process. But for Incel fanboys he is the Villain.
@@solo10041utter rubbish, Murthy was brought in in 2015 as even the 14nm Intel process was producing poor yields, 10nm didn't appear until 2019, Murthy was let go the following year. 7nm was just the nail in the coffin. 14nm yields were initially poor, Murthy was brought in to troubleshoot it and get intel back on track and begin the transition to 10nm and then to 7nm. Krzanich refused to believe AMD would be able to catch up as both Global Foundries and TSMC were at least an entire process node behind, especially considering Intel went from 32-22nm in 2011, whilst AMD with GF could only manage the switch from 32nm to 28nm as late as 2015. By 2014 Intel were already on 14nm and stayed there because Murthy fucked it up and lied about it which was why we got press releases for 3 or 4 years saying "10nm is coming along just fine and should arrive in the next year" when it clearly wasn't. When it DID arrive in 2018 the performance was so bad Intel HAD to release it in laptop only 1st, 10nm didn't come to desktop until 2021. 3 YEARS later... AMD had released Zen 1 on 14nm in 2017, Zen+ on 12nm in 2018 and Zen 2 in 2019 on 7nm. Intel had no choice but to switch to TSMC and abandon both 10nm and 7nm, Murthy was booted in 2020 as a result of turning Intels' foundries into a joke.
People think it was Intel getting stuck on 14nm. That misses the big picture that from 2000 to 2020 the cost of a new foundry was going up faster than any business that intel was in. The only way to survive in this scenario is to become a foundry and intel was too stubborn to accept outside chips for fabrication! Even Samsung knows this but Intel was too stubborn to ask outsiders for help (chip designs to fab)!
Technically, neither NVIDIA nor AMD are chipmakers. AMD chips are made by TSMC while NVIDIA chips are made my TSMC or Samsung. Intel make their own chips and can also make them for other companies. And given the geopolitical risks associated with Taiwan and potentially South Korea as well, Intel is basically the only advanced chip manufacturer in the "safe" parts of the world.
True the UA-camr is being in accurate. And Intel is far more than some of these other companies. Even though I think intel has stagnated let’s be least fair and elevate the company over others let’s be honest chips and x86 wouldn’t be a thing without Intel
Intel is a long way behind TSMC and it is behind Samsung as well. Part of their strategy is deploy over $100 billion in capital to build new plants to catch up. But this will take many years if it is possible at all, because TSMC and Samsung won't be standing still waiting to be overtaken by Intel. The second aspect is developing AI chips, which many companies are currently doing. Intel's latest entry is the Gaudi 3 which is attempting to take on Nvidia's latest and greatest, so far with very limited success. Intel needs this to succeed in the near term, otherwise they will be hemorrhaging money while the wait for profits to emerge from their chip making business...
Not really. For one, Intel quality is still there. Second, they didn't just get complacent, this video is kind of far off the mark. They bit off way more than they could chew in terms of what could be packed into their process updates multiple times. They simply failed to execute on the new and aggressive fab nodes. They tried and failed swallowing big bites while TSMC succeeded by nibbling at the problem.
@@chrisbradley3224 Yeah this doesn't get mentioned often enough. It's not like Intel stopped innovating, they actually have tons of interesting R&D/engineering going on, where they failed was in bringing those innovations to market in a timely manner for quite some time. And even when they did with things like optane for better or worse they gave up on it and additionally sold off their storage business. Not that Intel hasn't made some mistakes but it's not as simple as complacency and C level ineptitude.
Well said! While I was in school they make consumer 4 core chips. When I graduated college they were still making 4 core chips( mainstream). Then the next year AMD made 8 core chips mainstream.
@@ramengahauhnar8474 Don't forget how intel literally just took the same 4 core chips, slapped two more cores on and called it a day when Ryzen came out lol.
What happened to Intel is they pulled a Netburst, when they found their chips couldn't cut the mustard instead of going back to the drawing board they simply started cranking the voltage to get higher and higher clocks. It is exactly what they did with Prescott P4s when they found the Athlon was beating them on performance and ended up creating space heaters that still wasn't as good as the AMD offering only this time they have cranked the voltages so high the silicon itself is degrading.
@@braixeninfection6312 When De8aur has to go to all the effort of a delidding AND liquid metal for thermal paste AND a custom milled waterblock AND a custom pump rig that can circulate an absolutely insane amount of water and even with all that he can't keep his 14900k from thermal throttling? Yeah stick a fork the current Intel arch is done.
@@braixeninfection6312 The issue isn't the cooling capacity, but rather the energy transfer rate limit from a tiny sand block to the rest of the cooling system.
Intel has spent more than 20 years marketing clock speed as their primary indicator of performance. That’s what lead to “Netburst” in the P4, and it’s what has driven power usage to insane levels. The only real step away from that was the Tualatin/Core/Core2 designs initiating out of their Israel design center. That of course has formed the basis of all subsequent Intel chips, but since then (~2006) it’s mostly focused on clock speed, larger cache, and in the high end chips, more cores, all of which have increased power/heat. Performance of Intel x86 (per core) has roughly doubled in the last decade, but with it, power usage has increased significantly, only being brought back to similar levels by their 10nm (Intel 7) process. Being stuck on their 14nm node contributed to that significantly, but wasn’t the only factor. Now Intel is split, pushing lower power “efficiency” cores paired with very power hungry P cores pushed to even higher clock speeds, just so they can claim to be faster than AMD. Lowering the max clock speeds just 300-400MHz reduces performance by 3%-5% while reducing power/heat by more than 25%. Same old marketing driven pursuit of clock speed regardless of power (and apparently stability) concerns.
I got a job offer from Intel for GPU Design Engineer. I'll be working in Intel for about a year and then going for masters. I used to love Intel but it's insane to think they ignored all 4 big new markets. I can't fathom how naive they were
Man, Intel's a sinking ship right now. You made this right before Raptor Lake got exposed as horribly defective and unstable to the point they fail within months if used at factory settings.
The only reason amd came back was because they found loophole to get out of their contract with globalfoundries (thats why their chips sucked) and was able to make chips with tsmc.
Yeah apparently they were stuck with their Bulldozer and Piledriver multi core cpu architecture because of Global foundaries. A lot of programs and games still used single cores. That’s why Intel was so successful with their Duo2 and later “i”series. AMD could not afford to manufacture the newer architectures due to your mentioned contract. Jim Keller and TSMC saved AMD.
Hell no lol, AMD came back because they put time, resources and care into making a good CPU architecture......and AMD is still doing good now because they continued to pump resources into research and development with multiple leapfrogging teams because their CEO is an engineer that understands you need to have good product that people wants to buy and the only way you can make a good product is to spend resources into R&D.....
This is completely wrong the 1700 x and 2700 x were produced with global foundries. @12nm. Globalfoundries is amd's original foundry. They spun it off and then it slowed down. Excavator, Bulldozer and Piledriver were stupid designs that tried to share one ALU with two execution pipelines, a moronic idea. So when AMD said they had two threads per core they actually had only one, a 30% deficit in each core!
What happened is Intel turned its back on extreme UV lithography for short term profit and investor's dividends but then this not only meant losses in the long term but also putting the USA strategic advantage in chips on the back burner.
Back in 2020, I resumed building custom PCs, after a several year’s hiatus .. built 3 based on the AMD Ryzen 5 (Zen 2, 3rd Gen) processor (5-3600x). All are still performing better than expected today. Thx to AMD for making such impressive strides forward, then and currently .. when in actuality, I didn’t think they could.
Actually the dod just canceled their largest contract with Intel this year. They have determined that the exclusive contracts that they had in place with them were not nearly worth the money compared to AMD Epyc
Their adoption of branding over engineering paints a clear picture of why Intel's latest 13th & 14th generation CPUs are crashing. For the first time in Intel's history their biggest consumers, desktop users, are frustrated and angry at Intel. If they don't manage this situation well, it may be lead to irreparable damage to this relationship.
intel is the only american cpu maker, not just designer. let that sink in. without intel foundry, the usa is no longer independent. uncle sam will not let that happen.
@@iLegionaire3755AMD used to be manufacturer but is not now it separated from that unit and went to tsmc. Nvidia probably didnt have one since the start. At the moment it appears intel is only one.
I've built with both and I'm on an AM4 build right now. I've gone as high as I can really with the 5800X3D. Any improvement will come from a new board, memory, CPU. It's kind of sad what is going on right now. If I were looking to build right now it'd be AMD for sure. I got lucky with AM4. AM5 looks to be doing great. Intel right now is scary. I'd be hard pressed to buy anything from them right now.
12:08 they lost their reputation when they disabled undervolting on laptops. Something that increased battery life, decreased heat and noise without losing any performance. Since they did, I switched to AMD.
One major problem with Intel's business model is this: they try to compete in both Design and Manufacturing at the same time, but they fall behind on both.
They fall because of their stuborness, but not so long ago they were at the top of the game in Manufacturing, lets not forget Intel invented things like Finfet
I wouldn't say "without any experience", they've been making iGPUs for many years now. The main issue with Arc was the bad drivers, not bad performance of the GPUs themselves.
It's both. The A770 is supposed to compete with a 3070. It competes a whole class or two below it while having a larger die size. Alchemist is architecturally flawed to the point where it relies on ReBAR to even function.
@@TabalugaDragon intel used to put igpus in the chipset or alone in the motherboard, then it was inserted into the cpu socket when size and ram access made it a better idea, specially useful for laptops
Intel is actually investing in research and development instead of just lining the pockets of their shareholders. They have taken a risky move lately on cutting edge technology, and the speculators are not happy with it.
Marketing has kept them in tight collusion with the OEMs though. Imagine if the likes of Dell only cared about performance and energy costs, it would be a struggle to find an intel laptop on their site instead of the other way around. The partner program aka bribery that keeps intel pushed by msps, 20 years of collusion with microshit, they have no legitimate reason to be #1 today.
@@jamescarter8311 while the hype does exist, they never said a thing and all their cpus looked like a copy and paste for many years they still have that problem and their node is useles, tsmc is making their latests designs, intel fell behind, hard.
Intel simply ran into the "wet van de remmende voorsprong", which very loosely translates as "the law of contemplacency". Intel at a time was só far ahead of AMD, that they decided they could afford to take it easier on their development, thus saving hundreds of millions of dollars per year. Then AMD made a break-through with their Ryzen chips and for Intel that kind of came as a huge surprise and they've been struggling to catch up ever since. But Intel has made this mistake before and weirdly they don't seem to have learned much from that experience. Remember the time when AMD's XP CPUs wiped the floor with Intel's CPUs because AMD's CPUs were way more efficient than Intel's CPUs, so despite Intel having CPU's that worked on much higher clock frequencies, AMD's CPUs outperformed them. That seems to have happened again when AMD introduced their Rysen CPUs.
It's much more because of Zen chiplettes, not chips. It's all about the chip yield. First numbers I heard before TR's release was that they were getting 80-90%... which was sick. What do you think the yield is on a 28 core Xeon, all cores up and over spec? Single digits. And if just one of those 28 cores is a little weak, bzzt. Meanwhile AMD pulls four chips from their best bins and effortlessly spawns another perfect monster. This leads to Intel pulling all their top silicon they'd normally want $12,000 for, selling it for $2,000 and bleeding out the ass. Meanwhile AMD sells theirs for $1,000, outperforms, and still makes money. And still even now, Intel won't abandon monolithic architecture.
As someone who worked at Intel from 1991 to 2014, I can testify that the company generally has terrible management (both upper and lower). Their first major misfire on a computer chip design occured all the way back in the late 90's with the Willamette-based Pentium 4 .. a superpipelined monstrosity that was quickly replaced (right hand turn) with muti-core. That P4 design was the dead opposite of ARM SOC evolution, a screwup that rapidly became increasingly evident.
Intel, in some ways, tried to become a "lifestyle brand." In the early-mid-90s, AMD was seen as a low-cost chipmaker, making budget versions of Intel chips. By the end of the 90's, the K6 was a compelling choice, wooing gamers.
Forget about Moore's law mate, as it come to nanometer level microchip production that do not apply. It was not supposedly dying, the law was not a law the begin with. If you do not get that, make a research on how lithography works and how ASML operates for example.
They are not in a dire state... They're just a little behind but fighting to catch up and with very promising research being finalised to get them back on top. And thats while they still has very VERY deep pockets to invest on that.
@@Allen-L-Canada as an investor, I'm waiting to see how low they can go before buying up all stock I can to see them rise multiple hundreds of percentile points in a few years. Because RN they are undervalued by the market and once they start showing promising signs then the market will do what it does best and overvalued Intel probably too much, but hey, as long as it makes me money right? It's interesting to research about the companies m especially tech companies, the signs that they will do well usually come years before they actually do, because they release research papers and from there you can see thay are ahead of other companies or behind other companies, it's not what they are selling now but what they re researching and investing now to get the upper hand later and when they fall to such lows like this I find it a lucky point in time indeed as I can capitalize on that, just like how happened to AMD, altough AMD was much more of a risky move because they really were on the brink of bankruptcy.
@@marcosvictor4935 I see your main point: buy it now while it's at one of the lowest points. It might just work for you. For me, I have invested in it 4 years ago, and had high hope for Pat, who said Intel will turn around by 2024-2025, but now he said will be 2026-2030. I lost faith in him. I sold 90% of my holdings at a big lost, and cut my sunken cost. If I keep buying it now, then I would be buying for hope, not for high probability of turn-around. There is low probability or any certainty that Intel will catch up with TSMC and Samsung in manufacturing. They have never demonstrated they watched up anyone before. So buying on hope is probably speculation. I personal prefer to buy on high probability and predictability. You don't need to agree with me, I with you the best.
The only way for Intel to get back to what it was is to buy Hi EUV machines (as it is doing) from ASML. TSM and Samsung have not yet acquired that machine. Maybe that can save it and become a cutting-edge company again. ASML was the one that made TSM and Samsun grow. It all depends on it, the design is the least complicated.
In general the Intel GPUs are not bad for the price, Arc is really good it is only somewhat held back by general driver optimizations, especially the ARC A770
You missed a few major points. First, a big reason Intel did so well for so long is because AMD was either A) getting blocked from OEMs by Intel or B) Emptying a revolver into their foot repeatedly. When the whole market is a duopoly, and one of the major players is just absent, then of course Intel will dominate. Second, your PC market share data is outdated. Ever since Intel released their 12th, 13th, and 14th gen chips that far outperform AMD, Intel has been regaining market share, and recent data shows it's closer to 66/33 split in favor of Intel. Third, it's worth noting that, although PC/Computer sales have slowed, Data centers that run everything still need CPUs, and this market has blown up, and continues to grow. Intel also has a heavy market presence in storage, network, edge, and car chips through mobileye. Intel is also making big advancements into the GPU market, and controls nearly 5% of the GPU market. It's worth noting that a big reason why Intel never made these moves into gaming and smartphones is because it'd require massive restructuring in the way they make chips. For AMD, Nvidia, and Qualcomm it's really easy for them to shift into a new market. They design a new chip, send the designs over to TSMC, and then a few months later they get their chip batches done. For Intel, they not only need to design the chips, but also change the manufacturing process, which would require a lot of investment and time to put into fabs, and to Intel that just didn't make sense at the time. The old Intel really hated investing into manufacturing and such. Fourth, a big reason why crypto and Ai are so heavily favored towards Nvidia is because Nvidia has spend years investing in software, such as CUDA. CUDA is the biggest reason why Nvidia is so dominant in Ai currently, as AMD and Intel have both made hardware that is arguably much better than Nvidia's offerings at a cheaper price, but the software for each still lags behind. Software has been a long standing issue for Intel, as they haven't made major investments into the sector, and still underpay both software and hardware engineers, although this is likely to begin changing with Pat at the helm(When he was CEO of VMWare, SWE pay was much better than Intel currently pays). Fifth, the reason Intel is losing so much isn't because they're investing a lot, it's because revenue declined, but fixed costs stayed the same. If you read the 10k, you'd notice that the operating expenses over the last few years has all been at about 50 billion dollars a year, while revenue has dropped. This is mainly because consumers aren't buying as many computers currently due to already having bought a lot of tech during the pandemic, and now with fears of a recession people are buying less. This is combined with higher interest rates, which means companies aren't borrowing as much to upgrade servers, which has led to a massive slowdown in revenue there. Yes, companies are upgrading servers with Ai chips, but that's only the top companies, and the vast majority of companies aren't spending much on chips currently. Nvidia only needs a few players buying a ton of chips, but Intel needs a lot of different companies buying, simply because servers aren't much better simply by adding more chips. Luckily for Intel, both these trends are only temporary. It's already been noted that consumer demand is rising again, and a lot of people have been waiting for 15th gen to upgrade, which will be coming out later this year. Also, most servers are on a timetable to upgrade, and servers only last about 5 years MAX, meaning the 2020-2021 servers that were bought will need to be replaced, and Intel will likely be the ones reaping these rewards. The massive Capex actually isn't that much more than historical amounts, and the foundry business has always operated at a loss to serve the chipmaking part of the business, it's just that now, people are getting independent results about this and freaking out. Intel has already stated that their current capex will likely come down in the coming years, as their 18A fabs are getting finished, and those will have the most advanced chipmaking process in the world. It's also worth noting that the 2026 figure you stated isn't about Intel in general, but about Ai. Intel doesn't need to convince the market about CPUs, because they're already the market leader. There are still tons of people that will only buy Intel chips, because AMD before Ryzen was a gamble on whether you'd be getting a brick, an underperformer, or a thermal disaster. Intel's main spending right now is all about their future foundry business, which you didn't mention for some reason. This foundry business is projected to be second in the world by 2030, and the revenue from that alone should help drive the company forward. Overall, the company is in a better position than it's been in over a decade. Pat is easily one of the best leaders in the world, and what he did with VMWare should show people just how capable he is as a CEO. I have no doubts that 10 years from now, Intel will be one of the largest and best chip companies in the world. I'd say the video was good overall, and it made a lot of solid points about how Intel's past was markedly bad due to leadership while not ignoring the fact Intel has a strong future. It seems like most videos on Intel are only extrapolating past performance and suggesting that, even though the entire companies management and vision is different, that somehow the company will remain on this downward trajectory. I'd also like to see a video talking about their fab business and a deeper dive into that arm of the company. It has a very bright future, and with Pat at the helm, Intel will be a great investment for anyone willing to ride the wave.
On your second point - the Intel 12/13/14th gen chips don't outperform AMD, they mostly just match AMD. In some cases they outperform, but in many cases they also use a lot more power for the same performance, at least on higher loads.
@@lycanthoss The reason Intel chips use more power is because Intel engineers have given complete freedom to the user to overclock the CPU to death. On a pound for pound rating, Intel does beat AMD, but this is only slightly, and even with a power cap and under-volting, the performance slightly favors intel(on some applications, some still favor AMD). The two are very close, but we'll have to see what 15th/8th gen brings, and I really believe Intel is bringing their A-game, as compared to 10th and 11th gen, this turnaround in competitiveness is insane.
solid points but pat lied about the dividend although i give him credit for buying shares but investing in intel comes at the opportunity cost of investing in nvidia which jensen has a proven track record of being leagues above pat.
@@laCruz40oz nvidia is overvalued. Intel is undervalued. Pat doesn't have enough time to prove what he can do, while Jensen has had decades. It's not a fair comparison
As a current chip designer at Intel I can say that Pat is fighting on both external and internal sides. Intel middle management is a mess and we have been steadily losing top talent to companies like Apple, Cadence, Microsoft, Google and Amazon these last 3 years. A lot of them have been director and VP level as well as Principal Engineers. Retention goals have been set and we are still not even near on hitting retention targets. I do hope that Intel does well, but it seems like it will be due to becoming a foundry for all these other companies that are now creating their own chips. Middle management is still led by the same people that made Intel failed before. On the other hand, the older gen top tier engineers seem to be willing to retire as a lot of them have ~30 in the company and morale is so low at the individual contributor level.
Intel put a MBA in charge of a technology company and he completely failed to develop new smaller semiconductor fab technology. They went from tech leaders to being stuck on 14nm for a embarrassingly long time.
2 months after this video and Intel reputation is in a bottomless hole with with 2 generations of defective cpus by which is far worse than the series of security vulnerabilties from recent years....
Honestly,when they made their Pentium 2 they were at their peak then. Everything started to go downhill from there. If you ever saw Pentium 2 you would know what im talking about.
Investing in stocks is like bagging the ugly girl that’s gonna glow up after the summer break. Intel’s problem is that their fab was their sole customer which gave it slack. While TSMC was busy Fabing for nvidia, Apple and AMD so they had to earn it. But Intel remains the only US based major fab. It’s crazy that it’s valued under applied materials (who they constantly buy wafers from) and TI (who is stuck making sensors and components with processes from 2 decades ago). Intel is spending in opening their fab for everyone to use. I would not write them off
As someone who works in the automotive industry, I can tell you that while not the biggest player, intel is still in the game there thanks to Mobileye aquisition years ago. However they are not the biggest player there and often have issues with propper supply of chips (although it seems it's almost common for everyone).
It's still a shame what happened to Optane storage given it's price tag, low latency and extreme endurance. I strongly suggest you guys pick one up as a caching drive (NOT the tiny ones, the 256GB+ models) before they disappear forever. I got a 905p 1TB for my system and it slots perfectly into the last available 3.0 x4 slot :)
Time to buy Intel. People are way overestimating what AMD is capable of and way underestimating Intel. Everyone stumbles and the media loves to pile on.
Intel appointed Christoph Schell at the head of their sales division. Since then, sales plummeted, parts of their balance sheet went into negative, good products fail to sell. Not sure if the guy let go of technical talent who knew the products well. Intel must get better leadership who really understand the technology.
A few days ago Intel’s CEO was here in Costa Rica. I have a lot of friends whom work for them. They are not dead, far from it. Intel is probably getting super strong right now
Well, when I was working for Intel it seemed their priority is more regulatory issues, diversity and data security. I never have worked for a more formal US company. Furthermore, many acquisitions like Altera / FPGA and Infineon - Radio Networking - was not really successful. They did not really improve their acquisitions but just managed the status quo. Now they rebrand Altera again.
didn't even mention arm that is munching on their server business for many reasons. energy, ARM processors consume significantly less power compared to x86 chips, resulting in lower cooling requirements and reduced energy costs for data centers. By distributing workloads across a larger number of lower-power ARM cores instead of fewer high-power x86 cores, ARM servers can achieve better energy efficiency. High Density The system-on-chip (SoC) design of ARM processors allows for higher core densities, enabling more compute power in a smaller footprint. ARM servers like those from Gigabyte with up to 48 cores per chip can provide high performance while occupying less physical space in a data center. Cost-Effective Scaling ARM processors are generally less complex and cheaper to manufacture than x86 chips. This makes it more cost-effective to scale out ARM server farms by adding more nodes as needed, rather than scaling up with more powerful but expensive x86 servers. AWS (with Graviton), Alibaba, and others are already deploying ARM-based server instances in their data centers to leverage these benefits
This is just not going to be true. ARM chips have so many weaknesses. They can't run most x86 apps, they can't run processes as fast. The only reason x86 isn't being built by Qualcomm, Apple, and Nvidia is because it's open source and just requires royalties.
And that's why I sold my NVDA at $950 and shifted it to ARM. Damn I'm a clever yob eh? ;) The reason we haven't seen things like 3D memory integration is because that stacking = heat. I expect this will be less of a problem with ARM arch. Is compute IN memory possible with ARM? That would be a truly massive breakthrough.
Nvidia acquiring ARM & AMD developing ZEN architecture was an important move that affected Intel's fall, intel used to be good, but amd & ARM became more important
Even in gaming, Intel is losing ground quite fast. The X3D AMD CPUs are quite fast for gaming whilst consuming way less power. Meanwhile, Intel CPUs are offing themselves due to crazy power budgets. Then we have the datacenter, Epyc is slowly but surely eating Xeon's share. It's crazy to see a mega giant like Intel falling like this. From the king to the underdog in less of a decade. I can't wait to see Intel picking themselves up with their amazing technology. Maybe this is the final push that they need to achieve that.
Years ago I was told that Apple should just close up shop. This was right at the time Steve Jobs was returning to Apple. Intel is making all the right moves at this time.
7:48 Ever heard about Zen X3D? Intel is not competitive in gaming. They need insane power consumptions to even approach AMD its best gaming CPUs. Those CPUs are more expensive and require more expensive motherboards. On top of that, their 13th gen and 14th gen CPUs are experiencing stability issues or downright failing when used for gaming. The incoming fix seems to be to clamp down on power usage. Intel has allowed the motherboard vendors to push settings to the maximum to extract a few more percent performance out of those CPUs. And now that is backfiring. Intel only has laptops left and that is AMD its next push.
That's called deadwood in Engineering Departments. That's when you know the culture is rotten from the top down. It's going to require a complete top to bottom leadership change. I wish Intel luck not gonna be easy.
@nazar1744 Plus... they continue to fire people, 15k now.... bc they have to reduce the work force by 15%... and guess who makes these decisions.... the managers. So if you get along well with your manager, you're saved.
In the late 80's, early 90's there was a slew of Harvard business laws paper's which argued that Fortune 500 companies are best led by financier's and bankers over technical led companies. Since the Fortune 500 companies run in packs, this led to all the Fortune 500 companies putting in bankers into CEO positions. This move absolutely destroyed engineering type companies, though it took many years for this mistake to manifest itself. Companies such as Boeing, GE, AT&T, HP and now Intel have suffered when banker CEO types took over. Now, the hottest companies are technically led CEO companies. When you are a technical company, engineering development is a massively expensive area where the payback for engineering development isn't for 5 years or even more. The issue is, if you miss an engineering development product cycle it can/will have massive implications on your future product roadmap. If you are a banker type CEO, it is easy to see that your present yearly profit can be maximized by cutting engineering development and focus on marketing and finances. In truth you need a CEO or C's who master engineering development and financing/marketing NOT just financing/marketing.
The plans to extend intel's business model to include a silicon foundry open to 3rd parties will turn Intel into one of the most important players in the AI race. Almost 70% of Silicon foundries are in potential conflict zones Taiwan, South Korea, Israel. The US has been making moves to limit the access to advanced AI chips, once the US has the ability to manufacture them in the US (or EU) they will implement restrictions.
it will turn them from designers they once were into just a fab, again, intel beginings was like this, well, at least we will not get more snake oil, no more amd glue bs and no more bribes and criminal behavior, so it is something, because if amd put intel to make their chiplets intel must shut the f up
I think the complacence of the Intel stagnation era is somewhat overstated. They invested quite a lot into architectural improvements. Trouble is they came to focus heavily on new SIMD instructions, which are extremely efficient in theory, but difficult to work with in practice, so these innovations took a long time to seep into consumer applications. As a consumer you'd never notice if your CPU had the latest set of AVX 512 extensions. Meanwhile their EUV lithography got badly delayed, so they got stuck endlessly polishing their 14nm process node. So what we got was year after year of very modest performance gains, stagnant core counts and rapidly increasing power consumptions. Then off course there's Optane, the extremely fast and wear resistant persistent storage, which looked really promising when it was announced, but release got delayed, and flash memory evolved too rapidly, making optane pretty much pointless by the time it got into production. Not to say they weren't complacent! It's pretty stark how much more effort they put into real performance after AMD had clawed back their entire lead and then some. With no fire under their feet Intel management clearly weren't in a hurry to improve real world performance.
for me one of the biggest mistakes was abadonning of intel quark microcontrollers, they had very high clock rate and single PCI lane for very good price, but not only they closed resources behind wall, they never promote that in sufficient way, they made just opposite to what microchip is constantly doing, thats is spamming free just-download IDE, toolchains, programmers, dev-kits and all kind of support for everyone who wants to use their platform
Intel is going to need to pull off a miracle to turn things round. Once NVIDIA starts ramping up their CPU business, that will surely start to eat in further to their server market.
If the channel exists in 20 years, there'll be the exact same video about NVIDIA. Maybe Google, too. No one stays on top forever. Companies slow down, age & die just like their leaders do.
People say intel isn't innovative, but the truth is that they are capable of creating chips using the most cutting edge and advanced technologies of 2007
I mean, nowadays 14th gen is the best CPU on the market for consumers, and Intel is beating AMD across the board from a pure performance standpoint
Edit: Man, I can't have shit. Intel can eat my ass. 15th gen better not have all these issues again
@@moomie1634lol lmao. Sure at 4x the power.
@@moomie1634 LOL best? Not sure what rock you're under but I guess you haven't heard about all 13/14th gen degradation, overheating and crashing? Regardless AMD is cheaper, generally better gaming performance, more efficient less power hungry and better upgrade path. There's hardly any reason to buy Intel.
🤣🤣🤣
@@moomie1634 I mean, if you want a space heater in your room, sure.
Marketing over engineering is also Boeing slogan
Facts
The frightening thing is, even as Boeing shows so many potentially catastrophic technical problems, customers can do nothing about it. It's not like I, a third-world islands country citizen can choose what plane I want to fly with, more than 90% of our aviation industry uses Boeing. It's either Boeing or you take the ship (travel times between 3 days to weeks compared to hours), so when people suggest you to pray before you board the plane, you better pray hard.
@@MrWillypanda88 Firms from Canada, Brazil and more recently China do make aircraft. It's just that Boeing's effective lobbying on both sides of the Atlantic will ensure those firms never get the required licenses to provide full-fledged competition.
I said the Boeing acquisition of McDonald Douglas was a bad idea, for both companies and for the aviation industry.
@@MrWillypanda88 your fears are a result of the idiots in the media overblowing anything to do with Boeing, whether it was their fault or not. The 737 MAX and a few of the latest 787s are Boeings ONLY aircraft affected, and will likely end up super safe as a result of so much scrutiny. The reality is, most Boeing aircraft you fly on are at least a decade old, built before Boeing truly fell to the low it is in today. The 737 is the most popular narrowbody aircraft on the market and has been for decades, issues like landing gear collapsing or fires are NOT boeings fault they are the airline's. As someone deeply interested in aviation and the Boeing scandals I dislike what Boeing has done to themselves, but I fucking DESPISE the media and how they use the company name for clicks. It's fearmongering and affects people like you when you really shouldn't worry
"When Apple was founded, we were 10 years ahead of the competition. What happened in 10 years? The competition caught up and Apple stood still. That is why Apple is failing." Steve Jobs, 1997
Pretty much what happened with Intel
Apple could be innovative, if they put the proper cooling, repairability ram, storage etc into their products, and come on, they use the same low cost labor ad anyone else, lower those prices
@@Sykxezninnovative they realised doesn't earn them as much money, or rather, user expandability.
@@stuartburns8657 "What is the solution to Apple's problem? It's not to slash and burn. It's to innovate out of the problem." Steve Jobs, 1997
And this is sadly why Nvidia is dominating because the competition never caught up and Nvidia never stood still.
They had "business" leaders instead of actual engineers and people who understand the industry. Unfortunately a problem with a lot of companies nowadays.
the company went back to the hands of engineers though, its costing them, but i say they can come back from the bs corpos did.
BK was a legit engineer. Hell, he used to brag about his overclocking exploits internally at Intel. Bob Swan, on the other hand, I was glad to not be around for.
Many folks I talked to inside Intel thought Pat was a natural choice to be CEO even when I was there a decade ago, and were extremely happy to see him come back to take on the reigns. The company obviously has its challenges in front of it, but I think they understand that they need to balance out their investment on the semiconductor and software sides to get back in the good graces of consumers.
@@AdamPippertBK has never done any serious engineering work, even less so for the past 30 years at least.
Business leaders should be important, so they can place focus on people who matter. Many place focus on own survival instead.
Exactly. AMD is doing so well because Lisa Su started her career as an engineer and then moved into management.
ASML is NOT a chip maker. They are a chip tool maker!
chip maker maker : )
lol
yep
Same with applied materials
A “tool” maker??! TSMC wouldn’t exist without ASML’s Lithography Systems. Go check out what other Fabs are making while limited to Steppers and scanners from other legacy lithography and wafer production equipment manufacturing specialists.
Nobody is even close.
I mean technically if you want to get right down to it ASML is absolutely a chip manufacturer because they do indeed produce wafers and chips in the design and production’s processes of theirs machinery, but that’s not production for the outside of sales but it’s chip manufacturing nonetheless
Whatever state Intel is in right now, they are definitely not going bankrupt.
Government will bankroll it if needed, too critical
Neither is Blackberry.
Yeah, bankruptcy is unlikely, but definitely a fading giant.
@@LogicallyAnswered they have a market cap of 120B. That's quite a lot, I wouldn't consider it a fading giant when it is in most consumer laptops.
Short intel and Nio thank me later
I used to work as a Compiler Engineer in Intel for 3 years, and I was laid off last year, along with a bunch of other people. I can say that it was good while it lasted. I don't know anything about financial concerns, I just think that Intel is a decent place for employees to work.
Fair enough, sorry to hear that man
That's what you get for working for such a complacent company. I wouldn't be surprised if you worked for Peloton next.
I am sorry about your job, I hope you are doing alright. Even after they laid you off, you say it was a good place to work? No judgement, I wanted to ask what the culture was like and why you enjoyed it.
It tells you something about the company when staff who are involuntarily terminated say it was a good place to work, no hard feelings.
@@sagejpc1175 It was alright. I was paid well, I worked remotely, although sometimes I couldn't handle the assignments that they gave me, and I had to ask for extensions multiple times. My job was a rather difficult one. I was also provided very sparse documentation. A hands off mentality, figure it out on your own, can be both a good thing and a bad thing depending how you look at it. It was my first job out of uni too. They had no formal program for onboarding new hires, just explained how their development environment works, pick out a jira issue, and start working. Nothing bad happened there, and I was treated well.
I guess that's what 7 years of Sky Lake did to Intel. Just imagine - they were making CPUs with the same archtecture for 7 years straight and only increased frequency of their chips. Moreover, they could not move from 14 nm to 10 nm in 6 years until they made very questionable Ice Lake low power mobile chips. Bob Swan was a wrong CEO to choose, for sure.
Entire BoD needs to be fired.
Yea. That whole series of chips then were almost the same every gen. Then AMD swung hard at Intel. And they panicked lol. Slapped a couple more cores on the same junk and called it brilliant. It wasn't lol. Intel is still stuck on the same thought really. Super hot running chips. That's why I like AMD. They seem to be trying year after year to improve and the gains they have made since 2018 have been amazing! Can't say the same for Intel lol. Unless you count them also doubling as space heaters lmao. Won't miss using Intel in my pc when I build a new one. Plus I remember Intel never being able to move from 14nm to 10nm. And they basically just gave up on 10nm and went for 7nm. They clearly got way too comfortable to actually compete until they were forced to.
That's still their design strategy 😂
😅❤❤❤❤❤❤❤❤
@braixeninfection6312 Intels stagnation was horrible. I cannot blame fully the EV node. It was both issues, failing to move past 14nm and getting 5% improvement year over year. They still had room for improvement.
Yet, I own a 13900K and it's amazing! AMD launching Zen 5 was a horrible mess and the pricing was too high. I went for Intel because the pricing was a ton better. But my original choice was AMD.
AMD sometimes it's obnoxiously expensive in some skus, that I don't find them worth it nor competitive
vs Intel.
And I am not talking about high end either. I5s are much better the price on Intel. 3D chips they launch them later to milk more the consumer.
And AMD failed after Zen 3 to be more price sensitive. Except in server tho.
Anyhow. when AMD raise the count of cores, to 24 I would buy them.
Intel fell behind because of one primary reason: They bet against ASML's EUV technology when it was first introduced. Intel believed that EUV systems are extremely complicated and expensive (which they are right about), but they also thought the yield on EUV systems would be too low for it to be worth investing in (They are wrong about this one however). Intel instead spent massive dollars on different patterning techniques such as multiple lithography to squeeze out smaller feature sizes in nanoelectronics with legacy DUV lithography systems.
Eventual TSMC, and Samsung caught up as they were the first adopters of EUV technology. Intel actually has spent much more money in R&D than any semiconductor industry company including Nvidia, TSMC, AMD, Samsung, etc. Intel actually spends more than Nvidia, AMD, and TSMC combined, but they are constantly researching crazy technologies that do not pan out well like Intel Optane. They also were stubborn and continued to manufacture older 14 nm process ICs using their own foundries instead of using TSMC. Intel Chip designs are arguable much more innovative than AMD's, but they are stuck on an older process node. Only recently has Intel begun to use TSMC under Pat Gelsinger.
Intel is on a come back. They will be the first to utilize ASML's new High NA EUV machines in their upcoming 20A, and 18A process nodes in late 2024. In addition, they will be introducing back power routing in those same nodes - something TSMC does not plan to introduce until 2026. These two technologies are gambles much like TSMC, and Samsung gambled in adopting EUV, but it could pan out well for Intel.
Learning from past mistakes they are aggressively investing in High-na euv.
😂
Well said, i like their shift in focus to become a foundry. I look forward to seeing them grow their customer base.
@@nick-dogg Them trying to become a foundry is mainly due to Washington pushing them, as it tries to wean off Taiwan.
You are delusional, they have not spent much in investment. They pay highly experienced scientists lower than fresh graduates in the market with notions saying the engineers are getting to work with such a big brand.
Intel bankrolled ASML to create EUV machines. They own a significant stake in ASML as well. Nice try with revisionist history, though.
The biggest mistake of Intel, by far, was it's failure to recognize the potential in smartphone SOC in the mid 2000. Apple was turned down by Intel, wich led the start of the development of their own SOC, wich became the foundation of the M cpu of today. This decision alone open the door for the ARM SOC era of today instead of X86 SOC in smartphone, to wich i'm pretty happy.
I don't think intel fail to recognize the potential but around that time intel also have issues with their primary bussiness. Netburst happen. So putting a resource towards what apple want back then can be big gamble and risk for them. Just look what happen with AMD around that time. They were very successful with their CPU and then they spend a fortune to acquite ATI for themselves. That move end up putting AMD in a world of hurt financinally a decade to come. Plus the mobile soc world also have it's own juggernaut.
I'm ex-Qualcomm with 12Y experience - Intel tried - TWICE - to purchase smartphone chip makers & startups. They ended up shutting BOTH purchases down! It takes lots of wireless system design skill to build smartphones and Intel has NONE. The real issue is the cost of new foundries was going up faster than any of their markets! They were trying to expand into new business areas so they could afford their next foundry. The many attempts at expansion from 2000 to 2020 they all failed. Intel was too selfish and too stubborn to accept outside chip designs for fabrication as a foundry. They felt it was beneath them. This arrogance was their doom ...
it also didn't help that the few x86 mobile devices that existed were straight up e-waste
One of their biggest mistakes was not backing up from Pentium 4 fast enough, and then not learning and doing the same with Skylake. Apple M1 showed them the door, and they still didn't get it.
Dunno, their decision to not buy into openAI 7 years ago, thinking AI was kot the future might also be their biggest mistake? Or their choice to avoid EUV? It’s funny that there are so many bad mistakes to argue which was the worst.
Intels biggest problem wasn't the CEO it was Murthy Renduchintala who Krzanich poached from Qualcomm in 2015 with the express purpose of upgrading Intels design efforts and foundries.
Pre 2011 Intel typically had 2 chips on each node before switching to the next, these changes happened every 2 years roughly.
e.g. 32nm in 2009 for Westmere & Sandybridge, whilst AMD was still using 45nm.
22nm in 2011 for Ivybridge and Haswell, whilst AMD was now on 32nm. This was when the complacency started 3 years on a single node. AMD was on it's knees with sheare prices hovering between $2-5 per share and hitting an all time low of $1.81 by 2014.
Intel was on 14nm in 2014 whilst AMD was now using 28nm, this was the start of the rot in their foundries, instead of just 2 CPU's per node we got this lot. Broadwell, Skylake, Kaby Lake, Coffee Lake, Cascade Lake, Comet Lake, 6 CPU models on a single node in 5 years, 2 years into this timeline they dropped their "Tick Tock" path of advancement, new architecture 1st year with a refinement the next (hence the 2 CPU lines per process node. we all remember the 14nm++++++++++ memes don't we? With 5-10% generational performance increases along with prices going through the roof.
This really only happened as Intel took 5 years to get their 10nm node to produce high enough yields to be profitable, this is where Murthy comes into play..
He constantly was telling Krzanich "10nm is coming along just fine" whilst hiding the outrageously poor yields and even worse performance of the entire node. Then we get to 2016..
Intel launches the 6900K 8 core 16 thread CPU STILL on 14nm launched for $1089US, followed less than a year later by Zen 1, Both on 14nm but intel were on their own foundries whilst AMD were using Global Foundries or TSMC
Intel had squandered their process node lead thanks to the 10nm delays and so, Zen arrived @$499 and beat Intel in non gaming benchmarks at less than half the price, yeah Intel still lead by a large margin in gaming in AOTS but a mere 3-20% in fps terms? For double the cost but no where near double the gaming performance?
Intel compounded the issue by then launching the 10 core 20 thread 7900X still at a whopping $999US and AMD released the 2700X @ just $329US.
Intel were relying on customer awareness to inflate costs (Not to mention dirty tactics in their past like forcing OEM's to not advertise AMD CPU's or risk losing access to Intel CPU's despite those AMD CPU's actually performing better)..
As AMD closed the gap Intel stood still until Murthy was ousted in favour of Bob Swan, Intel wasn't helped by Murthy encouraging inter departmental infighting, Bob Swan ended this by basically telling dept head to stop pissing in other departments ponds or face the axe.
Swan wasn't an engineer but he DID solve the internal strife at Intel before Gelsinger took over in 2021.
Intel is STILL paying the price of Murthy's game playing, this is made worse by the current Intel fiasco over what is Intel "in spec" which is now seeing Intel HAVE to push out much lower wattage BIOS versions due to poor stability because they're THAT desperate for the win they'll do practically anything until they get found out...
AMD have constantly innovated since Jim Keller rejoined them in 2012 to design Zen the infamously "glued together chips" which Intel derided so vociferously then desperately followed suit???
AMD also adopted Intel's previous tick tock regimen with 2 entire teams leapfrogging for each new major architecture, a policy reintroduced at Intel in 2021-22 after being abandoned in 2016 right in the midst of the 10nm debacle.
Intel only has themselves to blame for their complacency arrogance and short-sightedness right when AMD were upping their game.
Put it this way... In 2015 AMD's shares were at an all time low of just $1.82 whilst Intel were at $34.94, how are things now? Intel @ $30.51 down from a high of $68.26 over the last 3 years, whilst AMD? $150.56 and an absolute high of $227.3 in March this year.
I gets worse when yo look at it this way... Intel's lead has fallen from 15 times the price per share in Q3 2015 and swung completely in the opposite to now being worth just 20% of AMD's current share prices....
This'll only become worse due to these new to the public stability issues which remained hidden for several years, iirc 11th 12th 13th and 14th gen are ALL affected and I wouldn't be too surprised to hear 10th gen are in there as well so that's 4-5 years of CPU's with instabilities cause by Intel chasing the gaming crown by pushing excessive watts and amps through their CPU's.
Holding on to foundries was a mistake. Intel was reluctant to spend billions to upgrade foundries every few years, it was not a one person decision. It was a decision made to save money by the leadership. Upgrading to smaller nodes keeps getting expensive and for producing chips economically they have to produce larger volumes. Intel had to produce 14nm chips more for them to get a decent ROI. AMD can choose the process based on the price they wanted to sell and not consider about the capital investment.
@@solo10041 Sorry but Intel only ever debated dumping the foundry business AFTER Murthy was fired, beforehand there was no point as they HAD the process advantage until Murthy screwed it all up and created an atmosphere of infighting at Intel.
Bob Swan was brought in to lay down the law.
Raja Koduri was brought in for the GPU division and saw the furore and thought he stood a chance of getting Murthy's job but left after Gelsinger got the job and actually started returning Intel to the top.
With the news of Intel pushing their CPU's to unstable levels and claiming that's "in spec" becoming public in April Intels share prices have now tanked.
The exact same thing happened back in the 2000's when Intel was caught out forcing OEM's to not use AMD chips despite them being faster.
Intels share prices fell from an all time high of ~$70 to just $21 just after AMD filed the suit, now we have a similar situation where the instability news broke about a aonth ago and intels shares fell off a cliff losing 25% in the last 6 weeks..
The foundry losses of late are due to Murthy and his actions which allowed TSMC to catch and surpass intel's process advantage in 2015 which then forced intel to look to TSMC in order to match or beat AMD..
@@SRS13Rastus As per insider information Murthy had to step down due to delays in production of Intel's 7nm chips.
There was a rumor that Intel was going to skip 10nm and directly for 7nm at which point maybe Murthy had to clarify as Intel's Chief Engineering Officer that 10 nm was coming along really well.
But some fanboys think Intel should have gone straight to 7nm by ditching 10nm. If Intel had gone that route Incel fanboys could enjoy their 14nm chips and memes for much longer.
Instead Murthy delivered 10nm process which has similar transistor density as compared TSMC 7nm process. But for Incel fanboys he is the Villain.
@@solo10041utter rubbish, Murthy was brought in in 2015 as even the 14nm Intel process was producing poor yields, 10nm didn't appear until 2019, Murthy was let go the following year. 7nm was just the nail in the coffin.
14nm yields were initially poor, Murthy was brought in to troubleshoot it and get intel back on track and begin the transition to 10nm and then to 7nm.
Krzanich refused to believe AMD would be able to catch up as both Global Foundries and TSMC were at least an entire process node behind, especially considering Intel went from 32-22nm in 2011, whilst AMD with GF could only manage the switch from 32nm to 28nm as late as 2015.
By 2014 Intel were already on 14nm and stayed there because Murthy fucked it up and lied about it which was why we got press releases for 3 or 4 years saying "10nm is coming along just fine and should arrive in the next year" when it clearly wasn't.
When it DID arrive in 2018 the performance was so bad Intel HAD to release it in laptop only 1st, 10nm didn't come to desktop until 2021. 3 YEARS later...
AMD had released Zen 1 on 14nm in 2017, Zen+ on 12nm in 2018 and Zen 2 in 2019 on 7nm.
Intel had no choice but to switch to TSMC and abandon both 10nm and 7nm, Murthy was booted in 2020 as a result of turning Intels' foundries into a joke.
People think it was Intel getting stuck on 14nm. That misses the big picture that from 2000 to 2020 the cost of a new foundry was going up faster than any business that intel was in. The only way to survive in this scenario is to become a foundry and intel was too stubborn to accept outside chips for fabrication! Even Samsung knows this but Intel was too stubborn to ask outsiders for help (chip designs to fab)!
AMD's Ryzen lineup were such a breath of fresh air to the industry.
well AMD was pretty much dead compeny when first ryzen came, out, they had to beg companies to make motherboards for it,...
Technically, neither NVIDIA nor AMD are chipmakers. AMD chips are made by TSMC while NVIDIA chips are made my TSMC or Samsung. Intel make their own chips and can also make them for other companies. And given the geopolitical risks associated with Taiwan and potentially South Korea as well, Intel is basically the only advanced chip manufacturer in the "safe" parts of the world.
EXACTLY thank you. Comparing CPUs to gpus is like comparing apples with bananas. Boath are important at the end
I can imagine China Invading Taiwan to ban gaming chips to use in their war machine or spyware.
I think he meant yo say chip designers, in which case he should have excluded TSMC and included Apple instead
True the UA-camr is being in accurate. And Intel is far more than some of these other companies. Even though I think intel has stagnated let’s be least fair and elevate the company over others let’s be honest chips and x86 wouldn’t be a thing without Intel
Intel is a long way behind TSMC and it is behind Samsung as well. Part of their strategy is deploy over $100 billion in capital to build new plants to catch up. But this will take many years if it is possible at all, because TSMC and Samsung won't be standing still waiting to be overtaken by Intel. The second aspect is developing AI chips, which many companies are currently doing. Intel's latest entry is the Gaudi 3 which is attempting to take on Nvidia's latest and greatest, so far with very limited success. Intel needs this to succeed in the near term, otherwise they will be hemorrhaging money while the wait for profits to emerge from their chip making business...
So basically, Intel pulled a Boeing.
Yep
same complacency.
Not really. For one, Intel quality is still there. Second, they didn't just get complacent, this video is kind of far off the mark. They bit off way more than they could chew in terms of what could be packed into their process updates multiple times. They simply failed to execute on the new and aggressive fab nodes. They tried and failed swallowing big bites while TSMC succeeded by nibbling at the problem.
@@chrisbradley3224 Yeah this doesn't get mentioned often enough. It's not like Intel stopped innovating, they actually have tons of interesting R&D/engineering going on, where they failed was in bringing those innovations to market in a timely manner for quite some time. And even when they did with things like optane for better or worse they gave up on it and additionally sold off their storage business. Not that Intel hasn't made some mistakes but it's not as simple as complacency and C level ineptitude.
Like Boeing (and many other U.S companies), they decided that investing in stock buy-backs was more important than investing in there core products.
I still remember the Intel CEO laughing and joking about the new "Ryzen" CPU from AMD at the time....
When you laugh at others product shows arrogance and lack of respect to others genuine work. It is there the downfall starts!
Maybe if Intel wasn't giving us 4 core CPU's for 10 years straight they wouldn't get caught with their pants down when Zen came out? 😂😂😂
Exactly. And it was the same quad core for 10 years.
Facts
Well, actually not the same but very similair...
Well said! While I was in school they make consumer 4 core chips. When I graduated college they were still making 4 core chips( mainstream). Then the next year AMD made 8 core chips mainstream.
@@ramengahauhnar8474 Don't forget how intel literally just took the same 4 core chips, slapped two more cores on and called it a day when Ryzen came out lol.
Google is on a similar path. 😢
Agreed, they’re where Intel was 20-30 years ago - beginning of stagnation
True. Everything new they try to invent none of them are as useful as they should be. Even though they are a billion doller company.
They are not trying to invent something useful enough. No creative tool for Daily use. They are running on ads.
nahhh google is pure monopoly of the internet.... your comment is on google....
@@thepro08so what, people still use computer with intel chips.
What happened to Intel is they pulled a Netburst, when they found their chips couldn't cut the mustard instead of going back to the drawing board they simply started cranking the voltage to get higher and higher clocks. It is exactly what they did with Prescott P4s when they found the Athlon was beating them on performance and ended up creating space heaters that still wasn't as good as the AMD offering only this time they have cranked the voltages so high the silicon itself is degrading.
Silicon can't run at 100*c all day. But Intel thinks otherwise lol. These things are so hard to keep cool it's ridiculous really.
@@braixeninfection6312 When De8aur has to go to all the effort of a delidding AND liquid metal for thermal paste AND a custom milled waterblock AND a custom pump rig that can circulate an absolutely insane amount of water and even with all that he can't keep his 14900k from thermal throttling? Yeah stick a fork the current Intel arch is done.
@@TheHangarHobbit That amount of heat is pure insanity and to still be thermally constrained with all that cooling.
@@braixeninfection6312 The issue isn't the cooling capacity, but rather the energy transfer rate limit from a tiny sand block to the rest of the cooling system.
Intel has spent more than 20 years marketing clock speed as their primary indicator of performance. That’s what lead to “Netburst” in the P4, and it’s what has driven power usage to insane levels. The only real step away from that was the Tualatin/Core/Core2 designs initiating out of their Israel design center. That of course has formed the basis of all subsequent Intel chips, but since then (~2006) it’s mostly focused on clock speed, larger cache, and in the high end chips, more cores, all of which have increased power/heat.
Performance of Intel x86 (per core) has roughly doubled in the last decade, but with it, power usage has increased significantly, only being brought back to similar levels by their 10nm (Intel 7) process. Being stuck on their 14nm node contributed to that significantly, but wasn’t the only factor. Now Intel is split, pushing lower power “efficiency” cores paired with very power hungry P cores pushed to even higher clock speeds, just so they can claim to be faster than AMD. Lowering the max clock speeds just 300-400MHz reduces performance by 3%-5% while reducing power/heat by more than 25%. Same old marketing driven pursuit of clock speed regardless of power (and apparently stability) concerns.
Well, this didn't age well... Things got SO much worse.
It aged pretty well. What are you talking about? He’s shitting on Intel
@@dannytran1587 Intel is for people who have too much money and too little common sense.
@@dannytran1587 rip English, lol?
The charging issues with 13th and 14th generation Intel chips are telling of what the problems they are having now.
I got a job offer from Intel for GPU Design Engineer. I'll be working in Intel for about a year and then going for masters. I used to love Intel but it's insane to think they ignored all 4 big new markets. I can't fathom how naive they were
And how is it?
Man, Intel's a sinking ship right now. You made this right before Raptor Lake got exposed as horribly defective and unstable to the point they fail within months if used at factory settings.
The only reason amd came back was because they found loophole to get out of their contract with globalfoundries (thats why their chips sucked) and was able to make chips with tsmc.
Yeah apparently they were stuck with their Bulldozer and Piledriver multi core cpu architecture because of Global foundaries. A lot of programs and games still used single cores. That’s why Intel was so successful with their Duo2 and later “i”series. AMD could not afford to manufacture the newer architectures due to your mentioned contract. Jim Keller and TSMC saved AMD.
Sad things is global foundaries is AMD sub company... Before it broke off..
Hell no lol, AMD came back because they put time, resources and care into making a good CPU architecture......and AMD is still doing good now because they continued to pump resources into research and development with multiple leapfrogging teams because their CEO is an engineer that understands you need to have good product that people wants to buy and the only way you can make a good product is to spend resources into R&D.....
This is completely wrong the 1700 x and 2700 x were produced with global foundries. @12nm. Globalfoundries is amd's original foundry. They spun it off and then it slowed down. Excavator, Bulldozer and Piledriver were stupid designs that tried to share one ALU with two execution pipelines, a moronic idea. So when AMD said they had two threads per core they actually had only one, a 30% deficit in each core!
Too many layers of middle management. Lack of innovation, old thinking. Too focused on stock price.
It hurt them long term, now they need to rework what they're clearly failing at
In the 90s ppl use to ask "What computer do you have? Pentium 1, 2 or 3?"
Ah yes 586, so much Doom and DN3D.
What happened is Intel turned its back on extreme UV lithography for short term profit and investor's dividends but then this not only meant losses in the long term but also putting the USA strategic advantage in chips on the back burner.
imagine fading ASML.
and now they are putting all their chips into the high NA EUV pot. It might save them, had they not laid off half their research and development team
UserBenchMark’s reviewers won’t like this video 💀
or WILL they? DUN! DUN! DUN!
no
Userbenchmark has fallen out of grace and credibility since AMD launched the first generation of Ryzen back in 2017.
Back in 2020, I resumed building custom PCs, after a several year’s hiatus .. built 3 based on the AMD Ryzen 5 (Zen 2, 3rd Gen) processor (5-3600x).
All are still performing better than expected today. Thx to AMD for making such impressive strides forward, then and currently .. when in actuality, I didn’t think they could.
They're a military contractor, the Pentagon won't allow them to disappear.
Actually the dod just canceled their largest contract with Intel this year. They have determined that the exclusive contracts that they had in place with them were not nearly worth the money compared to AMD Epyc
They are perhaps the only US chip maker. Others are chip designers.
@@SilverKnightPCs They just got a $3B deal from military.
Their adoption of branding over engineering paints a clear picture of why Intel's latest 13th & 14th generation CPUs are crashing. For the first time in Intel's history their biggest consumers, desktop users, are frustrated and angry at Intel. If they don't manage this situation well, it may be lead to irreparable damage to this relationship.
intel is the only american cpu maker, not just designer. let that sink in. without intel foundry, the usa is no longer independent. uncle sam will not let that happen.
AMD was founded in Sunnyvale, California on May 1, 1969. Intel and AMD are both American CPU manufacturers.
@@iLegionaire3755 dude they said not just designer. AMD does not own any fabs in the US. All of AMDs chips are manufactured by TSMC in Taiwan
@@iLegionaire3755AMD used to be manufacturer but is not now it separated from that unit and went to tsmc.
Nvidia probably didnt have one since the start.
At the moment it appears intel is only one.
@@iLegionaire3755 AMD stopped making their own chips 20 years ago.
I still remember the intel fanboys back when AMD first released the FX-series CPU's.
Oh how the tables have turned. 😂👍
The turns have truly tabled...
Lol@@auritro3903
I've built with both and I'm on an AM4 build right now. I've gone as high as I can really with the 5800X3D. Any improvement will come from a new board, memory, CPU. It's kind of sad what is going on right now. If I were looking to build right now it'd be AMD for sure. I got lucky with AM4. AM5 looks to be doing great. Intel right now is scary. I'd be hard pressed to buy anything from them right now.
Mhm.
Intel is not actually "down" 57% (pre inflation) since 2000 peak. You need to add back dividends. It's down about 17% (still very bad tho)
12:08 they lost their reputation when they disabled undervolting on laptops. Something that increased battery life, decreased heat and noise without losing any performance. Since they did, I switched to AMD.
lol
@@SubLimation7 something funny about it?
they disabled undervolting after it was discovered that you could basically root any system by playing with voltage
@@AG-sx9ws and yet it's still allowed on HK and HX. This was never about security, this was about greed.
Clickbait thumb at it finest
Gotta get that views lol
how?
@@lightninghell4 Maybe it was a different thumbnail a few hours ago. I didn't notice any clickbait with this current one.
There’s no clickbait
@@lightninghell4 They aren't at any risk of going bankrupt any time soon.
One major problem with Intel's business model is this: they try to compete in both Design and Manufacturing at the same time, but they fall behind on both.
They fall because of their stuborness, but not so long ago they were at the top of the game in Manufacturing, lets not forget Intel invented things like Finfet
lol what?? real men have fabs???????
Simple……..Intel sat on their hands and chose not to innovate. Don’t need a masters degree to figure that one out.
I wouldn't say "without any experience", they've been making iGPUs for many years now. The main issue with Arc was the bad drivers, not bad performance of the GPUs themselves.
years? decades, since the i810 chipset they offered gpus made by them and that was back in 1998, closer to 3 decades now
@@betag24cn I didn't know how many years exactly hence why I said what I said. I saw their first iGPU in my first laptop in 2011.
It's both. The A770 is supposed to compete with a 3070. It competes a whole class or two below it while having a larger die size. Alchemist is architecturally flawed to the point where it relies on ReBAR to even function.
@@riven4121 while having more vram per dollar than Nvidia and having better ray tracing and upscaling than AMD does.
@@TabalugaDragon intel used to put igpus in the chipset or alone in the motherboard, then it was inserted into the cpu socket when size and ram access made it a better idea, specially useful for laptops
Nvidia ate Intel's breakfast.
AMD ate Intel's lunch.
TSMC is eating Intel's dinner.
This is one of the best and most accurate comments I've read today.
Intel is eating Intel itself.
If china sets foot in Taiwan all 3 of they will be back to US govt/Intel support
And the customers ate their own boots.
Intel is actually investing in research and development instead of just lining the pockets of their shareholders. They have taken a risky move lately on cutting edge technology, and the speculators are not happy with it.
invest is not tye same as produce, look at 14nm++++++++++++++
Marketing has kept them in tight collusion with the OEMs though. Imagine if the likes of Dell only cared about performance and energy costs, it would be a struggle to find an intel laptop on their site instead of the other way around. The partner program aka bribery that keeps intel pushed by msps, 20 years of collusion with microshit, they have no legitimate reason to be #1 today.
@@betag24cn While they were developing 7nm. They never actually fell behind. It was all media hype.
@@jamescarter8311 while the hype does exist, they never said a thing and all their cpus looked like a copy and paste for many years
they still have that problem and their node is useles, tsmc is making their latests designs, intel fell behind, hard.
Intel simply ran into the "wet van de remmende voorsprong", which very loosely translates as "the law of contemplacency". Intel at a time was só far ahead of AMD, that they decided they could afford to take it easier on their development, thus saving hundreds of millions of dollars per year. Then AMD made a break-through with their Ryzen chips and for Intel that kind of came as a huge surprise and they've been struggling to catch up ever since.
But Intel has made this mistake before and weirdly they don't seem to have learned much from that experience. Remember the time when AMD's XP CPUs wiped the floor with Intel's CPUs because AMD's CPUs were way more efficient than Intel's CPUs, so despite Intel having CPU's that worked on much higher clock frequencies, AMD's CPUs outperformed them. That seems to have happened again when AMD introduced their Rysen CPUs.
It's much more because of Zen chiplettes, not chips. It's all about the chip yield. First numbers I heard before TR's release was that they were getting 80-90%... which was sick. What do you think the yield is on a 28 core Xeon, all cores up and over spec? Single digits. And if just one of those 28 cores is a little weak, bzzt. Meanwhile AMD pulls four chips from their best bins and effortlessly spawns another perfect monster. This leads to Intel pulling all their top silicon they'd normally want $12,000 for, selling it for $2,000 and bleeding out the ass. Meanwhile AMD sells theirs for $1,000, outperforms, and still makes money. And still even now, Intel won't abandon monolithic architecture.
As someone who worked at Intel from 1991 to 2014, I can testify that the company generally has terrible management (both upper and lower). Their first major misfire on a computer chip design occured all the way back in the late 90's with the Willamette-based Pentium 4 .. a superpipelined monstrosity that was quickly replaced (right hand turn) with muti-core. That P4 design was the dead opposite of ARM SOC evolution, a screwup that rapidly became increasingly evident.
Intel, in some ways, tried to become a "lifestyle brand." In the early-mid-90s, AMD was seen as a low-cost chipmaker, making budget versions of Intel chips. By the end of the 90's, the K6 was a compelling choice, wooing gamers.
Forget about Moore's law mate, as it come to nanometer level microchip production that do not apply. It was not supposedly dying, the law was not a law the begin with. If you do not get that, make a research on how lithography works and how ASML operates for example.
They are not in a dire state... They're just a little behind but fighting to catch up and with very promising research being finalised to get them back on top. And thats while they still has very VERY deep pockets to invest on that.
Yup this video makes intel sound like an underdog of some kind, an underdog with 76% PC market of CPUs xD
As an investor, I would stay away from high capex-intensive company like Intel.
@@Allen-L-Canada as an investor, I'm waiting to see how low they can go before buying up all stock I can to see them rise multiple hundreds of percentile points in a few years. Because RN they are undervalued by the market and once they start showing promising signs then the market will do what it does best and overvalued Intel probably too much, but hey, as long as it makes me money right?
It's interesting to research about the companies m especially tech companies, the signs that they will do well usually come years before they actually do, because they release research papers and from there you can see thay are ahead of other companies or behind other companies, it's not what they are selling now but what they re researching and investing now to get the upper hand later and when they fall to such lows like this I find it a lucky point in time indeed as I can capitalize on that, just like how happened to AMD, altough AMD was much more of a risky move because they really were on the brink of bankruptcy.
@@marcosvictor4935 I see your main point: buy it now while it's at one of the lowest points. It might just work for you. For me, I have invested in it 4 years ago, and had high hope for Pat, who said Intel will turn around by 2024-2025, but now he said will be 2026-2030. I lost faith in him. I sold 90% of my holdings at a big lost, and cut my sunken cost. If I keep buying it now, then I would be buying for hope, not for high probability of turn-around. There is low probability or any certainty that Intel will catch up with TSMC and Samsung in manufacturing. They have never demonstrated they watched up anyone before. So buying on hope is probably speculation. I personal prefer to buy on high probability and predictability. You don't need to agree with me, I with you the best.
The only way for Intel to get back to what it was is to buy Hi EUV machines (as it is doing) from ASML. TSM and Samsung have not yet acquired that machine. Maybe that can save it and become a cutting-edge company again. ASML was the one that made TSM and Samsun grow. It all depends on it, the design is the least complicated.
In general the Intel GPUs are not bad for the price, Arc is really good it is only somewhat held back by general driver optimizations, especially the ARC A770
You missed a few major points.
First, a big reason Intel did so well for so long is because AMD was either A) getting blocked from OEMs by Intel or B) Emptying a revolver into their foot repeatedly. When the whole market is a duopoly, and one of the major players is just absent, then of course Intel will dominate.
Second, your PC market share data is outdated. Ever since Intel released their 12th, 13th, and 14th gen chips that far outperform AMD, Intel has been regaining market share, and recent data shows it's closer to 66/33 split in favor of Intel.
Third, it's worth noting that, although PC/Computer sales have slowed, Data centers that run everything still need CPUs, and this market has blown up, and continues to grow. Intel also has a heavy market presence in storage, network, edge, and car chips through mobileye. Intel is also making big advancements into the GPU market, and controls nearly 5% of the GPU market. It's worth noting that a big reason why Intel never made these moves into gaming and smartphones is because it'd require massive restructuring in the way they make chips. For AMD, Nvidia, and Qualcomm it's really easy for them to shift into a new market. They design a new chip, send the designs over to TSMC, and then a few months later they get their chip batches done. For Intel, they not only need to design the chips, but also change the manufacturing process, which would require a lot of investment and time to put into fabs, and to Intel that just didn't make sense at the time. The old Intel really hated investing into manufacturing and such.
Fourth, a big reason why crypto and Ai are so heavily favored towards Nvidia is because Nvidia has spend years investing in software, such as CUDA. CUDA is the biggest reason why Nvidia is so dominant in Ai currently, as AMD and Intel have both made hardware that is arguably much better than Nvidia's offerings at a cheaper price, but the software for each still lags behind. Software has been a long standing issue for Intel, as they haven't made major investments into the sector, and still underpay both software and hardware engineers, although this is likely to begin changing with Pat at the helm(When he was CEO of VMWare, SWE pay was much better than Intel currently pays).
Fifth, the reason Intel is losing so much isn't because they're investing a lot, it's because revenue declined, but fixed costs stayed the same. If you read the 10k, you'd notice that the operating expenses over the last few years has all been at about 50 billion dollars a year, while revenue has dropped. This is mainly because consumers aren't buying as many computers currently due to already having bought a lot of tech during the pandemic, and now with fears of a recession people are buying less. This is combined with higher interest rates, which means companies aren't borrowing as much to upgrade servers, which has led to a massive slowdown in revenue there. Yes, companies are upgrading servers with Ai chips, but that's only the top companies, and the vast majority of companies aren't spending much on chips currently. Nvidia only needs a few players buying a ton of chips, but Intel needs a lot of different companies buying, simply because servers aren't much better simply by adding more chips. Luckily for Intel, both these trends are only temporary. It's already been noted that consumer demand is rising again, and a lot of people have been waiting for 15th gen to upgrade, which will be coming out later this year. Also, most servers are on a timetable to upgrade, and servers only last about 5 years MAX, meaning the 2020-2021 servers that were bought will need to be replaced, and Intel will likely be the ones reaping these rewards. The massive Capex actually isn't that much more than historical amounts, and the foundry business has always operated at a loss to serve the chipmaking part of the business, it's just that now, people are getting independent results about this and freaking out. Intel has already stated that their current capex will likely come down in the coming years, as their 18A fabs are getting finished, and those will have the most advanced chipmaking process in the world. It's also worth noting that the 2026 figure you stated isn't about Intel in general, but about Ai. Intel doesn't need to convince the market about CPUs, because they're already the market leader. There are still tons of people that will only buy Intel chips, because AMD before Ryzen was a gamble on whether you'd be getting a brick, an underperformer, or a thermal disaster. Intel's main spending right now is all about their future foundry business, which you didn't mention for some reason. This foundry business is projected to be second in the world by 2030, and the revenue from that alone should help drive the company forward.
Overall, the company is in a better position than it's been in over a decade. Pat is easily one of the best leaders in the world, and what he did with VMWare should show people just how capable he is as a CEO. I have no doubts that 10 years from now, Intel will be one of the largest and best chip companies in the world.
I'd say the video was good overall, and it made a lot of solid points about how Intel's past was markedly bad due to leadership while not ignoring the fact Intel has a strong future. It seems like most videos on Intel are only extrapolating past performance and suggesting that, even though the entire companies management and vision is different, that somehow the company will remain on this downward trajectory. I'd also like to see a video talking about their fab business and a deeper dive into that arm of the company. It has a very bright future, and with Pat at the helm, Intel will be a great investment for anyone willing to ride the wave.
On your second point - the Intel 12/13/14th gen chips don't outperform AMD, they mostly just match AMD. In some cases they outperform, but in many cases they also use a lot more power for the same performance, at least on higher loads.
intel holds one of the top ten spots for consumer cpu sales on amazon right now. The other 9 are all AMD.
@@lycanthoss The reason Intel chips use more power is because Intel engineers have given complete freedom to the user to overclock the CPU to death. On a pound for pound rating, Intel does beat AMD, but this is only slightly, and even with a power cap and under-volting, the performance slightly favors intel(on some applications, some still favor AMD). The two are very close, but we'll have to see what 15th/8th gen brings, and I really believe Intel is bringing their A-game, as compared to 10th and 11th gen, this turnaround in competitiveness is insane.
solid points but pat lied about the dividend although i give him credit for buying shares but investing in intel comes at the opportunity cost of investing in nvidia which jensen has a proven track record of being leagues above pat.
@@laCruz40oz nvidia is overvalued. Intel is undervalued. Pat doesn't have enough time to prove what he can do, while Jensen has had decades. It's not a fair comparison
As a current chip designer at Intel I can say that Pat is fighting on both external and internal sides. Intel middle management is a mess and we have been steadily losing top talent to companies like Apple, Cadence, Microsoft, Google and Amazon these last 3 years. A lot of them have been director and VP level as well as Principal Engineers. Retention goals have been set and we are still not even near on hitting retention targets. I do hope that Intel does well, but it seems like it will be due to becoming a foundry for all these other companies that are now creating their own chips. Middle management is still led by the same people that made Intel failed before. On the other hand, the older gen top tier engineers seem to be willing to retire as a lot of them have ~30 in the company and morale is so low at the individual contributor level.
Intel put a MBA in charge of a technology company and he completely failed to develop new smaller semiconductor fab technology. They went from tech leaders to being stuck on 14nm for a embarrassingly long time.
That was one annoying thing 14nm while AMD went into the 12nm territory... I knew Intel are not serious
Damn that sound in the beginning is so nostalgic. Takes me back to time when the entire town had very few computers and most were in library.
2 months after this video and Intel reputation is in a bottomless hole with with 2 generations of defective cpus by which is far worse than the series of security vulnerabilties from recent years....
2 months and 8 days later Intel crashing 25% and fires 15000 employees lol
Honestly,when they made their Pentium 2 they were at their peak then.
Everything started to go downhill from there.
If you ever saw Pentium 2 you would know what im talking about.
Investing in stocks is like bagging the ugly girl that’s gonna glow up after the summer break. Intel’s problem is that their fab was their sole customer which gave it slack. While TSMC was busy Fabing for nvidia, Apple and AMD so they had to earn it. But Intel remains the only US based major fab. It’s crazy that it’s valued under applied materials (who they constantly buy wafers from) and TI (who is stuck making sensors and components with processes from 2 decades ago). Intel is spending in opening their fab for everyone to use. I would not write them off
As someone who works in the automotive industry, I can tell you that while not the biggest player, intel is still in the game there thanks to Mobileye aquisition years ago. However they are not the biggest player there and often have issues with propper supply of chips (although it seems it's almost common for everyone).
Which AI tool do you use to modify voice?
It's still a shame what happened to Optane storage given it's price tag, low latency and extreme endurance. I strongly suggest you guys pick one up as a caching drive (NOT the tiny ones, the 256GB+ models) before they disappear forever. I got a 905p 1TB for my system and it slots perfectly into the last available 3.0 x4 slot :)
Intel actually had atom z mobile chips back in 2013 and I had an asus tablet with atom z2520
Time to buy Intel. People are way overestimating what AMD is capable of and way underestimating Intel. Everyone stumbles and the media loves to pile on.
I think they are shifting to be a foundry instead. You don’t buy a couple ASML machines only to make the cutting edge chips for yourself.
Intel's first GPU was the Larabee project in 2009 but it was not released.
no, their first gpu came in like 1998, iirc it was a i740 or 750 and was not a good gpu
Wow that is right
Intel definitely is taking GPUs seriously. They just aren’t a gpu company at their heart and starting a new gpu branch is incredibly difficult
Complacency kills!
Intel appointed Christoph Schell at the head of their sales division. Since then, sales plummeted, parts of their balance sheet went into negative, good products fail to sell. Not sure if the guy let go of technical talent who knew the products well. Intel must get better leadership who really understand the technology.
A few days ago Intel’s CEO was here in Costa Rica. I have a lot of friends whom work for them. They are not dead, far from it. Intel is probably getting super strong right now
I am sure. You should load up as many as Intel stocks as you can and borrow money to buy them!
I am sure you are not disappointed!!!
@@PO-nb8qc that is a super stupid idea, but yeah… I want to invest in them, but not being a retard
Erm might needa reassess the situation after the announcement from intel ytd
Pat is a lying gaslighter
Well, when I was working for Intel it seemed their priority is more regulatory issues, diversity and data security. I never have worked for a more formal US company. Furthermore, many acquisitions like Altera / FPGA and Infineon - Radio Networking - was not really successful. They did not really improve their acquisitions but just managed the status quo. Now they rebrand Altera again.
Intel still has lots of amazing Cpu's. That is just a phase. They will be on top eventually
Engineers are the real asset of all tech companies but when a company treat their engineers unfairly, that's when sh*t hit the fan.
didn't even mention arm that is munching on their server business for many reasons. energy, ARM processors consume significantly less power compared to x86 chips, resulting in lower cooling requirements and reduced energy costs for data centers. By distributing workloads across a larger number of lower-power ARM cores instead of fewer high-power x86 cores, ARM servers can achieve better energy efficiency.
High Density
The system-on-chip (SoC) design of ARM processors allows for higher core densities, enabling more compute power in a smaller footprint. ARM servers like those from Gigabyte with up to 48 cores per chip can provide high performance while occupying less physical space in a data center.
Cost-Effective Scaling
ARM processors are generally less complex and cheaper to manufacture than x86 chips. This makes it more cost-effective to scale out ARM server farms by adding more nodes as needed, rather than scaling up with more powerful but expensive x86 servers. AWS (with Graviton), Alibaba, and others are already deploying ARM-based server instances in their data centers to leverage these benefits
This is just not going to be true. ARM chips have so many weaknesses. They can't run most x86 apps, they can't run processes as fast. The only reason x86 isn't being built by Qualcomm, Apple, and Nvidia is because it's open source and just requires royalties.
@@moomie1634 RISC-V is on the horizon as well
And that's why I sold my NVDA at $950 and shifted it to ARM. Damn I'm a clever yob eh? ;) The reason we haven't seen things like 3D memory integration is because that stacking = heat. I expect this will be less of a problem with ARM arch. Is compute IN memory possible with ARM? That would be a truly massive breakthrough.
@@artdehls9100 ARM!?!?!? DUDE, THE COMPANY IS TRADING AT 37X SALES
@@hyperturbotechnomike that's what they always say. Everything is the x86 killer. It just won't happen
Nvidia acquiring ARM & AMD developing ZEN architecture was an important move that affected Intel's fall, intel used to be good, but amd & ARM became more important
Even in gaming, Intel is losing ground quite fast. The X3D AMD CPUs are quite fast for gaming whilst consuming way less power. Meanwhile, Intel CPUs are offing themselves due to crazy power budgets.
Then we have the datacenter, Epyc is slowly but surely eating Xeon's share.
It's crazy to see a mega giant like Intel falling like this. From the king to the underdog in less of a decade.
I can't wait to see Intel picking themselves up with their amazing technology. Maybe this is the final push that they need to achieve that.
Years ago I was told that Apple should just close up shop. This was right at the time Steve Jobs was returning to Apple. Intel is making all the right moves at this time.
7:48 Ever heard about Zen X3D? Intel is not competitive in gaming. They need insane power consumptions to even approach AMD its best gaming CPUs. Those CPUs are more expensive and require more expensive motherboards. On top of that, their 13th gen and 14th gen CPUs are experiencing stability issues or downright failing when used for gaming. The incoming fix seems to be to clamp down on power usage. Intel has allowed the motherboard vendors to push settings to the maximum to extract a few more percent performance out of those CPUs. And now that is backfiring.
Intel only has laptops left and that is AMD its next push.
Our new workstations at work are AMD Threadrippers. Never thought I'd see the day we'd be using AMDs.
We're still on Nvidia though...
Just got laid off at Intel. Not gonna lie we only did about 1.5 hours of actual work per 8 hour day @ $75/hr.
fab or office? R&D or Marketing?
@@limpbalm we got extremely arrogant a while ago because we were basically a monopoly. Then came Nvidia amd tsm arm and wiped us out
That's called deadwood in Engineering Departments. That's when you know the culture is rotten from the top down. It's going to require a complete top to bottom leadership change. I wish Intel luck not gonna be easy.
That's why you got laid off.
@@jamescarter8311not after we milked the company for hundreds of thousands over the years. We made out like bandits!
Intel had a badass weapon in their foundry tech and didn't capitalize on it. What a shame.
We should not underestimate this giant. Stock valuation says nothing. Intel stil has worlds best engineers and much more financial power than AMD.
They are correlated in the long term. Only 4 giants remained in the nasdaq after 30 years. The rest were swap for new industry leaders
Good engineers but a lot of blind narrow-minded managers who unfortunatelly make all the important decisions and ultimately drive the company.
Nah the best engineers are elsewhere now, they're having difficulty attracting top talent because they don't pay competitive salaries any more
@nazar1744 Plus... they continue to fire people, 15k now.... bc they have to reduce the work force by 15%... and guess who makes these decisions.... the managers. So if you get along well with your manager, you're saved.
Key strength: Revux adaptability positions it for market changes!
Not all the companies of the first list are chip makers. Asml builds chip making machines
I have 13th Gen K processor almost 2yrs now. OC^d it ever since. No problem. Playing MSFS 2020 at 1440p since built. Smooth asf. No issue for me.
You guys are wrong. Intel chips are multipurpose. My I9 13900k cpu can also double as a space heater. Im not kidding 😂
In the late 80's, early 90's there was a slew of Harvard business laws paper's which argued that Fortune 500 companies are best led by financier's and bankers over technical led companies. Since the Fortune 500 companies run in packs, this led to all the Fortune 500 companies putting in bankers into CEO positions. This move absolutely destroyed engineering type companies, though it took many years for this mistake to manifest itself. Companies such as Boeing, GE, AT&T, HP and now Intel have suffered when banker CEO types took over. Now, the hottest companies are technically led CEO companies. When you are a technical company, engineering development is a massively expensive area where the payback for engineering development isn't for 5 years or even more. The issue is, if you miss an engineering development product cycle it can/will have massive implications on your future product roadmap. If you are a banker type CEO, it is easy to see that your present yearly profit can be maximized by cutting engineering development and focus on marketing and finances. In truth you need a CEO or C's who master engineering development and financing/marketing NOT just financing/marketing.
The plans to extend intel's business model to include a silicon foundry open to 3rd parties will turn Intel into one of the most important players in the AI race. Almost 70% of Silicon foundries are in potential conflict zones Taiwan, South Korea, Israel. The US has been making moves to limit the access to advanced AI chips, once the US has the ability to manufacture them in the US (or EU) they will implement restrictions.
it will turn them from designers they once were into just a fab, again, intel beginings was like this, well, at least we will not get more snake oil, no more amd glue bs and no more bribes and criminal behavior, so it is something, because if amd put intel to make their chiplets intel must shut the f up
ASML is not a chip manufacturer. They make the machines that make chips. They should not be compared to Intel, Intel buys machines from ASML.
Yeah this video is not very informative.
1:10 Wow! Intel is below even Boeing!
500th
BOEING BETTER!
I think the complacence of the Intel stagnation era is somewhat overstated. They invested quite a lot into architectural improvements. Trouble is they came to focus heavily on new SIMD instructions, which are extremely efficient in theory, but difficult to work with in practice, so these innovations took a long time to seep into consumer applications. As a consumer you'd never notice if your CPU had the latest set of AVX 512 extensions. Meanwhile their EUV lithography got badly delayed, so they got stuck endlessly polishing their 14nm process node. So what we got was year after year of very modest performance gains, stagnant core counts and rapidly increasing power consumptions.
Then off course there's Optane, the extremely fast and wear resistant persistent storage, which looked really promising when it was announced, but release got delayed, and flash memory evolved too rapidly, making optane pretty much pointless by the time it got into production.
Not to say they weren't complacent! It's pretty stark how much more effort they put into real performance after AMD had clawed back their entire lead and then some. With no fire under their feet Intel management clearly weren't in a hurry to improve real world performance.
AMD won't kill Intel. The ARM and RISCV combo will.
ARM is not going to kill anything.
@@jamescarter8311 cope.
for me one of the biggest mistakes was abadonning of intel quark microcontrollers, they had very high clock rate and single PCI lane for very good price, but not only they closed resources behind wall, they never promote that in sufficient way, they made just opposite to what microchip is constantly doing, thats is spamming free just-download IDE, toolchains, programmers, dev-kits and all kind of support for everyone who wants to use their platform
Yeah intel is collapsing the 120 billion market cap, 75 percent market share in consumer laptops, and promising new reasearch
Intel is going to need to pull off a miracle to turn things round. Once NVIDIA starts ramping up their CPU business, that will surely start to eat in further to their server market.
64 percent of x86 computer processor they arent going anywhere anytime soon
In my opinión the fall started when they shelved RISC (Itanium). True, it has a ton of problems, but instead of solving, they opted to exit
Are you just making clickbait now? Dude. Last couple videos been pretty sus.
Little did he know the $10USD per stock drop this week after Intel "Shed" all their reputation axing 15000 peoples jobs ... yikes
If the channel exists in 20 years, there'll be the exact same video about NVIDIA. Maybe Google, too. No one stays on top forever. Companies slow down, age & die just like their leaders do.
On a much longer time scale, so do empires rise and fall. History shows its almost guaranteed.
NVIDIA never rests on their laurels, they will ALWAYS remain #1.
because of arrogantnity
Excellent video! Many thanks for sharing these insights! 😊
Thank you Punditgi!
The fanboys will save them.
missing the mobile business was the no 1 reason
please stop begging for us to subscribe to your email newsletter. literally nobody wants more emails
Inspiring vision: Revux commitment to decentralization stands out!