Almost all stock investors have heard of Warren Buffett's famous saying:

"Others are fearful when I am greedy, and I am fearful when others are greedy!"

But would you dare to do so when the time comes?

For example, this semiconductor giant, Intel, has fallen by as much as 40% from its high at the end of last year. If we calculate from the high point in 2021, its stock has been cut in half.

Looking at the slope of the decline, it really resembles the moment in 2022 when everything collapsed.

As a big company that has dominated the global semiconductor industry for decades, has Intel reached the moment of bottom-fishing?

Fundamentals can be summed up in one word: terrible.

Why has it fallen so much?

Just take a look at Intel's finances. In 2021, Intel had an annual net income of $19.8 billion. By 2022, this had fallen to $8 billion, and by 2023, it was only $1.6 billion, a year-on-year drop of nearly 70%.

It's no wonder the stock price has plummeted when net income has fallen so drastically.

So what caused such a sharp drop in net income?

There are two main reasons:

First, revenue is decreasing. In 2021, Intel had a revenue of $79 billion, but by 2022, this had decreased to $63 billion, a decrease of more than 20%. And by 2023, it had decreased further to $54.2 billion, while the decline narrowed slightly to 14%, the fall in the past two years was still astonishing.

Secondly, costs have risen sharply. Overall, there hasn't been much change in sales costs, increasing by nearly $1 billion in 2022 compared to 2021, and decreasing by only $3.6 billion in 2023 compared to the previous year, which is significantly slower than the decline in revenue, dragging down the gross margin. In 2023, Intel's gross margin was only $21.7 billion, half of what it was in 2021.

After gross margin, market, sales, and management expenses did not change much, maintaining a level of $6-7 billion. But research and development expenses have remained high, reaching $15.1 billion in 2021, rising to $17.5 billion in 2022, and still remaining at $16 billion in 2023.

The continuous maintenance of high costs combined with the decline in revenue ultimately undermined Intel's profitability. In 2023, Intel's net income was 60% lower than it was in 2009 during the last financial crisis, and it was probably the darkest moment in the company's history.

How did the former global semiconductor leader fall into such a state?

In short, Intel is paying the price for past 'squeezing toothpaste' efforts.

Around 2013-2014, Intel was still a behemoth-like existence. Although it missed the wave of smart phones, it was still absolutely the 'big boss'.

The famous "Tick Tock" strategy, that is, two years to improve chip research and development, two years to improve advanced processes, made Intel a source of everything, monopolizing the PC chip market with one hand, and monopolizing advanced processes with the other.

And the old rival AMD was also struggling with debt and on the brink of bankruptcy, while Taiwan Semiconductor was still struggling to catch up with Intel's processes, not as powerful as it is today.

Since "Jiangshan" is already calling the shots, the enemy can no longer shake their position, so Intel decides to lie down, put away their weapons, and go into seclusion. This begins the "toothpaste squeezing" era, which is a slow research and development process that relies on previous technological advances.

By 2018, Intel's CEO had been replaced by Bob Swan, a financially minded executive. They say that your butt decides your actions, and Swan not only failed to make any progress in technological research and development, but instead focused on issues such as buybacks and increasing shareholder returns, which further compounded Intel's technological backwardness.

In the business world, it's like a battlefield, and you have to keep moving forward when the tide is against you. The seeds of decline are often planted in times of great prosperity.

While Intel was busy boosting ROIC through buybacks, their competitors were not idle. AMD, led by Su Ma, managed to climb out of their difficulties, and Taiwan Semiconductor was the first to make a breakthrough in ASML Holding's EUV lithography technology.

Suddenly, two of their biggest competitors ran ahead of them, and Intel entered a long period of decline, being beaten by newcomers.

Change of leadership, dramatic transformation.

The sudden outbreak of the pandemic in 2020, coupled with a large number of salaried workers working from home, led to a sudden surge in demand for PCs and data centers, easing Intel's pressure somewhat. However, problems are still problems and will not change because of a momentary change.

By early 2021, Intel's decline could no longer be concealed, and they could only replace financially minded Swan with the technically savvy veteran Pat Gelsinger, heralding the beginning of Intel's massive business transformation.

This transformation plan was very controversial at the time.

First, a hedge fund manager wrote a letter to Intel's management before the plan was released, suggesting that Intel emulate AMD and divest its manufacturing business, completely transforming itself into a fab company, i.e., a chip design company, and listing many reasons that would benefit both the company's future development and investors.

From the perspective of profit and return, what the fund manager said made sense. Even rumors spread that Intel's management was seriously considering this proposal, as AMD's success experience was ahead of it, and now the well-performing chip companies are mostly based on the fab model, which once ignited the capital market's enthusiasm for Intel's transformation, causing the stock price to rise for a period of time.

However, when the transformation plan came out later, this hedge fund manager, as well as the entire market, was surprised, because Gelsinger, the new leader, not only did not adopt the hedge fund's suggestion, but also went against the grain, heavily investing in chip manufacturing, which later became known as IDM2.0:

Under the plan, Intel will invest $200 billion in building wafer factories over the next decade, with facilities in the US, Europe and Asia, and will catch up with the world's most advanced process technology at the fastest pace, returning to the state where it dominated Taiwan Semiconductor.

Why did the story go the other way?

This has to be considered in conjunction with the geopolitical situation at the time. With the emergence of the US-China trade war, the autonomous security of the chip industry chain became the top priority for countries, especially major powers, and the US naturally did not let it pass, for which it threw out billions of dollars in chip subsidy policies.

Looking around the entire US chip manufacturing industry, only Intel was left as the only one. Gecheng, which was split off from AMD years ago, had already announced the abandonment of advanced process research and development, unable to bear the burden. Therefore, although Gelsinger knew he had to burn money, in the face of national interests, Intel had no choice but to risk it.

This leads to another question: Did Intel take the right step? Will this rebirth Intel or sink it deeper into the quagmire?

Strategic differences.

The market's view of Intel's IDM2.0 strategy is very divided.

On the one hand, the construction of a wafer factory is time-consuming and takes years to complete, and it takes several more years to achieve breakeven. Advanced processes take even longer. Where's the head start? On the other hand, even if the advanced process development and wafer factory construction go smoothly, will the customers that Taiwan Semiconductor has already won over be willing to switch to Intel? What's more, many customers are already competitors with Intel, such as AMD and Nvidia. At the same time, how long can Intel's stock price support the company's earnings power during the input period?

Looking at its later performance and stock price, the market is currently holding a negative attitude.

However, is Intel's strategic direction of focusing on manufacturing industries really wrong if viewed from a long-term perspective?

Actually, it's not necessarily wrong.

First of all, the fact that the US government values chip manufacturing localization so much is the biggest support for Intel. Although TSMC and Samsung also have factories in the US, these companies are not US companies, and they are not as secure as their own people. Especially TSMC, we won't talk about the regional situation behind it.

At the recent Q1 earnings conference, Intel announced that its 20A advanced process had been secured by the US Department of Defense, its biggest customer. It's needless to say why the US Department of Defense didn't give the order to TSMC.

Secondly, there are many semiconductor design companies in the United States, and the orders each year are very large, such as Nvidia, AMD, Apple, Tesla, and so on, all of which are TSMC users. If there is a US chip maker as an option, they can also re-select, even if they have worked with TSMC for a long time. Changing suppliers will bring additional costs, but from the perspective of supplier diversification and security, Intel can be completely used as a second option as long as Intel has the corresponding process and capacity.

Imagine that one day, if Intel really catches up with TSMC's process and capacity can be released, the company's operating difficulties can be reversed.

Is this possible?

Yes.

TSMC was able to surpass Intel in the past by daring to use a new lithography technology, which was the EUV lithography machine from ASML. This technology was risky at the time, but fortunately TSMC succeeded, and ASML's EUV lithography machine became a hot commodity sought after by global chip factories. We should know this best.

Now Intel is also starting to follow in TSMC's footsteps and buy a lot of ASML's EUV lithography machines. With Intel's R&D strength, it is very likely to make 5nm, 3nm, and 2nm. More importantly, there is a successful path to rely on, to possibly shorten the development time, which is also the reason why Intel's advanced process development progress was considered too hasty and even mocked for greedily grabbing subsidies and ignoring the company's future development.

However, all of this is only within our linear reasoning range, and there are many linearly unrelated factors in the development of things, let's wait and see.

Dare to copy?

It has been three years since the launch of IDM 2.0 in 2021.

Intel has invested a lot of resources for this, and with the decline of the epidemic, the decline in demand for PCs and data centers, the downturn in the semiconductor cycle, AMD's step-by-step push, the new AI, which has always been Nvidia's solo, and Intel has no products to enjoy the industry's development dividends, resulting in a drop in Intel's performance and stock price.

The former leader is now paying for the past's lying flat.

In the short term, Intel has nothing to boast about.

Manufacturing business still needs to continue investing and remains the biggest burden on profits and cash flow. Q1 performance was also very ordinary, and Q2 guidance was not particularly outstanding, so the stock price fell 9 points on the day of the earnings report.

In addition, ARM architecture continues to erode Intel X86 architecture's market share in the data center and AI PC sectors, including Microsoft's build conference a short while ago, where the latter vigorously launched AI PC products with Windows on ARM, which can also be regarded as a bad face for the decades-long wintel cooperation.

However, this year may also be the bottom for Intel.

On one hand, the company is expected to reach its peak investment in manufacturing this year, and the stand-alone accounting manufacturing business loss will also reach its maximum. The relevant output will be gradually released next year, contributing to revenue.

On the other hand, under the boost of AI, the PC business will gradually release the demand for replacement and become the main contributor to the repair of Intel's traditional PC business. Even with competition from ARM, the X86 architecture has been used since the late 1970s and still has a relatively large advantage in terms of software ecology. It is highly probable that the two architectures will coexist in the future and the possibility of being completely replaced by ARM is low.

At the same time, the competitiveness problem caused by the company's outdated process will be improved with the transition of the design department to Taiwan Semiconductor and the breakthrough of its self-owned process, regaining competitiveness. Therefore, the company maintains its annual guidance of shipping 40 million AI PC chips.

Now, Intel's market cap is only about 130 billion US dollars, with a dynamic PE ratio of 33 times, which is not in the same league as Nvidia (27 trillion), AMD (269.7 billion), and Taiwan Semiconductor (783.3 billion).

However, this also shows that Intel is actually not expensive.

If valuing by different parts, manufacturing is worth 30 billion by comparing with Qualcomm, and the remaining 100 billion is purely regarded as design business. If valued at 10 times, net income needs to reach 10 billion.

From 2010 to 2021, besides 9.6 billion in 2017, Intel's net income in other years exceeded 10 billion. If it really returns to its former glory, a net income of 10 billion is not out of reach.

Therefore, the final conclusion is that this year may be the lowest point for Intel. After surviving, with the gradual recovery of profitability, Intel's value will be rediscovered.

But it should also be reminded that this process may be relatively long and requires more patience. And what needs to be focused on now is Intel's performance next month and the guidance for Q3.

Editor/ruby