news

produced by huxiu technology group

author | wandushan

editor | miao zhengqing

header image | visual china

on september 21, the wall street journal published an article saying that qualcomm plans to acquire intel and has contacted the latter regarding the acquisition.

as of now, intel's market value is us$93.388 billion. if both parties confirm their intention, it will undoubtedly become the largest acquisition in the history of the semiconductor industry, and even the largest acquisition in the history of the technology industry.

more importantly, once the acquisition is completed, qualcomm will become the industry's first super chip company spanning mobile, pc, car and data centers.

but is it really possible?

let me first state the conclusion.the possibility of qualcomm acquiring intel is infinitely close to zero, and even acquiring a single business unit of intel is unlikely.

the most realistic situation is thatintel will sell a "secondary business unit" to qualcomm - just like it sold its mobile phone (baseband) modem business to apple.

standing at their respective crossroads

on september 16, intel ceo pat gelsinger issued a letter to all employees, focusing on the company's current reform plan, which includes laying off 15,000 employees worldwide by the end of this year; eliminating non-core business departments; and spinning off the chip foundry business and operating it as an independent subsidiary.

it is not difficult to see that these measures are all formulated around one goal, that is, to cut costs. after the release of the second quarter financial report, intel's loss of $1.6 billion in a single quarter was shocking, and its stock price plummeted 26% the day after the financial report was released. the core reason why intel fell into this predicament is the uncontrollable capital expenditure brought about by the two core strategies of foundry business and ai pc.

according to kissinger, the company's annual costs will be reduced by $10 billion by 2025.

however, this letter to all employees also revealed a key message: intel will not sell its wafer foundry business in a package for the time being. in other words, intel will still need to continue to provide capital to several wafer fabs that cost tens of billions of dollars.

this is not because intel is determined to hold on to the wafer fabs, but because there are currently no buyers who can take over these wafer fabs.

one piece of background information needs to be added. all of intel's domestic chip factories that start construction in 2022 are within the subsidy scope of the chips and science act. although this subsidy of up to us$8.5 billion has not yet been received, it is still difficult for intel to sell the wafer fabs directly to "foreign manufacturers."

against this backdrop, there are only two chip foundries in the united states that have the financial resources to take over: globalfoundries and texas instruments. however, these two companies have long since withdrawn from the competition for advanced processes and have no experience in developing advanced process-related technologies. therefore, they are unlikely to take the initiative to find themselves in a situation where they are "too big to fail."

this is also the reason mentioned at the beginning that qualcomm will never be able to acquire intel as a whole - although the chip foundry business will operate independently in the future, for intel or qualcomm, it is an asset that cannot be disposed of.

what if you just acquire a single business unit?

at present, in addition to the two independent departments of mobileye, the autonomous driving department, and altera, the programmable solutions group, intel also has three major business departments, namely ccg (client computing group, which can be understood as the pc chip department), dcai (data center and ai, which can be understood as server chips), and nex (edge ​​computing).

qualcomm is most interested in the ccg division. although intel is criticized as a "toothpaste factory", the company is still the absolute overlord in the pc cpu industry. according to the latest report from market tracking agency mercury research, intel's cpu market share in the pc side was 78.9% in the second quarter of this year.

on the other hand, qualcomm has also been trying hard to get a share of the pc cpu market in recent years.

at the microsoft build developer conference held in may, microsoft announced that the first batch of "copilot+pcs" products would all be equipped with snapdragon x elite/x plus chips, and almost all the well-known oem manufacturers came to the scene to support qualcomm.

now the question is: can qualcomm, which is currently in the limelight in the pc industry, take over intel's ccg?

one bite makes you fat

qualcomm's current market value ($188.1 billion) is almost twice that of intel, but a company's market value reflects investor confidence rather than the company's true value.

at least in the pc cpu industry, the value that intel ccg can generate cannot be underestimated.

on the one hand, ccg can bring intel a stable cash flow. looking through the company's financial report for the second quarter of recent years, you will find that when the company's overall loss was $1.6 billion, the ccg business unit contributed $2.497 billion in operating profit, while revenue was also able to maintain stable growth.

on the other hand, due to its size advantage, intel still maintains strong bargaining power among oem manufacturers.

not long ago, i had a conversation with the product manager of a first-tier oem manufacturer about qualcomm's pricing issues on cpu chips.

the other party said that qualcomm's initial offer for the x elite was higher than intel's, but the company responded directly that it was not sure whether the end market would be willing to pay for qualcomm chips. if the price of this chip could not be lower than intel's, there was no need to discuss cooperation between the two parties.

this story also reflects qualcomm's current embarrassing situation in the pc cpu industry: despite its good product strength and continuous investment in this field, it is still difficult to shake the foundation of the x86 architecture camp led by intel.

as for intel, what reason does the company have to sell its ccg business, which is likely to generate stable profits regardless of market conditions, when it is unable to dispose of its wafer foundry business?

in addition, from a regulatory perspective, major regions/countries in the world will launch antitrust investigations into this acquisition, accompanied by strict security reviews.

therefore, even if qualcomm is interested in intel ccg, it is difficult to directly bring it under its control. it is more likely to acquire a "secondary business department" to improve its own cpu technology reserves, such as the adaptation of the most typical arm architecture instruction set.