An Attempt to Explain Intel’s Situation

Over the past five years, the tiny rivulets and whispered rumors about problems at Intel have transformed into a torrent. Everyone seems confused as to what is going on over there. After all, there is a reason why Intel is called Chipzilla. So what has happened?

Whatever happened in the previous millennium is likely too far in the past to have much impact on the current situation, so let’s start from the first decade of this century. And of course, we cannot talk about Intel without mentioning AMD – they are the Yin and the Yang, or the David and Goliath, of the processor world; their histories intrinsically intertwined.

So lets start our journey at the great battle between the Intel Pentium and the AMD Athlon processor families. At that time in the early 2000’s, the focus was solely on the megahertz war which was soon to move into the gigahertz territory. Both chip manufacturers were staring each other down in a high stakes game of Texas Hold’em, with each party raising the stakes more and more, upping each other’s frequency. And of course, like all good stories, the era was also rife with shady tactics, dirty marketing wars and deals made behind closed doors. If all of this is news for you, then I would recommend you to watch this YouTube video. The situation even reached absurd levels where certain retailers and OEMs were refusing free CPUs from AMD, for fear of losing the financial sweeteners from Intel’s deep pockets. Without diving too deep into every detail, the gist of the story is that the first decade of the millennium oversaw a fierce fight between AMD and Intel over the technology supremacy and performance leadership.

However, regardless of the fact that performance leadership was periodically changing hands, Intel was always far ahead commercially and financially. They had a better brand image and they were not shying away from shady tactics in order to secure lucrative contracts with large scale OEMs. Eventually things started changing on the technological side as well as cash-starved AMD began to lose competitiveness around the mid to late 2000’s. Late in 2011, we saw AMD and Intel embarking on different roads. While Intel continued to iterate and improve their Core micro-architecture, AMD decided to start from scratch with a brand new architecture.

Intel’s Sandy Bridge architecture turned out to be a good step forward, while AMD’s Bulldozer architecture turned out to be less impressive than they hoped. And from then onwards, besides being the financial underdog, AMD would be chasing Intel on the technological front as well.

Pentium versus Athlon

From this point on, if I may, I would like to take you on a highly subjective and speculative logic trail, trying to explain why we are where we are today.

So now let’s jump past the middle of the ’10 decade. The world has changed in the last 10 years. There was a world-wide financial crisis. We saw the meteoric growth of those that were the new-kids-on-the-tech-block like Facebook, Apple, Google, Amazon, Tesla, Uber, Tencent, Alibaba and others to gigantic sizes, some of them even passing the trillion dollar market capitalization.

These new “kids” – young adults by now – are a different kind of beast compared to the dinosaurs like IBM, Microsoft, Intel, AMD, Oracle, and the other stars of the ’90s and early 2000s. They injected a lot of energy into improving technologies such as neural networks and machine learning, which were up to this point mostly fringe academical projects. Within a single decade, we saw the birth and growth of AI technology at such a fast rate that it’s hard to keep up with developments.

They pushed us into a world where humanity can’t compete with a machine at the most complex game ever invented (Go). Not to mention that the “machine” has learned the game by itself and – even more astonishing – we don’t even understand how it thinks and is able to beat us. Having a computer that makes decisions based on self-learned algorithm that we humans do not understand is a new era of computing, and is some scary new territory indeed.

Today, each of us has in their pocket an automated personal assistant, and we are sometimes unable to discern when we speak over the phone to a machine instead of a person. Machines can track and recognize us as we walk down the street, and they are hooked into everything now that the Internet of Things paradigm has become more and more prevalent. We are deep into an age where physical wars are being fought with ever more sophisticated and autonomous drones. The Internet has changed from a network between people to a network between machines.

Skynet is truly getting closer to being born any day now.

Coming back to our main topic, at the core of all this technological perfect-storm sits the mighty (or tiny) silicon chip. Demand for it has surged in this decade, to numbers which would be unbelievable just 10 years before. And at the big boys’ table there was only one company truly ready to meet the demand: Intel. Intel’s market position became so dominant, that the price of a single server-grade 28-core Xeon Platinum 8180M processor was $13,000 back in 2017.

Intel was selling chips like there was no tomorrow. Its leadership, undisputed. A certain victorious aura was etched on their foreheads and they started fooling themselves over their indestructible supremacy. They started setting overly ambitious goals, something to demonstrate to the world once and for all the technological prowess of a company worthy of the name Chipzilla. And for the time being – since there was no competition anyway – it was enough to continue milking the existing product line. Every year, more or less the same processor was casted and re-casted, sold at ever increasing prices, the biggest difference between consecutive iterations being a couple of hundreds of megahertz increase in the core frequency.

Meanwhile, AMD was in dire straits. Down in the ditch and fighting tooth and nail for their very survival, beating off the creditors asking for their money back. The stakes were never higher. AMD entered the all-in phase of the above-mentioned Texas hold’em poker game. So they hired a new CEO, sold off their manufacturing capacity, which later became known as GlobalFoundries, and hired the best micro-architecture talent they could find. It’s not too hard to imagine that maybe some of that talent came from Intel’s ranks – people still motivated to achieve something and to leave their imprint upon this world. What could a talented young engineer show for, after working at Intel for these past 10 years; while the company leadership was aiming for the stars? Another “plus” symbol which follows a number on a box? Another couple of hundred of megahertz increase in CPU clocks? The challenge was gone at Intel, meanwhile AMD concocted and executed a roadmap which would re-imagine and re-invigorate computing with their combination of a new chiplet-based micro-architecture and a new manufacturing process that can use chips from different nodes on the same package.

A Zen moment indeed.

However, this time AMD is not facing the same Intel of old days. This new Intel has become a giant with feet of clay, staring absently at the horizon, daydreaming of the next big thing they’ll achieve, too busy to look down in the gutter to observe what were the rats doing. In the first two years after the launch of the Zen architecture, Intel did nothing. It’s almost as if Zen didn’t even register. They failed to notice the tide was about to change. They were too confident in their plans, content to continue iterating their (still) successful 14nm product line, while busying themself with corporate power struggles and overly ambitious technological goals. And most importantly, Intel felt safe. It was only natural for Intel to feel safe. After all, money continued to pour in through all the cracks in the wall.

But AMD proved that they are serious this time. That they really do have a plan, and they are able to execute it. Year after year, one milestone after another, AMD would show commitment to its plan, achieving its goals and shaking the foundations of Intel’s dominance. By 2018, early signs of the incoming panic over at Intel started to become visible. The giant was nudged out of its daydream, and worry started to creep in. One such moment was the embarrassing 2018 Computex fiasco, where Intel demonstrated its “revolutionary” all-core 5 GHz, 28-core behemoth, ready to crush the competition.

Then came another major distraction. Keeping around the same architecture for almost a decade allowed researchers to find security weaknesses in Intel’s CPUs across many “generations” sharing the same micro-architecture. The same fundamental weakness found in a processor from 2015 is still found in the latest one. And micro-architectural security issues cannot be fixed on shipped or soon to be released products. They can only be patched in firmware or software afterwards, which comes with (sometimes large) performance penalties.

But Intel were still breaking their own quarterly earning records on a regular basis, even as the storm clouds rolled in over the horizon. That is mostly because the majority of people do not understand technology or are too complacent to spend some time scratching below the shiny polished surface that marketing departments are paid to present to us consumers. We get blinded by shiny boxes and by cool-sounding three-lettered features, and we are willing to pay premium to have them simply for prestige.

And Intel is still enjoying financial success, despite the fact that the competition had got their act together and is now clearly dominant on the technical front again. There is still no immediate financial advantage for Intel to rush out and improve on a product that still sells like hot cakes. Until it’s too late, that is. Exactly like the Titanic that noticed too late the approaching iceberg ahead. By the time Intel started to do something to try to maneuver that huge ship of theirs away from harm’s way, it was too late and the collision was inevitable.

Unfortunately, reality has a nasty way of slapping you harder, the longer you choose to ignore it.

It’s not impossible to think that Intel might have to slim down and follow AMD’s footsteps through the “impassible labyrinth of razor-sharp rocks” and the “festering stinking marshlands as far as the eye can see” road to success. Intel might have to shed some weight and sell off their manufacturing business (just like AMD did), in order to allow them to focus on design. Early rumors certainly indicate an intention to “diversify” the sources of some of their chips, with TSMC favourite to take over more and more of Intel’s manufacturing.

Intel vs AMD stock prices compared since 2005

To conclude. Who is to blame?

In our western culture, we have an obsession with finding the culprit(s). We want justice to be served. We spend countless hours reading books and watching movies which depict scenes where the guilty is found, judged and punished. So the word is out: we want some blood. Not because most of us care about Intel, but because we want to enjoy the scandal. The Romans were experts and knew how to control the masses, inventing and providing “bread and circuses” (from Latin: panem et circenses) to keep them content and submissive.

The reality is though, that very likely no nobody is really to blame for this. It’s only natural that these things happen. Such events occurred over and over again throughout human history. Great empires were born in the crucibles of countless wars, and then when they finally conquered the entire world, they became busy cannibalizing themselves out of boredom. The same can be said centuries later about a few technological ex-juggernauts that also got lost when the world became their oyster.

Looking at the picture above, I cannot stop thinking that AMD is set to become top dog, for a while at least. This thought should send chills down their spines to some key people back at AMD. If those key AMD people can see past the next quarterly result, they should understand that there are lessons to be drawn from this situation.

We Adore to see competition in the field of technology, because we know that fierce straight competition is our best friend. We’ve lived too long under a monopoly and would hate to see a reversal of roles between AMD and Intel. Changing one despot for another won’t do us any good. But we also believe in rewarding hard work, and recently it seems that only AMD has been doing that. We are looking forward to the moment when Intel will come back to us with the fruits of their hard work.

Unless they instead decide to resort to dirty tricks again in their quest to regain supremacy…

Liked it? Take a second to support Anubias on Patreon!
Become a patron at Patreon!