news

author｜xinxin

editor｜jingyu

intel, the undisputed giant of the global semiconductor industry, is in trouble.

intel's stock price has halved over the past year, making it one of the worst-performing technology stocks in the s&p 500. the company's latest quarterly financial report also disappointed investors, and expectations were tepid, causing the market value to evaporate by more than $30 billion.

not long ago, there was even news in the market that qualcomm was interested in acquiring intel. although it is not finalized yet, qualcomm has only made a proposal, but it also shows intel's dilemma.

looking back on the past two decades, intel can be said to have missed a series of key opportunities, including ignoring the explosion of the mobile phone market, being slow to apply euv technology, and canceling the general-purpose gpu project early, resulting in its current failure to keep up with the artificial intelligence boom. .

how did intel, once a marvel of engineering technology, fall to its current situation step by step? what enlightenment can its failure bring to people?

miss the times

the first key mistake,missing the wave of the mobile computing revolution is intel's most high-profile setback。

in 2007, apple released the iphone, and the smartphone quickly became a global computing platform in the following years. however, despite dominating the pc era, intel stuck to its x86 architecture and failed to seize the opportunity to provide chips for mobile devices. the then intel ceo reportedly rejected apple's chip order for the first-generation iphone because apple's bid was lower than intel's budgeted cost.

subsequently, competitors based on arm architecture such asqualcomm, mediatek, with its energy efficiency advantages and low cost, quickly dominates the mobile field. although intel launched the atom processor in 2008 in an attempt to enter the mobile market, it was basically unsuccessful due to poor energy efficiency and performance.

the company canceled its smartphone business outright in 2016。

misjudgment in the mobile field not only cost intel billions of dollars, but also paved the way for the rise of companies such as qualcomm. moreover, the arm architecture, once limited to mobile devices, has now even spread to intel's traditional laptop and server markets.

qualcomm’s rise in the mobile field｜image source: qualcomm

the missteps in the mobile market are just the beginning.

intel's second mistake was a mistake in technology decision-making, and it occurred in the manufacturing area that it is most proud of.。

intel is hesitant to adopt the advanced process of extreme ultraviolet lithography (euv) technology, preferring to continue to use its own original technology, causing its process technology to gradually fall behind in the competition with competitors tsmc and samsung.

ironically, this expensive chip manufacturing process was originally funded by intel itself.because intel acquired some shares of asml in 2012, but intel’s ceo at the time was unwilling to use its newly launched euv equipment。

since intel co-founder moore proposed "moore's law", intel has been promoting smaller and more efficient processes. in the post-leader era, this advancement has encountered unprecedented delays. in the transition to smaller processes, the advantages of past manufacturing processes appear to be stretched thin. the 10nm process was originally scheduled to be launched in 2016, but experienced multiple delays and did not come out until 2019. the subsequent advancement to the 7nm process also encountered delays.

at the same time, tsmc and samsung have ASML it purchases euv equipment in large quantities, continuously reduces the process size of chips, improves chip efficiency and performance, and seizes the market for advanced technologies. amd also uses tsmc's manufacturing process to launch chips with superior performance. for intel, the price of delay is painful, as its technological advantages gradually disappear and its competitors gradually eat away at market share.

in addition,mistakes in manufacturing also lie in decisions made in the oem business。

in the past, intel has stuck to its idm model, which was its advantage in the glory days by controlling both design and manufacturing. however, with the rapid changes in the market environment, this persistence has become a shackles for the company. compared with competitors such as tsmc and samsung that focus on oem manufacturing, intel's model limits its flexibility in the supply chain and makes it difficult to adapt to changes in market demand.

jealous of tsmc's huge profits from its foundry model, intel finally decided to copy this model in recent years and launch intel foundry services. however, this decision was like missing out on the mobile market.it came too late. not only did we not make any money, we even lost billions of dollars in the contract manufacturing business last year., only 1% of the revenue from oem business in the first half of this year came from external customers.

a view of intel foundry｜image source: intel

just as intel continued to decline, even apple chose to abandon intel.

in 2005, apple ceo steve jobs publicly embraced then-intel ceo paul otellini at the worldwide developers conference and announced that mac computers would fully switch to using intel chips. at that moment, intel leaders were very proud.

but in 2020, apple announced that it would abandon intel processors and switch to self-developed m series chips based on arm architecture. intel has since lost a large part of the high-end market for personal computers, further weakening the influence of the x86 architecture.

jobs launched the mac with intel chips in 2005｜image source: apple

intel's third mistake was to give up on gpu graphics cards. as a result, when ai became popular, the company was completely misfired and could not reap any dividends from artificial intelligence.。

intel made an early attempt to enter the gpu market with the larrabee project, but the project was canceled in 2010, along with intel's entire gpu plans. the consequences of this decision were officially revealed ten years later, with intel missing out on the gpu and ai markets.

this now makes intel's current ceo pat gelsinger very unhappy. he worked at intel for 30 years from 1979 to 2009. he was "squeezed out" of the company a year before the gpu project was canceled. in his view, if larrabee had been launched back then, there would have been no need to spend huge sums of money to acquire some ai companies, and intel would not be so lagging behind in the ai ​​arms race.

it is said to be a prototype of intel's larrabee gpu and was photographed by a collector on ebay in france | image source: leodanmarjod

why are some projects and plans being cancelled? in addition to the judgment of strategic technology and capability issues, another problem for large companies is thatfor the sake of financial indicators, the professional managers at the helm of the company are very focused on short-term returns and spend hundreds of billions of dollars on dividends and buybacks instead of investing in research and development., and, in order to increase profits and cut costs, projects that were not yet profitable at the time were often cut off, or layoffs were laid off.

in the past two decades, intel has conducted many large-scale layoffs, including tens of thousands of layoffs in 2006, 2016, and 2022, as "part of a cost-saving plan." in the early years of layoffs, many key engineers involved in the development of next-generation processes and architectures were laid off. the industry was once considered to have caused long-term damage to intel, especially its innovation capabilities, making it unable to cope with technological changes.

at the same time, while intel continued to miss opportunities, competitors never stopped. amd is making a comeback in the cpu market, tsmc is producing chips for top companies inside and outside the industry with its pure foundry model, and nvidia is dominating the gpu and ai fields.

intel has acquired several ai companies, such as nervana and deeplabs, but many of these acquisitions have not yielded the expected results. for example, nervana's products were canceled just a few years after the acquisition.

how to recover?

at the moment, intel can be said to be in a lot of troubles. there are rumors that intel may be acquired from time to time. in september this year, qualcomm was rumored to be a major potential buyer.

although intel has not confirmed relevant plans, the news that this giant that once dominated the market may be acquired shows the gap from its peak to its struggle.

intel's current ceo pat gelsinger returned to the company in 2021. his strategy is to invest heavily in manufacturing capabilities and refocus on technological innovation.to this end, intel invested us$100 billion to build a new factory in the united states and promised to open foundry services to the outside world.。

intel also recently took the important step of spinning off its foundries into independent entities. in addition to poor business performance, the reason for the separation is to improve operational flexibility and facilitate access to potential external funds. more importantly, it may be to let potential external customers believe that intel is willing to use the most cutting-edge technology to provide services for the company. "outsiders" make chips.

in the past, intel reserved its most advanced node technologies for its own processors, but pat gelsinger believes this was a mistake. he is very concerned about intel's foundry business and plans to continue investing even if it is suffering serious losses.

intel’s current ceo pat gelsinger｜image source: intel

intel has also begun to increase investment in artificial intelligence and next-generation chip design, including the launch of the ai ​​chip gaudi3 in an attempt to compete with nvidia and amd in the ai ​​market, as well as the development of new ai pc chips such as lunar lake, arrow lake and panther lake , the goal is to have cumulative shipments of more than 100 million units by the end of 2025.

in addition, in september this year, intel announced that it would expand its cooperation with amazon to customize ai chips for amazon's cloud computing department.

intel wants to customize ai chips for amazon｜image source: intel

however, there are still doubts in the market about intel's ability to manufacture chips for external customers like tsmc. citibank analyst christopher danely once said sarcastically: "i'm five feet six inches tall and over fifty. even if politicians all over the world want me to play in the nba, it's almost impossible."

intel's revenue and net profit have declined significantly in recent years. full-year revenue in 2021, 2022, and 2023 is us$79 billion, us$63.1 billion, and us$54.2 billion respectively, while net profit has dropped from three digits to double digits.

in august this year, after its second-quarter financial report still disappointed the market, intel announced the implementation of a comprehensive cost reduction plan, including cutting 15,000 positions, accounting for 15% of the total number of employees, and planned to reduce non-gaap r&d and administrative expenses in 2025. control it at us$17.5 billion and reduce capital expenditure by more than 20%.

the one who is still willing to provide intel with the greatest support now is the us government.

intel is a direct beneficiary of the u.s. chip act, and funds from the policy have provided the company with a temporary buffer, allowing it to continue technological innovation and factory expansion.specifically, it includes up to us$8.5 billion in subsidies, us$11 billion in loans, and up to us$3 billion in funds, etc.。

however, policy financial support alone may not be enough to restore the giant's lost market share. funds from policies are limited and not forever.

as for qualcomm's potential acquisition, it could theoretically create a larger chip giant, but even if it does not encounter antitrust crackdowns from u.s. regulators, there are other challenges. because qualcomm itself does not have manufacturing experience, it may be difficult to swallow intel with its current cash reserves. there are also potential headwinds from geopolitics.

in any case, intel is at a crossroads and its fatigue is showing. how to turn around the situation and even regain its past glory can be said to be a huge pressure for the current ceo.

facebook (now meta) has had a thumbs-up button at the entrance to its workspace in menlo park for many years, and the same location once had the signboard of sun microsystems, which once dominated the the entire silicon valley, until it finally fell into decline and was acquired by oracle.

no matter how powerful a giant is, when it encounters a technological paradigm shift and fails to seize the opportunity, it will be surpassed by smaller but faster opponents. fortunately, at least intel is aware of this.

header image source: intel