Intel has had a string of high-profile blunders with their newest processors. If these trends continue, Intel could go bankrupt by 2035.
I've gotten into PC building content over the past year or so. I started watching Phone Repair Guru on YouTube Shorts back in around 2021 or 2022 and it gradually snowballed from there, with my shorts filling with Salem TechSperts, UFD Tech, Carter PCs, and Zach's Tech Turf alongside VTuber, LawTube, and Rebecca Rogers content. I also bought PC Building Simulator 1 and then 2, and eventually bought a Dell Optiplex 5040 from the Purdue Surplus Store and converted it into a VM box. Now it runs KVM on Linux Mint, with some minecraft server VMs running Debian along with a maintenance VM running Linux Mint I use exclusively for running VMManager over SSH X11Forwarding. That was after I built a PC from scratch using a new NZXT H5 Flow, new Gigabyte H610M, new FSP Hydro 550W, new Arctic i35, two used Corsair Vengeance DDR5 16GB-5200MHz sticks from Ebay, a suspiciously cheap factory sealed I5-12600KF from Ebay, a used MSI Ventus 2060 12GB I had laying around, a Samsung 990 PRO 2TB lifted from the laptop it was replacing, and one of two new WD Blue 2TBs from Best Buy.
That was a lot of words and they probably mean something. But there were two sagas I have been following in the shorts feed and eventually the longs feed. These were the duopolies between Intel and AMD, and AMD and NVIDIA. Zach's Tech Turf kept summarizing the AMD and NVIDIA duopoly very simply: on price to performance, AMD usually wins, but if ray tracing or DLSS are a factor, NVIDIA is better. As for Intel and AMD, it was always that AMD has a better upgrade path because AM4 has outlasted any chipset from Intel. But in terms of head-to-head, Zach was a lot less clear, and that's because there's more room for comparison. Intel is often more expensive, but not always. The Ryzen 7 7800X3D is probably the best chip for gamers on the market right now, but Intel chips like the i5-12600KF or I3-13100F still provide pretty good price-to-performance. Ultimately, in the gaming space, a lot of gamers are only concerned with FPS, and FPS is mostly a function of the GPU, and the biggest job of the CPU is to make sure it isn't bottlenecking the GPU.
But where there isn't a comparison is in the high-end Raptor Lake chips. If you haven't been following what's been happening, the i9-14900K has gained a reputation in recent weeks for failing. There was a debacle when it was found that some ASUS motherboards were overclocking the 14900K by default, and Intel later released a power profile for the 14900K that meant reducing performance by 9% according to some sources. But even underclocking the 14900K only bought time, and as of this month several sources started reporting that servers have been failing as the 14900K chips finally crack. One company reported that the 13900K and 14900K constituted 100% of their server failure rate. Most of the high-end 7800 chips from AMD don't hold a candle to how badly Intel chips are doing, while still performing better.
While the 14900K and the 13900K are just two chips, they're still cause for concern. Every company has screw ups, but most companies don't let two of their products fail this badly and this back-to-back. When AMD's 7000 series started exploding motherboards, AMD started working on fixes and returns within two days of the story breaking. But Intel said they would release a statement about the 14900K in May. It is now July, and they haven't released a statement except for leaked internal documents which indicate that corporate is confused and disorganized. Ultimately, the problem with Intel isn't just two bad chips, it's a lack of proper communication and accountability.
This echoes a lot of the same problems that Boeing has had with their 737-MAX lineup. The 737-MAX was a rushed rollout in order to compete with The A320-NEO that was also set to release, and as such the 737-MAX made a number of compromises. One compromise was that Boeing wanted to fit bigger engines on the 737-MAX, but their plane was too low to the ground for that. So instead, they moved the engines, but this meant that the plane started tilting too far to the back. But rather than train pilots on this new setup, they just added a program to the flight controller that would control the pitch of the aircraft without telling the pilots about it. And two months later, two airplanes fell out of the sky after takeoff because they had never been told about this change, and the Federal Aviation Administration grounded the 737-MAX8. Then earlier this month, a 737-MAX10 had a door blow out soon after takeoff from Portland. Boeing has since turned to silencing whistleblowers and a settlement to avoid taking responsibility for these disasters.
The reason I bring up Boeing is that Intel is having a lot of similar problems. The corporate culture at Boeing was chasing stock price above quality, and it seems that Intel may have done the same. Intel rushed a poor quality chip that wasn't actually that different from the lower-end chips it was supposed to be competitive against, but rushed it out the door with a bandaid solution. Their hardware design was poor, so they simply fixed it with software. Boeing implemented a new program to fix their bad hardware and didn't tell anyone, and Intel implemented new overclocking to fix their bad hardware and didn't tell anyone. It's a sign of Intel doing sloppy work to try to compete, and ultimately hurting their reputation in the process.
In today's economy of competition, customer loyalty is one of the biggest intangible assets a company has. Especially for big purchases that don't happen often, it's much better to keep a single customer coming back than it is to try to offer the absolute best product. Look at Huey Fong Sriracha, look at Google Cloud vs AWS, look at Starbucks Loyalty Card, talk to the your car repair mechanic or computer technician or waitress. One of the big reasons AMD felt the need to come out within 2 days is it inspires customer loyalty, that AMD is willing to recognize a problem and make it a priority. One of the reasons that AMD is comfortable with making their AM4 socket last even longer is it inspires customer loyalty since they only have to buy a new CPU and not an entire motherboard to go with it. The biggest problem with Intel not releasing any statement is the breakdown of trust. After their third time replacing their CPU and motherboard, the high-end customer is going to give up on Intel and look for alternatives and consider switching to AMD. Business customers have already talked about dropping support for Intel because of this fiasco. Intel's repeat customer base is shrinking faster and faster by the week.
The other major category in which AMD consistently wins is in power efficiency. AMD has always galloped Intel in the power efficiency realm, often using a full digit less wattage for the same or better performance. Power efficiency isn't a big factor to most gamers, but it is a factor for businesses to take into account. And some enterprise customers have already dropped Intel, in part because of the 14900K disaster but also because of power efficiency concerns. Compare Intel's newest MeteorLake chipsets with AMD's 9000 series. Reports are that AMD has focused on trying to match the 7800X3D, not beat it. Rather, AMD is trying to beat 7800X3D on power efficiency while staying in the same performance tier and chipset. This makes it much easier for enterprise customers to stay on the AM4 motherboard and upgrade their chipsets for better power efficiency, an investment which can easily pay itself back over a few months.
In my opinion, Intel could do well to follow the same playbook. Rather than chase CPU benchmark scores, Intel could do well to focus on quality and power efficiency. For the really powerful desktop and server CPUs, AMD has its TR4 and TRX40 sockets, and Intel has it's X299 and LGA 2066. Intel could do well to segment their LGA 2066 and LGA 1700 chips to target different market segments, marketing the LGA 1700 to the vast majority of end customers and LGA 2066 to the highest-end customers. For LGA 2066 customers, it makes much more sense to market a more experimental chip that may be more unstable, as these are largely for enthusiasts and the highest-end customers. But LGA 1700 customers have significant variance in needs, and while there are enthusiasts who just don't have the money for chips with an X at the end, most LGA 1700 consumers just want a good computer that runs well and doesn't break too often. Intel could do well to focus their marketing and goals with the LGA 1700, focusing instead on overall quality and reliability rather than chasing higher benchmark numbers and higher stock valuations. Intel releasing a lower power profile and more reliable chip could compete well with existing AMD models while helping to restore consumer confidence in Intel chips as a whole.
Alternatively, Intel could continue it's switch to an entirely new market. Intel has announced that it is building what it calls Intel Arc Battlemage- an attempt by Intel to move into the discrete GPU market. While Intel has always been in the GPU space with integrated graphics, Intel Arc Battlemage would move Intel into the discrete GPU space as well, directly competing with not just AMD, but also NVIDIA. There's a lot of room for a third market player in the GPU space, as AMD and NVIDIA both have their problems and there are ways to compete. AMD has gained a reputation for having terrible driver support, and NVIDIA has been dragging its feet on producing Linux game drivers for years. AMD's reputation may or may not be unfounded, and NVIDIA does have a Linux driver for workstations, but the point being that there are market segments that Intel can plant its feet into. There is space for Intel to grow if it wants to, and this could provide jobs for RND employees who might otherwise be laid off.
There is also one other thing I picked up before I bought any PC building related parts, except for the 2060 and the 990 PROs I already had. That other thing I picked up was distinctly unrelated to PCs, which was why it was not on the shelf of Best Buy. I found it in the store next door, Barnes & Noble. I picked up a 450 page book titled Enron: The Smartest Guys in the Room, subtitle The Amazing Rise and Scandalous Fall of Enron, by Bethany McLean and Peter Elkind. It was a 10th-Anniversary Edition, as that was the book the store manager happened to have, along with zero books on Polish history, Turkish history, or Nortel.
I picked up the Enron book for a couple of reasons, one of which was that I was inspired by a 2-part Youtube documentary series about the collapse of Nortel Networks, Canada's largest tech company. The documentary was very enlightening, and I've already watched it a few times on my surplus store Dell 1080p60Hz or just on my phone. While the collapse of Enron doesn't have much to do with Intel's current situation, the collapse of Nortel does. Enron was, at its core, an accounting scandal, as debt was moved off balance sheet and mark-to-market accounting hid just how bad the numbers actually were. As for Nortel, the documentary drew comparisons between Nortel and a sinking cruise ship, a behemoth whose massive momentum carried them through the best of times, but ultimately ended in their destruction. Nortel had a lot of problems, but one of its problems were that it grew too quickly, and another problem was that it did so using bogus accounting numbers. As enterprise customer after enterprise customer dropped Nortel, the stock price soon took a plunge and Nortel entered Chapter 7, and then years later Chapter 11. Intel could be having the same problem, especially if its investors and management truly are putting stock price above quality. If this continues, Intel could ultimately find itself following in Nortel's footsteps.
Ultimately though, this essay is mostly just the musings of one man. Not just one man, one kid who hasn't finished college yet. And I do have some sources, but a lot of them are just YouTube videos. And many of these kinds of problems have plagued companies like Microsoft, NVIDIA, AMD, OpenAI, and others. Samsung quickly gained a reputation for its exploding phones in 2018, and yet they're doing fine today. Google has had more than its share of problems with its own AI bots, and they continue to weather that storm, but their problems don't stem from a few bad products, but rather poor monetization.
The title on this essay is a little bit clickbait, in that I don't actually believe that Intel will go bankrupt by 2035. But it wouldn't surprise me. I feel that if we were to enter a Second Great Depression, Intel would not come out the other side without at least one Chapter 7 filing. Intel survived the Great Recession, but I don't have confidence in its ability to survive in a worse economic downturn. AMD, for its problems, would fair much better. AMD chips have cheaper operating costs and thus enterprise customers especially will be more willing to purchase them, even at slightly higher prices than similar Intel chips. Upgrade costs for AMD are much cheaper, both in terms of lower power profiles requiring less powerful PSUs and because of the comparatively ancient AM4 socket. And AMD has a much more loyal customer base, and while YouTube comments love to shame people for using Intel, there is actually much reason to prefer AMD such as more effective customer support, the aforementioned AM4 socket, and better support for over- and underclocking. In short, AMD management focuses on customers and customer relations, with good customer support and chips that use less energy and sockets that last so much longer. Intel's management has been disorganized, uninformed, slow to act, slow to speak, and leaky, with chips that fail even when underclocked and chipsets that last as long as a typical master's program. While Intel can continue to skate along on its existing customer base, that ice is melting fast.
There are comparisons to draw with Boeing and Airbus here. Boeing's MAX generation has seen some success, but lax regulations have allowed Boeing to cut costs at the expense of overall quality, and Airbus has been fairing much better. Intel's 14900K and 13900K CPUs have seen some success, but lax regulations have allowed Intel to cut costs at the expense of overal quality, and AMD has been fairing much better. But at the end of the day, I don't want to see Intel die on the vine. I don't want to see Intel layoff its employees, downsize massively, leave thousands of families without a stable source of income, and take a dent out of almost everyone's retirements. But I do feel that if Intel management continues to let this scandal blow over, as they have been doing, Intel could declare bankruptcy by 2035.