NVIDIA’s latest quarterly results have consolidated what was already a clear trend: the company is like a cannon, multiplying revenues and profits, thanks to the sales of its specialized AI chips. Their competition is scarce and the work of decades has culminated in an era that has earned them a place in the group of the technological elite, along with Microsoft, Apple, Alphabet and Amazon.
Intel has missed many key opportunities in mobile, computer, and lithographic leaps…
On side B of the album, Intel announced a couple of days ago its new roadmap with which it aspires to have “the best transistors” in 2025, after moving back in positions as a result of the advances of TSMC and Samsung. We also learned from Satya Nadella that Microsoft will work with Intel to produce its own chips and it’s very possible that that includes an AI accelerator, not just a CPU.
The return of veteran Pat Gelsinger in 2021 has allowed the company to move in a much more agile and ambitious way than in previous years, but it continues to carry an inertia of decades that has made it grow at a much lower rate than its competitors.
A review of a quarter of a century
Bad luck, questionable decisions, half-bets and the determined advance of competitors contextualize these last two decades of Intel.
One of the products that has changed the world in this time has been the iPhone. With asterisks, without discrediting Android, but that phone established the physical standard that is still in force almost two decades later. Intel could have been part of that first iPhone and years later its CEO confirmed that there were advanced negotiations for it, but he did not consider Apple’s conditions adequate and ended up discarding the operation.
Although it is true that few people at that time could foresee what that launch was going to mean, Intel ended up giving up being part of the winning horse . The iPhone itself and the entire mobile industry ended up choosing ARM and Intel completely missed that boat.
Samsung and TSMC were following their own path in mobile phones, parallel to Qualcomm, which, together with Mediatek, finished dominating the smartphone, the great technological figure of the last decade and which continues to be so in this one.
We started 2018 knowing that their processors had a serious design error, Meltdown , impossible to solve through updates, for which the operating systems themselves had to redesign their kernels to remedy it, something that implied a slowdown in performance. That error, which ended up affecting many other devices, and not just those with Intel chips, was another thorn in the side.
2020 was one of the critical moments. On the one hand, the AMD Ryzen 5000 proved to be far superior to any Intel product for desktop computers. On the other hand, Apple announced and executed the transition from Intel x86 chips to its own Apple Silicon ARM chips in its Macs. With spectacular results.
During all those years there was also another constant for Intel: that of delays when changing lithographic nodes , something continuous that has been weighing down its products in all the nm leaps of these years. Especially in the face of their competition, which already had the next one ready by the time Intel thought it had caught up to them. The 14nm bottleneck for years was especially problematic.
With the smartphone and the computer already in the hands of others, to a greater or lesser extent, Intel has been dwarfed. As if that were not enough… NVIDIA, who did not buy ARM solely for regulatory reasons and has still done extremely well all this time.
AMD, NVIDIA or TSMC have overtaken it to the right in several aspects
NVIDIA is the one who has practically achieved a monopoly when it comes to supplying chips specialized in AI tasks , something that has caught the industry on its feet in an era where interest in these tools has exploded. Microsoft has already made a declaration of intent with Intel , but it is still to be developed and its chipsets would arrive, at the earliest, at the end of this year.
Intel has also adapted to this stage: it has separated its chip design and manufacturing groups with the idea of taking advantage of the trend of different companies designing and building chips, like TSMC with almost all the others. Thus, it could work better with external clients (like Microsoft!), and perhaps break a trend that has left a devastating fact: Intel has only grown 5% as a company in the last 24 years .
It is no longer just that Apple or especially NVIDIA have skyrocketed in all this time, it is that AMD has also done so to a lesser extent… and above all, even TSMC, of whom we can only start counting from 2019 since it was that year when it began to be quoted.
A base 100 graph that represents this evolution. These types of graphs take as a reference the value at which everyone started and establish their growth from it, to homogenize actions with very different values.
And a simpler way of looking at it: the adjusted closing value for all of them. We discarded TSMC from the graph above because it has been trading since 2019 and we cannot represent it clearly in it. We have also not accounted for ARM in either of them since in its case it went public at the end of 2023.
A fact that condenses all the slippages of someone who was called to do much more and who will now seek to redeem herself. Maybe he will succeed, and even Europe can make it easy for him , but with each passing year the need to come back becomes more pressing. It’s either that or living off the aftertaste of the past. And, inevitably, lose weight.