Per reporter: Yang Hui Per editor: Dong Xingsheng
On the last day of the Spring Festival holiday in the Year of the Rabbit, the chip industry opened a round of "earthquake", the source of which was even more difficult than expected. According to data disclosed by Intel, in fiscal 2022, its revenue was $63.1 billion, down 20% year-on-year; Net income was US$8.0 billion, down 60% year-over-year. Among them, the loss in the fourth quarter of 2022 was $661 million, and the loss is expected to continue into the first quarter of this year. A number of analysts in an interview with the "Daily Economic News" reporter analyzed that Intel is too late to start the mobile terminal, coupled with the early ecological management strategy is too closed, the new field innovation is insufficient, its mobile chip architecture including the ARM architecture has obvious disadvantages, even in the X86 chip architecture, the "self-reserved land", but also suppressed by AMD.
Despite the current passive situation, Intel is still considered to have a chance to turn around. Chen Jia, an independent international strategy researcher, pointed out that as long as the force is correct, Intel has the opportunity to achieve the X86 architecture to reverse the mobile architecture, the key is whether it can get out of the comfort zone created by its own patent barriers, and at the same time, Apple, the mobile chip design leader, has begun to encroach on the X86 architecture and server market in the early stage, matching a strong enough counterstrategy. Otherwise, at the current market falling rate, relying on the chip design process to eat the old capital will soon be surpassed by competitors.
PC chip inventory backlog
"The tough environment was exacerbated by excess chip inventories, with both the company's fourth quarter (2022) results and (this year's) first quarter guidance below expectations." Intel CEO Pat Gelsinger said in a conference call after the earnings release.
On January 27, Beijing time, Intel released its fourth quarter and full year 2022 financial report. In the fourth quarter of fiscal 2022, Intel achieved revenue of $14 billion, down 32% from the same period last year; Net loss was $661 million, compared to $4,623 million in the year-ago quarter.
More seriously, compared with the current performance, the fog that cannot be cleared in front of the market is more worrisome. Intel's first-quarter 2023 performance guidance in this earnings report is revenue of $10.5 billion to $11.5 billion, far lower than the previous market expectation of $14 billion, and the first quarter will also be difficult to escape losses.
Affected by this news, on January 27, before the US stock market, Intel fell nearly 10%, down 6.41% as of the close, and the market value lost about 8 billion US dollars (about 54.3 billion yuan). Early on the morning of January 28, terms such as "Intel Thunder" appeared in some media reports and social platforms. In this regard, on the afternoon of January 28, Zhang Xiaorong, president of the Deepin Science and Technology Research Institute, told reporters through WeChat that the "sell-off" was triggered by poor performance and is not expected to last.
For the chip industry, although the industry has long felt the chill, Intel's performance still made everyone shiver.
Chen Jia bluntly said that even if the valuation of technology stocks has experienced a harsh winter, Intel's challenges are extremely prominent, combined with the low growth expectations of the entire PC market in the future, it is expected that this round of Intel's valuation loss will be about $10 billion.
There are two reasons given by the earnings report and Intel management: declining sales of personal computer (PC) chips and fierce competition in the server hardware market.
Sure enough, the fading of the PC market dividend had to be whipped. According to the financial report data, Intel's customer computing group, including the PC chip business, had net revenue of $6.625 billion in the fourth quarter of fiscal 2022, down 36% year-on-year. Kissinger also revealed on the conference call that in the short term, due to the destocking of PC terminal manufacturers, Intel's shipments in 2022 will be 10% lower than market demand. PC shipments are expected to be 270 million units in 2023 due to a larger-than-expected deterioration in market demand.
It's not just Intel, the entire PC industry has struggled with a backlog of inventory over the past year.
According to research firm Gartner, PC shipments fell 28.5 percent in the fourth quarter of last year, the worst pullback since Gartner began tracking the market in the '90s. In Zhang Xiaorong's view, although the PC market is still sluggish, it is expected to improve in the second half of the year, and next year may tend to normal, and Intel's performance is expected to improve.
However, Xiang Ligang, a well-known observer of the communications industry, pointed out in an interview with reporters on WeChat that under the influence of shrinking demand, it is still difficult for Intel to change the current pattern quickly, and it is expected that it will still be in a relatively difficult stage for a period of time.
Process + expansion still "burns money"
The industry has entered a cold winter, and there will only be more and more companies tightening their belts to live.
In October last year, news of "Intel layoffs" began to circulate on social platforms, and the specific layoffs were confirmed by the management in this conference call. According to Kissinger, Intel is making the tough decision to resize the organization and further strengthen the business focus of its business units by rationalizing its product roadmap and investments, with layoffs around 10,000 people. This also means that Intel has officially joined the "wave of Silicon Valley layoffs".
In addition to layoffs, Kissinger also revealed that in the future, the company will stop investing in network conversion product lines and focus on serving existing customers and existing products. In fact, since the second half of 2021, Intel has successively closed several businesses, including the Optane memory business. Kissinger said the shutdown of seven companies has cut spending by more than $1.5 billion for Intel.
These measures are only part of the cost reduction. Intel hopes to save $8 billion to $10 billion by 2025. Kissinger said that at the beginning of 2022, Intel experienced the worst supply tight environment in more than 20 years, which made companies have to adjust in the second quarter, that is, to create an internal foundry system, which will become a key factor in Intel's "tens of billions of money saving plan".
Under a number of measures, according to Kissinger, Intel directly offset the expected expenditure of $13 billion in 2022.
However, breaking down the financial report data, Intel's funds are actually "the province, the flower", and the investment is not ambiguous. Specifically, in 2022, Intel's total capital expenditures were $24.8 billion, up 33% from the previous year, accounting for 39% of Intel's total revenue in 2022.
There are several main parts to the "money spent on the cutting edge": expanding the factory and advancing the manufacturing process. However, it has recently been reported that Intel's chip factory construction plan may be affected by costs. In response, Kissinger responded on the conference call that Intel will not reduce long-term strategic investment, nor will it reduce the upstream and downstream capability capital investment of the industrial chain related to business needs.
As for the process progress, the "Daily Economic News" reporter learned from Intel that Intel will still be committed to achieving a long-term strategic roadmap, including completing five process nodes within four years. Intel 4 is ready for manufacturing and is expected to begin ramping up Meteor Lake capacity in the second half of 2023. On the Intel 20A and Intel 18A, Intel's first in-house test chips, as well as chips from a major potential foundry customer, have tapeed out along with the products being manufactured and are expected to regain the lead in transistor performance and power supply performance by 2025.
It should be noted that in the view of Zhang Xiaorong and many other industry insiders, the new chip foundry business developed by Intel is still in the investment period, has not yet produced an upward performance pulling effect, and will not have an impact on leaders such as TSMC in the short term. If Intel wants to improve the level of chip manufacturing, it is likely to increase its efforts to introduce TSMC's talents and technology, and at the same time increase its efforts to lobby the US government to introduce more chip policies that are conducive to its own development.
Chen Jia said bluntly that the current difficulties in the technology industry far exceeded expectations, coupled with the continued deep decline in European and American market demand, the US-led chip sanctions strategy against China is still fermenting and other factors, in addition to a small number of new energy vehicle chips, the overall growth momentum of the global chip industry chain this year will continue to slow down, Intel's chip strategy will take at least a few years to see preliminary results, and may not be positive.
How to break through PC architecture chips?
Money-burning products will undoubtedly need more transformation and rising projects to support, and Intel will need to accelerate the speed of nuggets in the future. In addition to the focus of chips, although the overall performance cannot be recovered, there are also some outstanding business performances in this financial report, such as Kissinger's foundry business after returning to Intel.
According to financial report data, in the fourth quarter of 2022, Intel Foundry Services (IFS) achieved revenue of $319 million, a year-on-year increase of 30%; Its Programmable Solutions Group (PSG) revenue grew 42% year-over-year, Mobileye's fourth-quarter revenue was $565 million, up 59% year-over-year, and design projects, including SuperVision, are expected to generate approximately 64 million systems and approximately $6.7 billion in revenue in the future.
Returning to Intel's old chip business, the current "main contradiction" is another issue mentioned in its earnings report: market competition. Although there are not many opponents, the intensity is not inferior to the popular track.
Kissinger took questions from analysts on the call and said that the company's data center market is being cannibalized by AMD, and the share of Intel products is declining. "We faltered, lost market share, lost momentum. We think it will stabilize this year. Kissinger said.
In Chen Jia's view, Intel is too late to start the mobile terminal, and its X86 architecture has obvious disadvantages compared to the ARM architecture. In addition, the early Intel ecosystem management strategy was too closed, and there was a lack of targeted response strategies under the newly emerging energy-efficient chip architecture and application scenario layout such as Qualcomm. "Although Intel has also recently worked on mobile chip architectures including ARM architecture, the latecomer disadvantages of the chip industry are very obvious, and Qualcomm and other institutions have grasped the initiative through patent barriers, whether it is R&D overtaking or building an ecology are difficult." Chen Jia said.
Talking about the constraints of the strategic model, Xiang Ligang also believes that a major problem of Intel is that it sticks to the basic system for a long time, and the breakthrough in new fields is not enough. "Intel still mainly relies on PCs and traditional servers, and lacks innovation in new fields, such as artificial intelligence and the Internet." Xiang Ligang said.
In addition, even in the "private land" of the X86 chip architecture, Intel has been constrained in recent years. "In the PC market, the pattern of Windows dominating the world was broken by macOS, ChromeOS and other systems, and the Wintel ecological advantages created by Intel in the early years have bottomed out, except for a small number of e-sports just needs, in most application scenarios, Intel is facing challenges, product lines and cost-effective advantages are not obvious."
Chen Jia pointed out that from the perspective of Intel's current strategy, the future may follow up long-term technology investment on the one hand, and integrate and shrink core business segments on the other. "But this also involves two key points", one is whether Intel needs to adhere to the X86 architecture and how to reflect on the long-term first-mover advantage in the product side by relying on its product design and development and market advantages, through the M1 and M2 series chips to eat the PC market, and how to reflect its long-term first-mover advantage on the product side, including cooperation and competition with the old rival AMD, and protect against strong enemies in the field of consumer electronics.
The second is how to deal with the relationship of competition and cooperation with Qualcomm and other global chip industry chain leaders such as Qualcomm in the field of baseband chips and mobile architecture chips. Further, whether it is necessary to dock the mobile chip process technology.
"Today, the naming convention for the N-nanometer process itself is derived from Intel chip manufacturing standards and refers to Moore's Law. But Moore's Law has long been invalid, and Intel's current X86 architecture process technology and AMR architecture process technology are completely different standards. Chen Jia said that the overall 3 nm, 5 nm and other process processes of the mobile terminal are more of a labeling method for chip design generations, not a real process value.
"This actually provides an opportunity for Intel to snipe midstream manufacturers such as TSMC in the process process, as well as head design companies such as Apple and Samsung." Chen Jia said that since Apple can realize the interconnection between the X86 architecture and the ARM architecture, Intel has the opportunity to achieve the X86 architecture to reverse the mobile architecture as long as it is correct. The key is whether it can get out of the comfort zone it has created through patent barriers, and match a strong enough counter-strategy in the early days when Apple has begun to cannibalize the X86 architecture and server market.
"Otherwise, with Intel's current market falling speed, relying on the current chip design process to eat the old capital will soon be surpassed by competitors such as Apple."
Daily economic news
![Goodbye, old system! | Geek time[fig]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)