Intel’s Struggles
Intel has been in the financial news headlines recently, but for all the wrong reasons. In fact, INTC just experienced its worst price drop in 50 years, with a 26% drop in just one day at the beginning of August 2024. This came on top of a steady decline the whole year, totaling almost -60% since January 2024.
Source: Google Finance
This is especially shocking, as this goes against the trend of the semiconductor industry, which has been rising to new highs continuously in the last few years. For example, the PHLX Semiconductor (^SOX) is up 5x since 2017.
Source: Yahoo Finance
On a surface level, it seems that markets are extremely pessimistic about Intel. However, it can be discussed whether this opinion is truly warranted or an overreaction to problems that are present but not as threatening as they seem.
The market might be starting to realize it, with the stock price jumping back upward by 9% on Friday, August 30th, 2024.
What Happened In August
The trigger for the 26% stock price drop was the release of the Q2 2024 results. Revenues were down 1% year-to-year, gross margin was 1.1% lower (to 38.7%), and EPS (Earning Per Share) was barely positive at $0.02/share. The Q3 2024 outlook had revenue, gross margin, and EPS all down.
To top it all, Intel announced it will lay off up to 15,000 people.
The tech and financial press had a lot to say about it, including piling on less than ideal optics of the company’s CEO’s social media posting: “Things Are So Bad at Intel That the Boss Is Posting Bible Verses” (even if Intel’s CEO is actually posting bible verses quite often, not just in a crisis).
Source: X.com
The consensus take during the stock crash was that Intel “lost its touch” when compared to its competitors like Arm (ARM), Samsung Electronics (005930.KS), or TSMC (TSM).
Competitive Pressure
At the core of the issue has been Intel lagging behind its competitors like TSMC and Samsung in the most advanced chips like 5nm processors.
Overall, it can be said that Intel got a little complacent and missed 2 technological innovations in the chip-making sector:
- EUV (Extreme UltraViolet) lithography is a method using high-power light beams to engrave extremely small transistors for advanced computer chips.
- Chiplets are smaller modular chips that can be assembled to boost performance at a lower cost.
- This design is especially popular in smartphones and other mobile devices.
- Chiplets are now a core advantage held by Intel’s competitors, ARM (ARM) and AMD (AMD).
Another segment that a competitor would end up dominating instead of Intel is GPU (Graphic Processing Units), a niche that almost the entire industry missed and left to Nvidia.
GPUs were initially a niche segment of the computing industry, mostly useful for PC and console graphic cards. However, it appears that GPUs’ slower and numerous parallel computing (compared to the rapid and “focused” processing of Intel’s CPU) were also very useful for mining cryptocurrency, making GPU leader Nvidia a new market darling.
The usefulness of GPU in AI computing would later turn Nvidia into the largest public company in the world by market cap.
Is it That Bad?
Losing technological lead and declining sales in a booming industry are certainly red flags. These issues are however overshadowing a few elements that make the picture about Intel a lot more nuanced.
A Firm Lead In PC & Servers
Intel had and still has strong leadership in the manufacturing of chips for PC and servers using the proprietary x86 core architecture. Since its inception in the 1980s, it has been, and remains, one of the world’s most widely used processor architectures.
This leadership extended to other “compute-heavy” markets, like servers, which are still a massive driver of Intel revenues today.
Source: Semiconductor Business Intelligence.
This does not solve the lagging of Intel in other sectors in itself but puts it into perspective.
For a few years, the market has focused on smartphones and GPUs. This means that any company underperforming in these segments has been excessively punished, while the winners have been richly rewarded.
However, there are signs that the smartphone market is now reaching maturity, with growth drastically slowing down. Overall, most people already have a smartphone and are satisfied with the existing capability. The new designs are also less variable.
The next big thing in mobile devices also keeps failing to materialize, with tablets still a smaller niche, smart glasses a failure, and VR headsets far from widespread adoption. So overall, mobile devices is a sector that might become a more “boring” mature segment, similar to Intel’s domain of x86 chips for servers and PC.
Catching Up
A New AI Leader?
The GPU and AI questions are different, as they are what made Nvidia an almost 3 trillion dollar company. AI is very different from smartphones, and it is probably the most explosive market in the industry for the decade to come.
So, ultimately, Intel’s ability to succeed in AI chips and AI-related products will likely determine whether the company can return to growth. Without it, Intel would risk becoming a classical “melting ice cube,” where the legacy x86 business slowly fades away to be replaced by nothing.
The central part is the plan for Intel to ship 100 million AI PC in 2 years. This is a solid idea, as it builds on Intel’s pre-existing strength in the PC market. The idea is to build a “One-stop-shop to bring AI everywhere”, built on open ecosystems.
Source: Intel AI
Players must be able to articulate a clear and compelling value proposition that resonates with enterprises across industries – one that paints a vivid picture of how AI PCs can drive tangible business outcomes, streamline workflows, and unlock benefits to productivity and innovation while improving the PCs total cost of ownership, security, and manageability.
So, instead of aiming directly at Nvidia and its lead in building GPU, Intel’s strategy will be based on offering a complete and well-articulated ecosystem for AI PC. This would also leverage Intel’s advantage in CPU and enterprise use cases, as small to medium-sized AI models are more suited to use CPU over GPU.
Source: Intel AI
Chiplets
Intel is also catching up in chiplets, having seemingly realized the potential of the segment, even if a little late.
“Integrating multiple chiplets in a package to deliver product innovation across market segments is the future of the semiconductor industry and a pillar of Intel’s IDM 2.0 strategy,” said Intel’s Chiplet Sandra Rivera, executive vice president – Intel’s data center and Artificial Intelligence Group
Leadership Delivers Industry-Leading Capabilities at an Accelerated Pace
Ramping Up Ireland
Lastly, a lot of short-term costs had been driven by the ramping of its new Ireland-based chip fab, which only started to use EUV at scale at the end of 2023.
The arrival of this important moment ushers in the future for products like Intel’s upcoming Intel® Core™ Ultra processors (code-named Meteor Lake), which will pave the way for AI PCs, as well as future-generation Intel® Xeon® processors coming in 2024 and produced on the Intel 3 process node.
So we can expect a return to profitability, or at least no significant losses soon, with Ireland becoming a key manufacturing facility.
EUV technology employed in the production of Intel 4 will be important for applications such as artificial intelligence (AI), advanced mobile networks, autonomous driving, and new data center and cloud applications.
EUV plays a critical role in driving Intel toward its goals of delivering five nodes in four years and regaining leadership in process technology by 2025.
A New Plan
The need to fire thousands of people and the stock crashing seems to have been the jolt Intel needed to start changing quickly.
In September 2024, Intel’s management came up with a new plan to both speed up the recovery of the company and reduce spending so that the current cash flow is sufficient for the needed investment.
One key part will be to speed up the adoption of EUV technology to compete with TSMC. However, this will be a very long-term process, ongoing until 2030, with the Irish fab only the first step.
Another element will be to bring back in-house the large majority of the silicon wafer production, which is currently partially outsourced.
Source: Intel
To cut costs beyond reducing headcount, they are considering selling Altera. This department is in charge of programmable chip units, which are simpler and less profitable than advanced chips.
Another cost-cutting could be the cancellation of a planned $32B new factory in Germany. It follows the cancellation of a $15B expansion plan for its chip plant in Israel (here more driven by the turmoil and war the country is engaged in).
This does not affect the plan to build new factories in North America, clearly becoming the center of Intel’s future strategy.
To meet the accelerating global demand for semiconductors and advance our IDM 2.0 strategy, we are expanding our manufacturing capacity, starting with approximately a $20 billion fab investment in Arizona and New Mexico, a more than $20 billion fab investment in Ohio, and the acquisition of Tower Semiconductor.
Geopolitical Advantage
The Past Geopolitical Hit
Many commentators focused on Intel losing chipset to ARM, GPU to Nvidia, and mobile CPU to TSMC. Few discussed that the current reduction in revenues was partially a collateral victim of the USA’s “chip war” with China.
For example, the USA revoked Intel’s and Qualcomm’s export licenses to sell to China’s Huawei in May 2024.
With China consuming half of the world’s semiconductors, sanctions have reduced the profit of many semiconductor foundries, including TSMC and Intel, causing a drop in Intel’s stock price at the time.
At the same time, this temporary pain could prove a saving grace for Intel.
The Coming Geopolitical Advantage
“Chips are the new oil” is a dominant idea in the renewed great power competition between the USA and the West versus China and Russia. This idea motivated, in large part, the US sanctions on the Chinese semiconductor industry.
It is also a true strategic threat to the US alliance, with most of the world’s chips produced in Taiwan and South Korea. Both countries are highly vulnerable to China in case of a military conflict, with China officially considering Taiwan a “rebel province” that should be brought home sooner or later.
In this context, Intel is a strategic asset of the US, having the only “safe” chip and semiconductor production far away from potential conflicts in Europe (Russia) and Asia (China).
In practice, any strategically important chip, like for military use, but also utilities, key infrastructures, AI data centers, healthcare, etc. will likely in the future have a strong “made in USA” mandatory clause. This should greatly benefit Intel, as it is already established and producing successfully in the USA.
And already, Intel is a prime receiver of money from the CHIPS Act, with $8.5B in direct funding, which is mostly still yet to be distributed to the company. This should also alleviate any concern regarding cash availability for Intel.
TSMC Struggles out of Taiwan
In addition, it seems that the push to get TSMC to move some of its factories to the US is not going very well.
Some staff told NYT the chipmaker is facing five to ten times higher operational costs in the US, compared with Taiwan. Others said they were reluctant to move to the US due to “potential culture clashes” and concerns around “difficult to manage” American workers.
Asia Financial
It seems that beyond cost structure, adapting the Taiwanese work culture to the US is difficult due to deep cultural differences.
So while this might be resolved in time, it seems it will take at least as long for TSMC to produce large chip volumes in the USA as it will take Intel to mass adopt EUV technology.
Future Computing Tech: Photonics, Quantum, etc.
Intel built its past success on innovation and technological excellence. Its recent blunders of missing GPU and chiplets / smaller smartphone chips should not distract from the fact that the company is still at the forefront of the most advanced form of computing.
Intel is especially ahead when it comes to new concepts of non-silicon-based computing. At the end of 2023, Intel decided to divest its photonics business to Jabil (JBL). (You can read more on photonics in our article “Computing at the Speed of Light with Silicon Photonics”).
But Intel is still very active in quantum computing, a type of computer using quantum effects to solve problems unsolvable with normal chips.
Intel Quantum Computing
Intel recently released “Tunnel Falls”, the “ most advanced silicon spin qubit chip”. What is remarkable is that it is not a prototype but a chip built at scale, with a 95% yield rate across the wafer and voltage uniformity. This opens the way to mass production of quantum computing chips, something for now elusive in a nascent and quickly changing industry.
Source: Intel
Faithful to its roots, Intel is also developing the software to utilize its chips, with the release of the Intel Quantum SDK. This provides the guideline for programmers to develop software for quantum computing compatible with Intel quantum chip design, which has historically been a very strong & profitable business moat for Intel’s conventional chip business.
Source: Intel
(you can also read our article “Top 10 Non-Silicon Computing Companies” for more information on these topics)
Intel Technological Future
Intel has stumbled along the way in its multi-decades-long history, becoming somewhat complacent from its outstanding success in PC and server markets with its x86-based products.
The recent stock price shock and decision to fire a segment of the workforce is a sign that the company is now getting serious about solving its deficiencies.
It will probably not suddenly dominate Nvidia in AI, ARM & AMD in smaller chips, or TSMC in advanced EUV-based chips. But it will likely start to catch up.
And might in the end have a much wider range of products, which can help it in a very quickly changing industry, where the winners of today are often the losers of the next cycle of innovation.
(Under?)Valuation
Intel stock price is now as low as going back to 1997 prices, ignoring completely the growth of the semiconductor and computing industry since a date even before the popping of the dot-com bubble.
Many investors are now starting to consider this could be an excessively low price, especially as the company’s management is (finally?) coming with aggressive moves to change direction, including:
- Reduction of capex to fund R&D and strategic spending.
- Divestment of non-core assets.
- Ramping up of EUV chip plants and refocus on US manufacturing.
- Attack the AI market through a holistic strategy leveraging the company’s existing position in enterprise, PC, and server markets.
This low stock price is also somewhat excessive due to Intel’s strategic importance to the USA and NATO in an era of intensifying great power rivalry.
Conclusion
Ultimately, Intel is the US’s insurance policy against a conflict in the Korean Peninsula and/or Taiwan / The South China Sea. So this provides an almost certain guarantee against bankruptcy or too strong of a decline in production capacity, no matter the short-term financial issues or cash flow limitations.
Intel is also far from an obsolete company, even if it has somewhat been outpaced by its competitors in recent years. This is something not rare in the tech industry and semiconductor manufacturing, where one wrong choice can take several years to correct.
So, Intel’s stock might be of interest to contrarian investors and value investors looking for a major semiconductor manufacturer which is selling at a discount compared to the rest of the industry.
It will however deter many other investors, due to the extreme volatility it displayed in the past few months, and the less-than-optimal performance and strategic decisions of the past few years.
