Intel missed three trains
Latest update time：2024-09-24
Reads：
Author | Gu Qiaoan, Editor | Evan
“
In the past 20 years, Intel has been at odds with the times in almost every key technological trend.
”
Pat Gelsinger has just become the eighth CEO in Intel's history in 2021. His first goal for the company after taking office was that Intel should become a leader in every business area in which it competes.
Three years later, not only did the imagined leadership fail to materialize, but as AMD seized the PC and server markets, and the industry shifted its focus from CPUs to GPUs that are more suitable for AI computing, Intel's market space suffered unprecedented squeeze. Compared with three years ago, Intel's revenue fell by more than 30%, cash flow turned negative, and debt exceeded $50 billion.
Pat Kissinger | Source: Market Realist
Since 2024, Intel's stock price has fallen by half. Its current market value is US$91.25 billion, which is only about 12% of Broadcom, 10.5% of TSMC, and 3.3% of Nvidia.
In September, Kissinger issued a letter to all employees, launching the company's biggest transformation in history to save itself. He had previously announced layoffs of 15,000 people and cuts of $10 billion in spending, and this time he made a move that was even more meaningful: spinning off the foundry business into an independent subsidiary, setting up independent directors, allowing external capital increases, and "more clearly separating and independent" from Intel.
The latest change is that on September 21, the Wall Street Journal and the New York Times quoted people familiar with the matter as saying that Qualcomm is negotiating with Intel on the acquisition, but Qualcomm has not yet made a formal offer to Intel. The two companies have not yet officially responded.
As the world's most powerful chip manufacturer at one time, and the only company currently equipped with platform, architecture, design, and manufacturing capabilities in the field of high-end logic chips, how did Intel get to its current situation step by step?
It seems that the rise of generative AI has made Intel the first giant to be crushed by the wheel of the times.
However, no snowflake is innocent in an avalanche. In the past 20 years, Intel has made four consecutive strategic decisions that are contrary to the trend of the times. Today, self-help measures such as the spin-off of the foundry business are more like a card that Intel must play in order to stay at the table, and it is too late.
01
The direct cause of Intel’s recent adjustments can be traced back to its second-quarter 2024 financial report released on August 2.
This is a financial report that can no longer whitewash. Data shows that Intel suffered losses beyond expectations in the second quarter of 2024. Although revenue only fell by 0.9% year-on-year, its net loss reached $1.61 billion, far exceeding the market's expected loss of $540 million. More importantly, there is no narrowing trend, and it is expected that the loss will continue to expand in the next quarter.
The decline in net profit is related to the decline in gross profit. Intel's gross profit margin for the quarter was 35.4%, far below the market expectation of 42.1%. Intel said that the decline in gross profit margin was mainly due to the increase in AI PC products, the adjustment of wafer production capacity and the increase in other non-core business expenses.
Intel mainly serves the microprocessor market such as personal computer servers.
In terms of sectors, Intel's business sectors include client business, data center and AI, network and edge domain,
its investment in autonomous driving company Mobileye and wafer foundry services.
Among them, the first two account for more than 80% of total revenue.
The client business is currently the only growing business segment, with revenue of US$7.41 billion in the second quarter, a year-on-year increase of 9.3%. However, the global PC is already a stock market and shipments are growing slowly.
Image source: Intel
What has attracted attention is the data center and AI business, which had revenue of $3.406 billion, down 14.9% year-on-year. In contrast, Nvidia's data center revenue grew 154% over the same period last year. In other words, the data center business, which is the core of Intel, not only failed to achieve the expected growth, but also saw a decline in revenue. This is in sharp contrast to the growth of the market demand for AI and high-performance computing, which is enough to prove that Intel has fallen behind in the AI wave.
In addition, the network and edge domain business revenue and Mobileye also declined to varying degrees. The wafer foundry service revenue was US$66 million, which is almost negligible if measured by the revenue standards of wafer fabs.
In other words, Intel's current situation is that its business base is shrinking and it has completely fallen behind in the competition for new technology trends. It is foreseeable that as AI competition intensifies, Intel, whose business is mainly concentrated in the traditional CPU and server fields, will only continue to lose market share to competitors such as AMD and Nvidia.
Along with this financial report, Intel announced that it would lay off 15,000 employees this year, equivalent to a 10% layoff, and would suspend dividends from the fourth quarter of this year, which is the first time Intel has suspended dividends in 20 years. As soon as the news came out, Intel's market value fell by more than $32 billion in one day.
Five days later, angry shareholders filed a class-action lawsuit in San Francisco, arguing that Intel "fraudulently concealed the company's problems," leading to a decline in performance and stock price.
02
Although shareholders who filed the class-action lawsuit say they were caught off guard by the company’s deception, the fact is that Intel was already lagging behind before the rise of generative AI technology.
If the development of a large company is compared to a relay race, key decisions are equivalent to the baton. Intel's fate is that in the past 20 years, it has stood on the opposite side of the times in judging almost every key technological trend. As a result, it has been in a catching-up state in every stage of the race and eventually fell out of the first echelon.
In 2006, Intel's management at the time made its first misjudgment and rejected the first generation iPhone as a foundry customer
, on the grounds that Jobs's extremely low bid was far below Intel's internal cost estimate. From the perspective of a business decision of a listed company, this choice was not unconvincing. Intel was in its heyday at the time, and its huge market share brought huge profits and resources, which were used to consolidate its leadership in the industry. Changing suppliers became costly and risky, and it was even more unacceptable to see a decline in profits.
Image source: Los Angeles Times
However, in retrospect, this choice was too costly: not only did Intel miss the most important opportunity to enter the mobile Internet, it also gave space to Samsung, which later became Apple's OEM, making it the largest chip foundry competitor in the future.
This kind of decision-making error has been repeated. Intel also failed to grasp the ARM architecture.
Because Intel insisted on making high-performance and high-power chips in the past, it was passive in the face of the low-power chip demand in the mobile market, and could not see the long-term value of the low-power ARM architecture. This led to the fact that although it explored the ARM architecture, it did not go deep, which ultimately further exacerbated its backwardness in the mobile market.
If the outside world still had illusions that Intel could overtake the competition when it lagged behind in the first two stages, then lagging behind in the competition for advanced processes became a watershed moment for the company.
In 2014, both Intel and Samsung achieved the production of 14nm process chips. However, TSMC had advanced to 10nm process in 2017. Intel was unwilling to adopt the latest EUV lithography technology, which resulted in its 10nm product being launched two and a half years later than TSMC. This also left Intel far behind TSMC and it has since fallen into the competitive echelon in advanced process.
The last major technology trend that Intel missed was that it failed to pay enough attention to the development of GPU. Intel relied on the success of CPU for a long time in the past, formed inertial thinking, and failed to discover the market demand for high-performance graphics processing in time. It was not until 2018 that Intel began to focus on the GPU business, but by then Nvidia had already cultivated the entire industry through the CUDA platform.
In addition to these four wrong decisions, Intel has another regret. Around 2017, OpenAI actively sought investment from Intel. This deal would not only bring financial returns to Intel, but also help it gain a first-mover advantage in the field of AI strategically. However, because the management at the time judged that generative AI technology would be difficult to scale in the short term, Intel eventually missed the most important ticket to the generative AI era.
Also in 2017, Intel entered the AI dedicated chip market by acquiring Nervana Systems. In the following years, Intel also launched several AI chip series, but these products have never made a splash in the market.
From missing the opportunity to cooperate with Apple, to ignoring the long-term value of the ARM architecture, to lagging behind in chip manufacturing technology, and to underestimating the development of GPUs, every mistake Intel made has laid the groundwork for its current predicament. In the early stages of the development of these technological trends, Intel also invested in pre-research and even took the lead for a time, but it ultimately failed to break free of the business inertia of large companies and the constraints of listed companies being responsible for the interests of investors and shareholders, and lost its lead. Under the current CEO Kissinger, Intel has faced huge challenges.
03
The reform measures that Kissinger took into account can be summarized into three main points: first, narrowing the chip process gap; second, using the US government's policy subsidies to build advanced wafer factories; and third, separating the chip design and manufacturing links to make the company more asset-light.
In terms of business model, by the time Kissinger was in office, Intel's integrated device manufacturer (IDM) model for chips had become the core of its pain.
Intel is the only IDM model company in the field of high-end logic chips. The IDM model refers to a company that simultaneously carries out the entire chain of chip design, packaging, manufacturing, and verification. Representative companies include Intel, Samsung, and Texas Instruments.
Intel once designed its own chips and manufactured them in its own factories, which was more conducive to cost control from a commercial perspective. However, with the continuous breakthroughs in chip manufacturing processes, the cost and difficulty of mass production of chips have also increased. The foundry model began to show its advantages. TSMC and other foundries have accumulated experience by serving multiple customers and gained more opportunities for technology iteration. As high-end process technology gradually matures and goes into mass production, foundries can charge a premium for advanced technology, thereby improving their profitability. This is sufficient for further expansion of the process and capital expenditures.
In contrast, under the IDM model, Intel's manufacturing customers are mainly itself, which limits its market adaptability and space for technological innovation. Intel has also long focused on mature markets and lacks the motivation to develop emerging technologies such as AI and mobile Internet. This in turn makes Intel more passive in manufacturing, with no extra profits to invest in the development of advanced processes, and the gap with industry leaders is gradually widening.
In the past two years, Intel's foundry business revenue has dropped from $27.49 billion in 2022 to $18.9 billion in 2023, and its operating losses have continued to expand from $5.2 billion in the previous year to $7 billion. It is foreseeable that if no measures are taken, the losses will only get worse.
Judging from the industry trend, the separation of digital chip design and manufacturing is almost a commercial necessity. First, the separation model of digital chip design and manufacturing is becoming more and more mature, and many digital chip designers do not even need to understand the process. Secondly, the R&D cost of the manufacturing process of high-end digital chips is getting higher and higher. Only independent foundries can widely accept orders to spread the cost, and at the same time, they can also further improve their manufacturing process and improve the yield by accepting orders widely. IDM factories do not have this ability and naturally have no advantage in the competition.
In the first quarter of 2024, Intel finally split its design and manufacturing businesses, and the business units became three product units: Client Computing Group (CCG), Data Center and AI Group (DCAI), and Network and Edge Group (NEX). Foundry became an independent operating unit, namely Intel Foundry, with independent financial statements.
In other words, the relationship between the wafer foundry business Intel Foundary and Intel's product departments has undergone a huge change: the product departments can choose foundries with more competitive advantages other than Intel Foundary; Intel Foundary will also be able to find external design customers at market prices and independently account for revenue from external customers and Intel products.
It seems that the independence of Intel Foundry can increase its possibility of finding external customers. The purpose is to improve capital efficiency and attract external investment to reduce losses and increase profits.
However, this idea also has obvious flaws: first, Intel is already facing serious problems of backward technology and products, and even a split does not solve the root cause; second, whether there will really be chip design companies that are willing to let a competitor with the same design business manufacture their own chips; finally, based on the above two points, the foundry part after the split will most likely still face the problem of insufficient capacity utilization, while the design part will face fierce competition from the outside world. In other words, the two may become more flexible or weaker after the split.
According to data from research firm TrendForce, Intel did not make the top 10 global foundry revenue list in the second quarter of 2024. In the past year, Intel briefly made the list in the third quarter of 2023, with a market share of 1%.
04
The most important move in Kissinger's recent letter to all employees is to split Intel's design and foundry businesses more thoroughly. In other words,
Intel has gone from being committed to becoming a leader in multiple fields to having to shrink in fact
.
After the second quarter financial report was released, Kissinger said, "We were previously trying to return to leadership, but now we have to focus on efficiency."
In addition, Intel's next important moves include: First, expanding cooperation with Amazon's cloud division (AWS). Intel has obtained a manufacturing contract for Amazon chips and will use its most advanced 18A manufacturing process (1.8nm node) to produce customized AI chips for Amazon.
The second is cooperation with the US military. According to the CHIPS Act, the Biden administration will provide Intel with an additional $3 billion in funding for the "Secure Enclave" program, which aims to expand the Department of Defense's supply of microelectronics products.
Image source: NBC News
The third is to suspend some factory construction investments in Poland and Germany.
Kissinger said, “We need to combine our cost structure with innovative operating strategies to achieve a fundamental transformation in the way we operate. At present, our revenue growth has not met our expectations, and we have not fully utilized the opportunities brought by cutting-edge technologies such as artificial intelligence. At the same time, our costs remain high, resulting in low profit margins.”
Next, there are two key points that will determine Intel's direction: one is the cooperation with Amazon. This cooperation will not only become a reference indicator for the outside world to Intel's competitiveness in the AI era, but will also bring Intel billions of dollars in revenue.
However, Reuters quoted an insider in early September as saying that Broadcom found that the yield rate of Intel's most advanced process 18A was not enough for mass production after testing it. Intel did not comment on this matter.
The second point is government funding. In order to reshape the US semiconductor manufacturing industry, the Biden administration proposed the Chip Act, and Intel is one of the only domestic manufacturers they can choose. In March of this year, the Biden administration gave Intel up to $8.5 billion in funding. However, the funds have not actually arrived, and as Intel's manufacturing business adjusts, it is generally believed that the funding will also change.
The company is already trapped both internally and externally. Intel's main business revenue has declined significantly in recent years, and its cash flow has been under pressure. At the same time, competitors have never stopped their efforts in R&D and market expansion, forcing Intel to push the company out of the quagmire and into a positive cycle in a short period of time. Whether it is the hasty launch of immature high-end products or its previous easy expansion and contraction in Europe, it shows that Intel has fallen into anxiety in the process of pursuing rapid success.
Intel's various self-rescue measures still need time to be tested.
The most certain effect at present is the layoffs. Intel's layoffs are already halfway through.
On many social platforms around the world, there are many posts about "graduating" from Intel. Some are angry, while others are regretful. An employee who left Intel a few years ago missed Intel's generosity to its employees. He said that compared with the high intensity of TSMC, working at Intel has freedom and good treatment. However, it was precisely because of this overly comfortable working environment that he felt the crisis in advance and chose to leave early.
-END-
京公网安备 11010802033920号