Intel and the 10nm problem

Intel is a giant, both its results and its market capitalization demonstrate this. We must also bear in mind that, within the technology sector, it is impossible to always hit the targetThat is, we cannot always do things well, or at least as well as we would like.

If we look at the sector of general consumer processors we will realize that both Intel and AMD have made many important achievements, and that both have at some point had the crown of performance on their “heads.” However, since Intel started the transition at 10nm, it has gradually entered an increasingly complicated situation.

When we talk about a “difficult situation” we must give it the proper context. This does not mean that Intel is in a bad moment, nothing is further from reality, but it does imply that is having trouble moving at the pace the company had planned, and that this is already starting to take its toll.

Let’s go back to 2017. In that year Intel had Kaby Lake processors on the market, a generation that used the 14nm + process, and that marked the end of the chip giant’s tick-tock strategy. AMD had nothing to compete with, and it looked like it would never be able to bounce back, but the Sunnyvale company chimed in with Zen, an MCM (multi-chip module) architecture that turned the sector upside down, and that supposed AMD’s return, through the front door.

Intel was calm, after all held the performance crown in terms of CPI, and its monolithic core architecture allowed much higher working frequencies. With Coffee Lake those in Santa Clara raised the core and thread count to 6 and 12, respectively. AMD’s response was Zen +, an architecture that was a slight improvement over Zen, and that did not end up putting Intel in a bind.

Zen 2 was the architecture that really became a real wake-up call for Intel. AMD had not only managed to achieve it in terms of IPC, but also had it completely coiled in core and thread count, and with a lower cost architecture at the wafer level, allowing it to offer much more competitive prices. It also advanced to the 7nm process, while Intel remained anchored at 14nm.

Intel

Intel and the 10nm process: a very complicated evolution

I perfectly remember that an Intel spokesperson said at the time something that, in my opinion, perfectly explains the problem that the company has had to drag for all these years, and that is when they decided to face the transition to the 10 nm process they sinned as optimists, they thought that the jump to advanced technologies, like extreme ultraviolet lithography, which will be present in the 7nm process, and in the end they found themselves in a dead end.

The launch of the first Intel 10nm processors was initially planned for 2016, just after the arrival of Skylake the arrival of Cannon Lake was to take place, but in the end this did not happen. Intel released Kaby Lake, a new “tock” on Skylake, and a twist on the 14nm process. In the following years the mechanics were the same, so much so that we already accumulated six generations in the 14nm process: Broadwell, Skylake, Kaby Lake, Coffee Lake, Coffee Lake Refresh and Comet Lake S.

I know what you are thinking, why doesn’t Intel abandon 10nm and go directly to 7nm? Well, very simple, because the world of semiconductors does not work that way. The company has invested a huge amount of money in the 10nm process, and you have to make it profitable before passing at 7 nm. The launch of the Ice Lake U-series and Y-series processors has been a move in the right direction, and the same will be true for Tiger Lake U-series and Y-series, but the high-performance 10nm processors are still “pending release.” .

Intel has no reason to really care yet. He is in a very comfortable position overall and has a lot of resources, so I don’t think he will have any trouble completing those pending transitions. All in all, it is important to note that AMD has also not stood idly by to see them coming. Zen 3 comes out later this yearAnd if Intel turns to the 14nm process again, it could end up in a position that is difficult to justify in front of users, and investors.

Now it’s your turn, do you think Intel will be able to recover in the medium term? The comments are yours.

Leave a Reply

Your email address will not be published. Required fields are marked *