news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

intel, which reaped the benefits of the pc era, is now in its twilight years, while qualcomm and amd, which have seized the benefits of mobile internet and mobile phones, are now in the limelight. however, the future belongs to nvidia, which has demonstrated the possibilities of ai. the waves behind push the waves ahead, and the waves ahead die on the beach.

text｜li jian

in recent years, intel has been streamlining its business lines. currently, intel's main business units are the client computer group (ccg), which includes desktop and notebook computer businesses, the data center and artificial intelligence group (dcai), the network and edge group (nex) and intel foundry.

from a financial perspective, ccg and dcai are intel's core business units, contributing 80% of total revenue in the fourth quarter of last year. however, in the new wave of ai, these two units do not seem to be the beneficiaries, as their revenue has declined. in the second quarter of this year, intel dcai's revenue was $3.05 billion, a year-on-year decrease of 3.5%. the market was extremely disappointed with this performance.

currently, despite facing competition from amd, intel still dominates the cpu market. in the second quarter of this year, intel controlled about 76% of the global cpu shipments, while amd's share was about 24%. in the pc market, intel even occupies 79% of the market share, while amd's share is only 21%.

however, intel's market share is won by sales. although intel's shipments are more than three times that of amd, its profits are far less than amd. the cpu revenues of both companies are basically around $3 billion, but intel has greatly sacrificed profits through sales.

in the second quarter of 2024, intel achieved revenue of $12.8 billion and a net loss of more than 200% year-on-year. in this regard, intel ceo pat gelsinger said that companies have not fully benefited from ai, the cost is too high and the profit is too low.

in august 2024, intel also announced a 15% layoff of its global workforce, affecting tens of thousands of people. this move will cut costs by tens of billions of dollars for intel. kissinger recently said that the layoff plan has been completed by about half.

in the view of "digital intelligence research society", intel is now in its twilight years, but still ambitious. it was the dividends of the pc era that pushed intel to the forefront of the times. qualcomm and amd, as latecomers, grabbed the dividends of the mobile internet era and are now entering middle age. today's nvidia, like an 18-year-old boy, has shown the world the endless possibilities of the ai ​​era.

the waves behind the yangtze river push the waves in front of them, and the waves in front of them die on the beach.

intel's decline began to show signs in 2021. that year, intel achieved revenue of nearly $80 billion. since then, revenue has fallen sharply at a rate of $10 billion per year. in 2023, intel's annual revenue fell below $60 billion, a sharp drop of one-third.

intel cannot sit still in such a situation. it has been seeking the direction and possibility of business adjustment. in the context of the development of its two core businesses, ccg and dcai, intel has chosen to vigorously develop its foundry business.

although this business has benefited from subsidies from the us chips and science act, it has become a bottomless pit for intel to burn money. recently, intel said it is creating an independent entity for its foundry business to help intel raise external funds.

as we all know, for brands, foundry business is a bottomless pit that can only burn money and may not be successful. from the current perspective, foundry business has become the main factor dragging down intel's profits. market analysts believe that intel's expenditure on foundry business has exceeded us$50 billion in the past two years.

although intel's foundry division generated $4.4 billion in revenue in the first quarter of this year, most of this revenue came from within the company. intel adopted the so-called internal foundry model at the beginning of this year, that is, its product divisions and external customers will purchase manufacturing and packaging services from independent departments within intel. most of the $4.4 billion in revenue came from foundry for intel's own chips. in theory, it is not wrong to count this part of the revenue as false revenue.

in response, a u.s. law firm claimed that intel's foundry service growth and profits were false and called on investors to join the class action lawsuit against the company.

intel seems to be sticking to its decision and is still investing in chip foundry projects. behind this is the huge subsidy supported by the us government. the us government is increasing its investment in semiconductor production. in early september, the us government provided intel with $3 billion in funding in accordance with the chips and science act. the bill also shows that in march 2024, intel received $8.5 billion in subsidies and $11 billion in credit loans from the us government. this is the largest subsidy granted by the us government to a single company.

at the same time, intel announced a multi-billion dollar partnership with amazon's cloud computing division aws. the cooperation agreement shows that intel will produce ai chips for aws based on intel's 18a foundry process node in the next few years. 18a is intel's most advanced process to date, and the company hopes that this chip can challenge tsmc's 2nm process technology. this chip will be produced in a factory in ohio, usa.

kissinger said he would fight for intel's future, but the industry is not optimistic about intel's idm foundry model. the idm model means that companies not only design chips themselves, but also produce them themselves.

the cato institute, a u.s. think tank, analyzed that intel's long-standing refusal to accept the new division of labor in the semiconductor industry has led to today's losses.

according to intel, it will continue to invest $20 billion in wafer foundries next year. however, based on intel's current operating conditions and financial performance, as well as the current financing environment, intel's ability to continue to invest in wafer foundries will be directly affected, and the rumored buyer, such as qualcomm, may be one of the options in the future.

"digital intelligence research society" learned that the cost of building a wafer fab is extremely high. a 2nm factory with a monthly output of 50,000 wafers costs more than $28 billion, and a 3nm factory with the same capacity costs more than $20 billion. more importantly, money does not solve all problems.

therefore, after missing many opportunities for transformation, intel's fate may be in the hands of others. recently, it was reported online that qualcomm was in formal negotiations to acquire intel's business. as of now, intel's market value has fallen below $100 billion.

in the view of "shuzhi research society", it is a good thing for qualcomm to acquire intel, whose valuation has hit rock bottom. as the world's largest mobile phone chip manufacturer, qualcomm's acquisition of the world's largest cpu manufacturer will increase qualcomm's valuation.

although various analysts are not optimistic about intel's future, intel is still valuable as an asset. for example, although it has been surpassed by nvidia, intel is still a leading player in the data field and still has a strong influence in the global pc market.

however, in addition to market considerations, mergers and acquisitions are also subject to government antitrust review. in 2017, broadcom's acquisition of qualcomm failed because it was stopped by the us government due to suspected monopoly. later, qualcomm's acquisition of nxp also failed.

furthermore, in the view of "digital intelligence research society", even if qualcomm can successfully acquire intel, it will not take over the loss-making foundry business. the best player to take over the foundry business is likely to be tsmc.

in short, no technology giant can win forever. the times will abandon intel and qualcomm.