Let’s go back to 2017, the year in which AMD launched its first Ryzen processors. After several years of infamous rehashes, the company completely changes the design of its processors in a heads or tails bet. The company was known to be on the brink of bankruptcy and these processors were its only lifeline.
To develop and launch these processors, AMD sold its foundries to GlobalFoundries and TSMC. In addition, signing agreements with GlobalFoundries to manufacture Ryzen chips, which later collapsed because the foundry stopped advancing in the miniaturization of transistors.
Where do Intel’s problems arise?
The origin of all evil lies in its monolithic processor design and 10nm lithography. During 5 years the company had to rehash its processors under the process of 14nm. For the first time, the so-called “tic-toc” cycle ended. This cycle establishes a first family of processors with a lithography and a second family with the more perfected lithography, to later be “abandoned.”
There the company was immersed in a runaway process of expansion into other segments. They had their own 5G chip division and were developing new memories for SSD and a new type of RAM with persistence. But, reality ended up having a strong impact on the company.
Its 10 nm process was not good, it caused many failures in production. That the rate of viable processors for commercialization barely reached 50% was horrible, since it prevented their sale. You should know that the rate of viable processors per wafer, to be “profitable” must be between 85-90%.
To save the situation and improve investment in this process, the 5G chip division falls first. This division is sold to Apple for practically what had been invested in its development.
It is still interesting that Manzana be an important actor in all this. The company announcessome time later, that they would stop working with Intel in favor of their own processors. Intel now loses a huge customer and, obviously, billions of euros.
A concatenation of problems and decisions
Years later, Intel would end up selling its SSD memory division in two parts. On the one hand, it sells the patents for its 3DXpoint memories and on the other the SSD division. In addition, it will end up closing the RAM memory division with persistence, a kind of hybrid between traditional RAM and SSD.
During this period two important figures have also left. First Jim Keller, who is the ideologue of today’s heterogeneous processors. Raja Koduiri, father of Intel Arc, the company’s new graphics cards, has left the company behind him.
In the meantime, Intel has opened up to acting as a foundry, something it had not done for almost two decades. We add to this that it has “gifted” the Intel NUC division to ASUS a few weeks ago.
We can see that there is a set of actions that speak to a major financial problem for Intel. It is possible that the company has become financially on the ropes and has had to make unpleasant decisions. Without a doubt, the worst was letting Keller and Koduri go.
