Welcome to the WeChat subscription account of “Sina Technology”: techsina
Text/Hernanderz
Source: ID: jiazhiyanjiusuo
“The brilliant superstar of the previous era is often the last person to adapt to changes. He is also the last person to succumb to the principle of strategic turning points. He will fail more tragically than most people.”
When I wrote this sentence in the book, I don’t know if Andy Grove, the legendary CEO and co-founder of Intel, would have expected that the company he founded would one day be trapped by transformation and become the frustrated giant.
On December 2, according to the British Financial Times, Intel will ask thousands of employees in Ireland to take unpaid leave for 3 months in order to reduce the company’s operating costs. According to public information, Intel’s Leixlip plant in Ireland currently employs about 4,500 people. If thousands of employees do leave their jobs, it will inevitably raise concerns about the day-to-day operations of Intel’s Leixlip factory.
Entering 2022, the prosperity of the semiconductor industry has taken a sharp turn for the worse, the PC market continues to be weak, and Intel has become one of the most severely affected semiconductor giants. For the immediate crisis, Intel is not without expectations, nor is it sitting still. Unfortunately, the hard work of the past few years has not paid off.
When the trump cards in the hand fail one by one, Intel may really have reached its darkest moment.
Profit and revenue plummeted
Throttling is the only option
It is not surprising that Intel Ireland employees have no shifts.
As early as the end of October, before the release of Intel’s third-quarter financial report, the news of layoffs of 20% spread like wildfire. According to foreign media reports, Intel’s move aims to do its best to cut operating costs and deal with the deteriorating PC market. In the previous earnings conference call, CEO Kissinger also publicly stated that Intel plans to cut costs by US$3 billion in 2023 and US$10 billion in costs by 2025.
The provision of three months of unpaid leave to Irish employees this time is believed to be part of Intel’s series of cost-cutting plans. It is reported that Kissinger had a video conference with his employees last month, frankly admitting that the company has problems such as high costs and plummeting profits.
And it can also be seen from the latest financial report that Intel’s problems cannot be solved by taking unpaid leave for thousands of employees.
Data show that in the third quarter of this year, Intel’s total revenue was US$15.338 billion, a year-on-year decrease of 20.08%; net profit was only US$1.019 billion, a year-on-year decrease of 85%. At the same time, various data such as gross profit margin and diluted earnings per share have declined, and the performance can be described as quite bleak.
Compared with historical data, Intel’s revenue has declined for three consecutive quarters, and the decline has continued to expand. In the first and second quarters of this year, its revenue was US$18.353 billion and US$15.321 billion, respectively, a year-on-year decrease of 6.71% and 21.96%. The situation on the profit side is even worse, with a rare loss of US$454 million in the second quarter.
For the fourth quarter performance expectations, Intel is also quite pessimistic. Intel expects total revenue in the fourth quarter to be US$14-15 billion, far lower than the 20.528 billion yuan in the same period last year; the gross profit margin is expected to be 41.4%, a decrease of 1.2 percentage points from the third quarter. If the revenue in the fourth quarter is in line with this expectation, then Intel’s full-year revenue in fiscal year 2022 will be fixed at US$63-64 billion, and diluted earnings per share will be about US$2, which is lower than market expectations.
From the perspective of revenue structure, the performance of the two revenue pillars of client computing business (CCG) and data center and artificial intelligence business (DCAI) in the third quarter was lower than expected. Among them, CCG’s total revenue was US$8.124 billion, a year-on-year decrease of 17%; DCAI’s revenue recorded US$4.209 billion, a year-on-year decrease of 27%.
The decline of the two main businesses is undoubtedly attributed to the continued weak PC demand.
According to the latest data from IDC, global PC and tablet PC shipments will drop by 11.9% to 457 million units in 2022, and will continue to decline on this basis by 2023. The report pointed out that the average annual compound growth rate of the PC market in the next five years is about 1%, and demand is expected to improve only in 2025-2026. At present, global suppliers are facing the impact of excess inventory and declining turnover rate, and the PC winter will continue for a long time.
(picture from IDC)
Of course, Intel is not the only one facing difficulties. The global technology and semiconductor industries are suffering. According to statistics from Crunchbase, a US data agency, as of November this year, the US technology industry has lost a total of 67,000 employees throughout the year, and industries such as the Internet, semiconductors, and cryptocurrencies are the hardest hit areas.
On December 2, the foundry company GlobalFoundries announced the layoff of 800 employees, accounting for 5.3% of its total employees; the American chip design company Micron Technology announced in October that it would abolish its overseas teams, and almost all the R&D teams in China were laid off; An American chip giant Texas Instruments also announced the abolition of the Chinese MCU R&D team in the second quarter of this year, and only a small number of engineers were transferred to the US headquarters.
The “Washington Post” analyzed in the latest report that the Federal Reserve appears to be less and less confident about whether it can avoid an economic recession. Federal Reserve Chairman Powell’s ambiguous speech at the press conference of the last federal market meeting made investors, companies All feel extremely worried. Focusing on the semiconductor industry, the contrast between last year’s glory and today’s desolation is even more embarrassing.
Revenues and profits have fallen sharply, and the PC market will not improve for a while. Intel has no choice but to lay off employees in disguise to reduce expenses. It’s just that to boost performance, we still need to increase revenue, and reducing expenditure cannot solve all problems.
So does Intel have any other backers?
Intel has no second hand:
Autopilot stalls, wafer foundry faces enemies
As mentioned at the beginning of the article, Intel has not thought about the future, and has also bet on several outlets in autonomous driving, graphics business and foundry. But the embarrassing thing is that Intel has passed by the glory days of these industries time and time again, and instead increased investment at inappropriate times, which eventually resulted in a wrong step.
Mobileye, an autonomous driving company, was once Intel’s most promising second pole of growth. Today’s performance cannot be said to be a failure, but it does not meet expectations. The lack of computing power of intelligent driving chips and failure to keep up with the development trend of sensor fusion sensing technology are the main reasons why Mobileye lags behind.
Statistics show that although the computing power of Mobileye’s first-generation product EyeQ4 is only 2.2TOPS, it was already at the top level in the industry at that time, and won the pursuit of many car companies such as Xiaopeng and Weilai. But in 2022, Nvidia’s chip Xavier has a computing power of 30TOPS, Tesla’s self-developed FSD is as high as 36TOPS, and it will explode the new generation of products such as EyeQ5 and EyeQ6L developed by Mobileye.
Although Mobileye has started to develop the next-generation ultra-high-performance chip Eye Ultra, the computing power is expected to reach 176TOPS, but Nvidia’s high-end product Orin can reach 200TOPS, and it is expected to be launched earlier than Eye Ultra.
The share price of Mobileye, which was successfully listed this year, has ushered in a wave of skyrocketing, which makes people see hope. Unfortunately, as of press time, its market value is fixed at 25.7 billion US dollars, which is still far from the original valuation of 50 billion US dollars. Mobileye has a long way to go if it wants to catch up to its once high hopes. It is even more difficult to catch up with competitors such as Nvidia.
As for the new business foundry, which is another key development, it is also ill-fated.
Intel’s foundry business will begin in March 2021, Kissinger’s second month in office. This new CEO with a strong technical background is a loyal supporter of “Moore’s Law” and believes that the foundry business has great potential. Intel’s technical accumulation and experience in chip design, manufacturing and other fields have laid a solid foundation for entering the foundry industry.
Under the impetus of Kissinger, Intel invested 20 billion US dollars to build two foundry factories in Arizona, USA, and established a foundry service business group (IFS), mainly for customers in the European and American markets. The head of IFS is technology tycoon Randhir Thakur, who reports directly to Kissinger.
However, Intel still overestimated its own capabilities and misestimated the prospects of the foundry industry.
On the one hand, Intel’s wafer foundry business is not progressing smoothly, and technology and yield are facing many doubts. The latest financial report shows that Intel’s foundry service revenue in the third quarter was only US$171 million, not only unable to compare with giants such as TSMC and Samsung, but even showing a 2% year-on-year decline.
Intel’s practice of abandoning the nm naming method and naming future generations of advanced process products with Intel 7, Intel 4, Intel 3 and Intel 20A also caused a lot of controversy. According to Intel’s plan, Intel 4’s benchmark is the 7nm advanced process, which will become Intel’s first process node using full EUV lithography technology, and is expected to be put into production in the second half of 2022. But until now, no relevant progress has been seen.
On the contrary, it is the change of internal personnel, which highlights the chaos of Intel’s foundry business. In November of this year, it was reported that Randhir Thakur had resigned as the head of the IFS business, and then Intel officially confirmed this statement. It is a big taboo for military strategists to be easy to be handsome. Intel’s wafer foundry business is full of twists and turns, and it is difficult to catch up with TSMC and Samsung.
On the other hand, demand from upstream customers such as PCs and smartphones has declined, and the foundry industry has not had such a good time.
Since the second half of this year, news of wafer foundry companies being cut orders has emerged in an endless stream: AMD cut 20,000 TSMC 7nm and 6nm advanced process chip orders, and Apple reportedly cut TSMC 10% advanced process orders. Samsung finally conquered the 3nm process, but was exposed to a lack of customers, and the capacity utilization rate dropped sharply.
Compared with other companies, the situation of Samsung and TSMC is not bad, at least there is still no shortage of business. However, the Matthew effect in the industry has intensified, and orders and customers are concentrated in leading companies, which is definitely not good news for new players like Intel.
The two carefully prepared back roads are blocked, what should Intel do next? Should I find another way, or return to my old job? Or take the trail to find an overtaking opportunity?
Judging from Kissinger’s recent speeches and the latest developments, Intel still believes in its old path: using technology to create new needs.
Sworn to defend “Moore’s Law”
Intel’s Last Stubbornness
Since Kissinger took office, technological innovation has become the focus of all Intel’s work. Recently, Intel has intensively released a variety of new products and technologies in an attempt to stimulate the market and user needs.
At the Intel On Technology Innovation Summit held at the end of September this year, Intel released the 13th generation Core processor product family, including 22 processors and 125 partner models including 6 unlocked desktop processors.
At the IEDM 2022 IEEE International Electronic Devices Conference, Intel also announced three new technologies including 3D packaging quasi-single chip, ultra-thin 2D material single chip integrating more transistors and 300 mm diameter gallium nitride wafer on silicon. According to the official information released by Intel, the next-generation 3D package quasi-single chip will increase the integration density and performance by 10 times, and the connection density and bandwidth can even be comparable to the single-chip SoC that is now popular all over the world.
People in the industry can see at a glance that Intel’s new technologies have a common goal: to increase the number and density of transistors. After all, Intel is swearing to the death to defend the “Moore’s Law” that has led the development of the semiconductor industry for 60 years.
As we all know, Intel co-founder Gordon Moore is the creator of “Moore’s Law”. Over the past few decades, “Moore’s Law” has also been regarded as the golden rule of the semiconductor industry, and Intel is a loyal supporter of this theory.
But now, the discussion that “Moore’s Law” is failing is endless, and even set off several rounds of debate between Intel and Nvidia.
In September of this year, Nvidia founder Huang Renxun said that “Moore’s Law” is dead. For the semiconductor industry, it has become an unrealistic expectation to achieve twice the performance expectation with fixed costs. At the opening ceremony of Intel Innovation not long after, Kissinger directly refuted Huang Renxun’s statement, insisting that “‘Moore’s Law’ is not dead, it has been alive and well.”
To some extent, the “Moore’s Law” battle between Intel and Nvidia, the two semiconductor giants, is also a battle between two routes, and it is Intel’s struggle to resist outside doubts and save the current decline.
Semiconductor upstarts such as Nvidia promote new technologies and materials such as artificial intelligence and non-silicon semiconductor materials, hoping to reshape the path of industry technology improvement. Huang Renxun has always emphasized the importance of architecture innovation and full-stack innovation. Nvidia’s GPU technology is the best solution to overcome the soaring cost of silicon wafers and the difficulty of advanced process technology research and development.
Intel believes that “Moore’s Law” has not failed, but the direction of technological breakthroughs has been changing. Since the 2010s, packaging and transistor expansion technologies have become the protagonists. Intel is now focusing on packaging technology, and is working hard to increase transistor density through embedded multi-chip interconnect bridge (EMIB) and Foveros technology.
Intel still seems to believe that as long as it can continue to make breakthroughs in technology, it can stimulate a rebound in market demand, or create new demand for consumers. As long as this belief does not change, Intel’s pursuit of “Moore’s Law” will not change.
But is this Intel’s own insistence, or the consensus of the entire industry? I am afraid that Kissinger did not dare to jump to conclusions.
Interestingly, Gordon Moore said in an interview with The New York Times in 2015 that “Moore’s Law” will not last forever.
In the final analysis, the market is unpredictable, whether “Moore’s Law” will completely fail, and whether Intel’s “stimulating the market with technology” route can succeed, no one can draw a conclusion. However, for the semiconductor industry, it is never wrong to delve into technology. Intel, which insists on its own technical route, definitely has the hope of a comeback.
write at the end
Intel, which puts all its future on technology, and the current CEO Kissinger can be said to be a match made in heaven.
In the first 30 years of working for Intel, Kissinger has been deeply involved in the technical field, all the way to the position of Intel CTO. The 12 years of running away in the middle did not cut off his relationship with his old club. He became the eighth CEO of Intel. Efforts to pull this giant ship back to the road of technological innovation.
Of course, today’s market environment and technological trends are very different from those of the past, and there are many doubts about whether Intel can regain its status through technological innovation. But Kissinger, who has participated in the development of more than a dozen microprocessor projects and promoted breakthroughs in USB and WiFi technologies, still believes that technology can change the world and save Intel.
Returning to technology is Kissinger’s plan for Intel, and it is also Intel’s hope for rebirth.
(Disclaimer: This article only represents the author’s point of view, not the position of Sina.com.)
This article is reproduced from: https://finance.sina.com.cn/tech/csj/2022-12-06/doc-imqqsmrp8722885.shtml
This site is only for collection, and the copyright belongs to the original author.