Intel, entirely Intel Corporation, an American manufacturer of semiconductor computer circuits. Its headquarters are in Santa Clara, California. The company name comes from the word “integrated e etronics.”
- How Intel Lost Its Way In a World That Loves Chips
- How Intel Missed Out on the Great Chip Shortage of 2021
- Opinion: How did Intel lose its Silicon Valley crown?
- Referenced Symbols
- Problem Two: Server Success
- Problem Three: Manufacturing
- Pentium microprocessor
- Expansion and other developments
- Wrap Up
- Intel has a plan
- A ‘pathway to parity’
- A word about 2025
- Conclusion
How Intel Lost Its Way In a World That Loves Chips
Every morning when I open my Macbook Air I am amazed how quickly it comes back to life. It’s almost immediate. Then I go through my workday without ever waiting for an app or document to open. As they say in the ads, it just works.
You could say that this is only part of what happens as technology gets better. The applications are optimized for speed and the operating systems are streamlined. Just another step on the ladder of progress. However, something special is happening in my MacBook Air.
As it turns out, my laptop uses a chip called M1 that Apple designed and manufactured with an external partner (TSMC). It also happened to replace the Intel chips that Apple had been using for over 15 years. Apple has made its way home with its new chips, and their performance can be felt throughout the industry.
However, this is not a story about Apple, although they will have an important episode. This is the story of Intel and how they lost their way. Intel continues to be successful and is one of the leaders in its industry, but there are cracks in the fortress that can destroy the entire company. To understand how we are here, we have to go back to the past.
How Intel Missed Out on the Great Chip Shortage of 2021
If you’ve read any business news lately, you know there is a global chip shortage right now. This is not expected to happen until the end of 2021 or the beginning of 2022. Companies like AMD and TSMC are in a jealous position. The demand for their products is so great that they cannot grow fast enough. Intel is also benefiting from this, but the mood in their headquarters is not that happy.
Intel was founded in 1968, but I will focus primarily on recent history. We are looking for the root cause of Intel’s current problems and have to go back some 15 years to find it. It starts in 2007 after Intel introduces its famous Tick-tock production model. Intel planned to consistently reduce the size of the chips every two releases, which is every Tick.
Intel started with 65mm chips and only reached 10mm. Everything seemed great at first. Intel designed and manufactured some of the best chips in the world. Their production model continued to improve and their chips were used in countless products. Intel has worked with companies like Apple (2005), Google (2012), and of course Microsoft even before that.
In 2006, Intel also announced its Core microarchitecture to widespread acclaim. He would put these chips in all kinds of devices, including laptops, phones, and tablets. Intel even had a brief entry into the smartphone market in 2011, but left the market 5 years later. The solution seemed to be manufacturing chips for other companies. Intel intended to continue designing and manufacturing chips, which is rare in their industry.
In 2016, things started to turn worse. The 10mm chips were full of production delays and errors. Their plan to make the 10mm chip was not based on what was possible, but simply what was next in their tick-tock model. The company did not know exactly how to make chips at this level, which caused serious production problems.
In 2018, Intel was affected by the Meltdown and Specter bugs. These two bugs have affected all Intel processors since 1995 – with the exception of pre-2013 Intel Itanium and Intel Atom. It was estimated that hundreds of millions of systems were affected by these errors. Eventually, more flaws came to light. It is not clear what the actual impact of these bugs was, but the impact on the Intel brand was significant.
It has changed in the industry as well. Competitors such as AMD have experienced a resurgence in the ability to design their chips. Other companies such as TSMC have gained the ability to make smaller chips, such as 5mm, and the production model has shifted. Competitors like AMD and Nvidia don’t even make their own chips. Instead, they design them and work with someone like TSMC to create them. Intel tried to do both and lost ground on both ends.
Finally, we get to Apple in 2020 and their M1 chip announcements. It was expected that Apple would eventually design its own chips as this would give them full control of their stack. Like AMD, Apple has partnered with TSMC to produce the chip. Companies like Google and Amazon are also exploring the design of their own chips as it gives them a further advantage in terms of performance.
The great chip shortage in 2021 is not the same for Intel. In early 2021, they appointed a new CEO and are working on repairing the ship. They are on track for the 7mm chip release, although questions arise regarding potential production delays. Nevertheless, Intel is not dying. Their stocks have actually performed well over the past 20 years (see chart below), but there are existential questions that Intel will have to answer in the next 5-10 years.
Roughly speaking, this is an average dividend growth rate of over 6%. Despite what we are seeing now in 2021, these dividend increases beat inflation fairly easily.
Opinion: How did Intel lose its Silicon Valley crown?
- Email icon
- Facebook icon
- Twitter icon
- Linkedin icon
- Flipboard icon
- Print icon
- Resize icon
Referenced Symbols
Intel Corp continues to be the dominant supplier of chips for personal computers and servers, but its reign as king of chips in Silicon Valley has come to an end and the pending question is whether its misfortunes are due to pushing the boundaries of technology, a human mix of both.
The chip giant announced this summer a delay of at least six months in the release of new chips designed using the next-generation 7nm manufacturing process, and unveiled a change that shocked longtime observers: Intel may work with a contract manufacturer to make some components of the first next-generation chip a graphics processor centered on a data center known as Ponte Vecchio. While Intel INTC, + 2.49% already works with a foundry for around 20% of its chips, the idea that the largest chip processor is likely to lease part of the production of one of the next-generation important chips has been seen by investors and industry as a staggering fall for the host Moore’s laws.
“I think Andy Grove would probably be in a grave,” said Nathan Brookwood, chief analyst at Insight64, referring to Intel’s famous paranoid co-founder and longtime CEO. “He considered the Intel manufacturing group to be the heart and soul of the company, so to see it collapse, especially with regard to new process technologies, which had never happened before.”
The delay in the 7nm process came after an almost four-year delay in the company’s transition to its current manufacturing process, the 10nm process. Intel is offering intra-node enhancements to its 10nm process, with new technologies and software optimization and other changes to stay competitive with Taiwan Semiconductor Manufacturing Co. Ltd., known as TSMC 2330, -0.78%, the world’s leading custom semiconductor foundry.
Intel’s prolonged downfall allowed rival Advanced Micro Devices Inc. AMD, + 2.78%, to overtake Intel in the manufacturing process with its partner TSMC, a once unimaginable turn of events. At the same time, Nvidia Corp. NVDA, + 5.09%, outperformed Intel by market value and made a bold move of $ 40 billion to buy ARM Holdings PLC from SoftBank Group Corp. 9984, + 1.72% 9984, + 1.72% 9984, + 1.72%, which would push a graphics chip manufacturer into the Intel core microprocessor market.
Three months after the first news of the lag, Intel has given little answers about its struggles, and analysts are still unsure if the company’s slippage was due to increased difficulties in fighting the laws of physics, the company’s staffing problems, or some combination of both. The only thing that may be more unknown is where Intel is going.
Is Moore’s Law dead?
“One thing we all know is that Moore’s Law is over,” said Jensen Huang, co-founder and CEO of Nvidia during a conference call last month to discuss the Arm deal.
Huang referred to a prediction made in 1965 by Intel co-founder Gordon Moore that the number of transistors in a semiconductor would double each year, which he later revised in the mid-1970s to two years. The highlight of this prediction was that computers would become more powerful and cheaper as part of this huge increase in the number of transistors in the future, and it served as a guiding light for the semiconductor industry.
But now the doubling of transistors and the associated increase in computing power have become more difficult as engineers struggle with the laws of physics, and the geometry of electronic transistors is becoming finer and more invisible to the human eye. It was Intel’s ability, like clockwork, to increase the number of transistors in its chips, in its own production facilities, with each new generation of chips that made both the company and its chips accelerate with each new product line every two to two and a half years.
All in all, without even looking at the charts at all, I found the INTC would quite easily break through to $ 60, and even much higher to $ 65 or even $ 70 by the end of 2021. A few days later I set up an exchange:
Problem Two: Server Success
Not so long ago, Intel was a breakthrough; while server space was originally dominated by integrated companies like Sun with similar prices, the explosion in PC sales meant Intel was rapidly improving performance, even at a lower price, especially in relation to performance. Sure, PCs didn’t live up to the reliability of integrated servers, but by the turn of the century Google realized that the scale and complexity involved in offering its services meant it was impossible to build a truly reliable stack; the solution was to build with failure assumption, which in turn made it possible to build data centers on (relatively) cheap x86 processors.
Over the next two decades, the Google approach was adopted by every major data center operator, and x86 became the default instruction set for servers; Intel was one of the biggest beneficiaries for the simple reason that it made the best x86 processors, especially for server applications. This was due to both Intel’s proprietary designs and excellent production; AMD, an IBM-commissioned competitor to Intel, has occasionally threatened the incumbent operator on desktops, but only at the lowest level for laptops, not at data centers.
Thus, Intel escaped Microsoft’s fate after PCs: Microsoft was not simply cut off from mobile devices, but also cut off its access to Linux servers, not Windows. Sure, the company tried to support Windows for as long as it could, both on the device side (via Office) and on the server side (via Azure); Conversely, what has been driving the company’s recent growth has been the end of Windows, as Office has moved to the cloud with endpoints on all devices, and Azure has adopted Linux. In both cases, Microsoft had to accept that their diversity has changed from having an API to being able to serve existing customers at scale.
The Intel capabilities I mentioned above would take a similar turn for Intel: while the company’s differentiation has long relied on the integration of chip design and manufacturing, mobility meant that x86, like Windows, was permanently relegated to a general minority by the computing market . It was an occasion, however.
Most chip designers aren’t fairy-tale like; create a design and then pass it on to a foundry. AMD, Nvidia, Qualcomm, MediaTek, Apple – none of them have their own factories. It certainly makes sense: semiconductor manufacturing is arguably the most capital-intensive industry in the world, and AMD, Qualcomm, and others were delighted to focus on higher-margin design work.
However, much of this design work is increasingly commodified in nature. After all, almost all mobile chips are ARM-focused. For the price of a license fee, companies like Apple can make their own modifications and hire a foundry to produce the resulting chip. Designs are unique in some respects, but design in mobile will never be dominated by a single player just as Intel has dominated PCs.
On the other hand, it is the production capacity that is becoming scarcer and therefore more and more valuable. There are currently only four major foundries: Samsung, GlobalFoundries, Taiwan Semiconductor Manufacturing Company (TSMC), and Intel. Only four companies can build the chips that are found in every mobile device today, and in everything tomorrow.
Problem Three: Manufacturing
Meanwhile, over the past decade, our focus on TSMC’s modules, driven by massive volumes from mobile devices, and a desire to collaborate – and thus share profits – with top vendors such as ASML, has exceeded Intel’s production capacity.
This threatens Intel on many fronts:
- Intel has already lost Apple’s business to the Mac, thanks in part to the exceptional performance of the latter’s M1 chip. It should be noted, however, that while some measure of this performance is due to Apple’s designs, the fact that it is manufactured on TSMC’s 5nm process is also an important factor.
- In a similar vein, AMD chips are now faster than Intel on desktops and extremely competitive in the data center. Again, some of AMD’s improvements are due to better designs, but equally important is the fact that AMD produces the chips using the 7nm TSMC process.
- Large cloud providers are increasingly investing in their own chip designs; For example, Amazon is on the second iteration of its Graviton ARM-based processor, which will be running Twitter’s timeline. Part of the Graviton’s advantage is its design, but part of it is – you know what’s coming! – the fact that it is manufactured by TSMC, also on the 7nm process (which is competitive with Intel’s finally introduced 10nm process).
In short, Intel is losing share of PCs even when threatened by AMD for x86 servers in the datacenter, and even as cloud companies like Amazon integrate backwards with the CPU; I didn’t even mention the rise of other specialized data center operations, such as GPU-based machine learning applications that are designed by companies like Nvidia and manufactured by Samsung.
What makes this situation so dangerous for Intel is the volume issue I mentioned above: the company has already missed out on mobility, and while server chips have delivered the growth that the company has had to invest in manufacturing over the past decade, the company cannot to afford to lose volume right now must invest more than ever.
putting it all together and to be clear I’m not buying or adding at the moment. And I will not sell. Therefore, I am neutral on INTC .
Pentium microprocessor
With the introduction of the Pentium microprocessor in 1993, Intel abandoned number-oriented product naming conventions for the trade names of its microprocessors. The Pentium was the first Intel chip for PCs to use parallel or superscalar processing, which significantly increased its speed. It had 3.1 million transistors, compared with the 1.2 million transistors of its predecessor, the 80486. Combined with the Microsoft Windows 3.x operating system, the much faster Pentium chip contributed significantly to the growth of the personal computer market. While companies continued to buy the majority of PCs, the more powerful Pentium computers allowed consumers to use the computers for multimedia graphics applications, such as games, that required more processing power.
Intel’s business strategy was to make newer microprocessors much faster than the previous ones to encourage buyers to upgrade their computers. One way to achieve this has been to manufacture chips with significantly more transistors in each device. For example, the 8088 found in the first IBM PC had 29,000 transistors, while the 80386 introduced four years later contained 275,000, and the Core 2 Quad introduced in 2008 had over 800,000,000 transistors. Itanium 9500, which was released in 2012, had 3,100,000,000 transistors. This increase in the number of transistors became known as Moore’s Law, named after company co-founder Gordon Moore, who observed in 1965 that the number of transistors in a silicon chip was doubling roughly annually.
Moore’s Law. Gordon E. Moore noticed that the number of transistors in a computer chip doubled every 18-24 months. As the logarithmic graph of the number of transistors in Intel processors at the time of their introduction, its “law” was obeyed.
To increase awareness of the consumer brand, in 1991 Intel began subsidizing computer advertising on the condition that the advertising contained the corporate label “Intel inside”. As part of the collaboration program, Intel set aside some of the money each computer manufacturer spent annually on Intel chips, of which Intel covered half of the company’s newspaper and television advertising costs over the course of the year. While the program cost Intel directly hundreds of millions of dollars each year, it had the desired effect, making Intel a conspicuous brand.
Intel’s famous technical prowess was not free from mishaps. His biggest mistake was the so-called “Pentium defect” in which an obscure segment among the 3.1 million transistors of the Pentium processor made an incorrect division. The company’s engineers discovered the problem after the product was released in 1993, but decided to remain silent and fix the problem in the chip updates. However, mathematician Thomas Nicely of Lynchburg College in West Virginia also discovered this flaw. Initially, Grove (then the CEO) declined recall requests. But when IBM announced it would not be shipping CPU computers, it forced a recall that cost Intel $ 475 million.
Though struck by the Pentium fiasco, the combination of Intel technology and Microsoft software continued to crush the competition. Competitive products from semiconductor company Advanced Micro Devices (AMD), wireless connectivity company Motorola, workstation maker Sun Microsystems and others have rarely threatened Intel’s market share. As a result, the Wintel duo have consistently faced accusations of being a monopoly. In 1999, Microsoft was found guilty in a US district court of being a monopoly after being sued by the Department of Justice, while in 2009 the European Union fined Intel 400.45 billion for alleged monopoly activities. In 2009, Intel also paid AMD 400.25 billion to settle a decades-long legal dispute,in which AMD accuses Intel of putting pressure on PC manufacturers not to use their chips.
Expansion and other developments
In the mid-1990s, Intel moved beyond the chip industry. Large personal computer manufacturers such as IBM and Hewlett-Packard have been able to design and manufacture Intel-based computers for their markets. However, Intel wanted other, smaller PC manufacturers to bring their products to market faster, and therefore Intel chips, so it began to design and build “motherboards” that contained all the essential parts of a computer, including graphics and networking chips. By 1995, the company had sold over 10 million motherboards to computer manufacturers, accounting for about 40 percent of the total PC market. In the early 2000s, the Taiwanese manufacturer ASUSTeK overtook Intel as the leading manufacturer of PC motherboards.
By the end of the century, Intel and compatible chips from companies such as AMD were found in every computer except the Macintosh by Apple Inc., which had been using Motorola processors since 1984. Craig Barrett, who succeeded Grove as Intel CEO in 1998, was able to close the gap. In 2005, Apple’s CEO Steven Jobs shocked the industry when he announced that future Apple computers would use Intel processors. Therefore, with the exception of some high-performance computers called servers and mainframes, Intel and Intel compatible microprocessors can be found in virtually every PC, and the company dominated the processor market in the early 2000s.
Paul Otellini replaced Barrett as Intel CEO in 2005, and four years later Jane Shaw replaced Barrett as President. She held this position until 2012, when Andy Bryant replaced her. The following year, Brian Krzanich became CEO. In 2019, CFO Bob Swan became CEO, and Intel was ranked 43 on the Fortune 500 list of the largest American companies.
Apple’s design of its own chips came as no surprise. The tendency for companies to do this has been going on for years. It’s unlikely that Intel was surprised by what Apple did.
Wrap Up
Fortunately, my “worst” mistake didn’t do much damage. I offset this setback in INTC trading with some amazing gains. Many of my actions have done exceptionally well this year. I “won” a lot more than I lost in 2021.
Also, my dividends are still pumping out – left, right and center. It’s just getting better every month. Trading is fine and elegant, but the long term dividend increase and folding really put a smile on my face. And as Growth Stock Renegade subscribers know, there were also some winners in big growth stocks in 2021.
I will continue to hold INTC and collect dividends. This is because, in general terms, I think the INTC will work fine. And I am not going to sell which would result in huge capital gains. I think CEO Pat Gelsinger could do a good job at INTC in the next 3-5 years.
putting it all together and to be clear I am not buying or adding at the moment. And I will not sell. Therefore, I am neutral on INTC.
Hutcheson noted that the retirement of Chief Production Officer Bill Holt in 2016, followed three years later by former Senior Associate and Process Architecture Director Mark Bohr in 2019, may have created a vacuum due to a lack of senior leadership positions, and instead some may have been focused on machinations inside Intel rather than on customers.
Intel has a plan
Gelsigner’s presentation was a strange mix of alarming and exciting. Intel is spending $ 20 billion building two state-of-the-art factories in Arizona that will be “EUV-capable,” meaning they will be able to produce chips at 7nm and below. It also plans to expand this with more plants in the United States and Europe.
At a time when global semiconductor supplies are both very tight and largely concentrated in Taiwan, this is big news. President Biden has US semiconductor supply chains, and while Intel was clear that there were no government incentives yet, he said he was “excited to work with the state of Arizona and the Biden administration on incentives that spur this kind of domestic investment.”
While you can expect to expand its chip manufacturing capabilities, it’s really interesting that Intel also plans to use its factories to produce chips for companies like Qualcomm and Apple, which are currently dependent on TSMC. However, rather than selling Apple its own processors, Intel would act as a foundry that builds the chips according to Apple’s design plans. This is not the first time Intel has tried a casting model, but from Gelsinger’s presentation, it now seems to be the cornerstone of his business plans. If you can’t beat them, build for them?
On the consumer side, Gelsinger spoke of the need to “restore tick-tock discipline” referring to an old Intel edict that had an improved processor every year, either with a new architecture or based on a fresh manufacturing process. The model was successful between 2006 and 2015, but has since completely collapsed. He argued that we would see Intel 7nm processors in 2023 and, a slight shock to those not following processor rumors, publicly admitted that Intel would have to outsource some of its processors to get back to the peak.
A ‘pathway to parity’
At a particularly telling moment, Gelsinger noted that Intel now has a “path to parity,” which is not the best proposition it can give potential customers and investors. The challenge for Intel is that TSMC is working to expand both its capacity and geographic coverage, including building a plant in Arizona. There is also Samsung, which is aggressively developing its foundry activities, including building the most modern factories in the USA. While Samsung is unlikely to overtake TSMC in terms of technology, it is working very hard to both stay competitive and expand its capabilities.
By the way: Samsung, TSMC and Intel equip their foundries with machines from ASML, which is currently the only supplier of EUV production equipment necessary for state-of-the-art factories. Thus, these companies not only compete for customers, but also compete for resources. ASML is expanding production but has so far only delivered just over 100 EUVs, another potential bottleneck. Last year we talked about EUV, Intel, TSMC and ASML in an episode of our Upscaled explainer show, which is worth checking out if you want to know more.
Coming back to Intel: Gelsinger wants the company to be the “undisputed leader” by the mid-1920s. How this happens is under debate. I mean TSMC is producing 5nm chips now. As I mentioned earlier, it is almost impossible to compare the “nm” values of different manufacturers, so it is difficult to predict where the Intel 7nm will stack up against the TSMC 5nm. The easiest value to compare is the transistor density where, if I were to guess, the Intel 7nm would probably land in front of the 5nm TSMC. It would be great if it was ready now, but TSMC plans to start mass production on 3nm technology in the second half of next year, so chances are it will be at the same stage before Intel is ready.
So things won’t change overnight. Indeed, the immediate future looks rather bleak. When Apple swaps its high-performance Macs for its own chips, it will skip from Intel’s 14nm to what’s most likely 5nm TSMC chips. Apple processors are, in every way, incredibly innovative and good in their own right – we did it again in our Upscaled program – but they definitely benefit from being physically more dense and energy efficient than Intel processors. While we don’t know the scalability of Apple’s architecture, the basic math suggests that these high-performance Apple processors should be able to beat the chips they are replacing. In the eyes of the public, this is likely to be another humiliating moment for Intel.
Away from Apple, AMD is booming, although the move towards the company has now stalled as global chip shortages in practice mean people can’t buy its processors. As with Apple, AMD’s strength didn’t just come from TSMC’s manufacturing strength. He switched to a heterogeneous chiplet architecture in 2017, which allowed him to quickly and cheaply scale the number of cores and SKUs in his product range. And since then, with almost every generation, it improves the efficiency of this design. The latest Ryzen chips have essentially increased performance by about 20 percent over the CPUs they replaced, despite being manufactured using the same TSMC process.
Intel also has great engineers – as evidenced by the fact that it was able to do so much with 14nm. It debuted with the first 14nm chips in 2015, centuries ago in technology. By the end of 2015, its line of consumer desktops peaked at the quad-core, 4-4.2 GHz 6700K. Today, the flagship 14nm chip for desktop computers is 8-core, 3.5-5.3 GHz 11900K. Intel’s 14nm production platform has improved over this time, but we mainly look at a group of very talented engineers working on some very difficult engineering problems.
What makes this situation so dangerous for Intel is the volume issue I mentioned above: the company has already missed out on mobility, and while server chips have delivered the growth that the company has had to invest in manufacturing over the past decade, the company cannot to afford to lose volume right now must invest more than ever.
A word about 2025
Of course, in our conversation with Tichelman, we talked a lot about 2025. Intel wants to be the absolute leader in the chip market again by 2025. It sounds quite ambitious, especially when you see that TSMC, for example, already has over 50 EUVs in production, and Intel is just getting started. Intel can never match this capacity and performance. Certainly not, considering that ASML can supply a maximum of 25 to 30 machines per year. Both TSMC and Samsung still have orders open for new machines.
Tichelman says Intel hopes to develop Intel labs. There are several features that have been in development for over 10 years, making Intel chips much faster. One of these innovations is RibbonFET. This is a new transistor architecture in which transistors can become even smaller and switch faster. Intel wants to start using this architecture in conjunction with the 2nm process from 2024. Intel also wants to power these chips from below with a technique they call PowerVia.
Tichelman says it will make a big difference for Intel and make the company a true leader again. We asked Tichelman how Intel could be so convinced. There is still some time until 2025, and the competition will not sit still. Tichelman revealed that people at Intel are convinced of the new technology.
In the meantime, we understand that IBM and Samsung have already designed chips with a similar transistor architecture. The market calls this gate-all-around (GAA) transistor. Next year, Samsung is looking to start using this in conjunction with 3nm chips, while TSMC intends to start working with GAA on 2nm chips (same format as Intel) from 2023. In any case, we are not convinced that Intel will overtake other parties with this new technology.
Intel’s belief could also stem from something else. ASML expects the first new high NA EUV machines to be ready by the end of 2023 or early 2024. High NA EUV will be the successor to the EUV machines. It’s possible that Intel isn’t very conservative this time, but has invested heavily to get the first machines. Thus, it can take the lead in the chip market.
Conclusion
ASML is the only company that can supply high-quality chip manufacturing machines. This will not change in the coming years. If ASML comes up with a new technology, you cannot refuse as a chip manufacturer. You have to participate because if you miss, you can suffer for many years.
This is exactly what happened to Intel, and through some reorganization, strategy changes, and other management, the company hopes to get back on its feet and end suffering. The lack of a chip limits damage. All chip designers at AMD, Apple, Samsung, Nvidia, and Qualcomm should take it as a big compliment that they have broken Intel’s hegemony. For now, there is nothing we can do but wait to see if Intel recovers in the coming years or if other chip makers continue to benefit from it.