would it be better to change the ceo?

arts｜"chinese entrepreneur" reporter yan junwen

edit|yao yun

image source｜visual china

while the outside world is debating rumors about qualcomm launching a tender offer for intel, intel ceo pat gelsinger is sitting firmly on the diaoyutai. in the early morning of september 23, he posted on social platforms quoting the gospel of mark and said:

“you shall love the lord and god with all your heart, soul, mind, and strength.”

some netizens left a message, "god can't save intel, but you can."

according to a report on the us "wall street journal" website on september 20, according to people familiar with the matter, chip giant qualcomm has recently been in contact with rival intel corporation regarding acquisitions. this move is expected to lead to one of the largest and most far-reaching mergers and acquisitions in recent years. .

in kissinger’s social platform updates, there are a lot of remarks about intel, such as receiving a us$3 billion subsidy from the federal government and releasing the most powerful x86 server, but most of them are sharing about beliefs. he mentioned in one of his speeches that he worked for faith, not for humanity.

kissinger was a devout christian who was born into a farm family in pennsylvania. pennsylvania is located in the northeastern united states and is the center of american manufacturing. it was a relatively traditional place, and for a time the most likely way out for kissinger was to become a farmer rather than an engineer.

now, this old man who was born in 1961 and is almost the same age as intel (founded in 1968) is watching this giant ship disintegrate and sink slowly.

gelsinger also experienced intel's most glorious period. in 1971, intel launched the 4004, the first general-purpose chip in human history - the world's first commercial computer microprocessor. as the advertisement at the time said, it was "an epoch-making work". it can be said that the computing revolution it brought about changed the entire world. at the same time, the "moore's law" proposed by one of its founders, gordon moore, affected the entire semiconductor industry.

since the invention of the x86 architecture, intel has relied on its design and manufacturing advantages to become the chip company with the most advanced design and the largest production volume in the world. in 2013, more than 85% of the global cpu market was occupied by intel.

in 1979, kissinger joined intel and worked directly under its three founders, robert noyce, gordon moore and andy grove. in 2001, he became intel's first cto, but left after failing to seek the position of chairman, and then served as ceo of companies including vmware.

later, when the intel board of directors found kissinger, who had been away for many years, he was immediately surrounded by a heavy sense of history and emphasized the mission, "for the legacy of gordon, andy and robert," "i wrote a letter for the board of directors strategy document. one day, it might be in the intel museum."

in february 2021, kissinger decided to resign as ceo of vmware. at that time, he was thriving at the enterprise service software giant. he was approaching his tenth year of employment. the company's annual revenue had more than tripled, and he was named "the best ceo in america" ​​for this. prior to this, kissinger expressed several times that he would not return to intel.

but in the end, gelsinger decided to accept the invitation from the intel board of directors and take over as intel ceo. in this regard, public opinion and the capital market gave kissinger the greatest goodwill. the outside world said that the "technical faction" returned to intel. on that day, intel's stock price rose by more than 12%.

at that time, intel was like an aircraft carrier sailing in the vast ocean, but this aircraft carrier was dilapidated and underpowered, but no one dared to innovate, because if they were not careful, the "symbol of american chips" would fall apart, and no one dared to undertake such a of infamy.

it is normal for enterprises to face crises and risks during their development, but this time intel was indeed on the verge of losing control.

the foundry's chip manufacturing process lags behind that of tsmc and samsung, but this is the most exciting "made in america" ​​story that intel has been telling recent u.s. presidents; the basic cpu market in the past has continued to beAMDcannibalization; it has not caught up with the gpu craze brought about by ai and large models, which is what the market is most concerned about now.nvidia, all customers are waiting for the goods, and there is even a slogan that gpu is more precious than gold, while intel's latest gpu is still in the laboratory.

the second quarter financial report for 2024 released in august this year showed that intel’s revenue was us$12.8 billion, a year-on-year decrease of 1%, and its net loss attributable to the parent company was us$1.6 billion, a year-on-year decrease of 200%. intel ceo gelsinger said the company's performance was disappointing. within one day, the stock price fell by 26%, leaving a market value of just over us$90 billion.

despite this, gelsinger is optimistic about intel's chip production advantages in the context of new demands triggered by ai pcs. he feels that intel can still reap rewards in the future as long as it bets on advanced technology.

according to convention, intel's ceo will retire at the age of 65. this year, kissinger, who was born in 1961, is 63 years old, and there is not much time left for him to save intel.

would it be better to change the ceo?

in early august, after intel released its disappointing second-quarter financial report, during a tv show, ed snyder, an analyst at the market research company charter, made a surprising statement. he commented:

"intel should abandon current ceo pat gelsinger and seek a reformist ceo from outside, otherwise the company will be in trouble."

however, schneider's remarks were countered by some internet users. they left messages saying that schneider's statement was nonsense, unfounded and unrealistic. no one except kissinger had the ability and willingness to take over intel. if you don't pay attention, this veteran chip giant will reach a dead end.

indeed, what intel needs most now is not to catch up, but to maintain, and then carefully perform surgery to remove the lesions.

in february 2021, kissinger, who had just become the helmsman of intel, issued an internal letter laying out the intel 2.0 plan. at that time, kissinger used a lot of nice words to praise intel. he said: no semiconductor company in the world can combine the advantages of design, manufacturing and large-scale work like intel. in the future, intel will become a leading player in a huge change. the world's leading semiconductor company.

to be fair, gelsinger is already the most reform-minded ceo at intel apart from the founding team.

on september 16, u.s. time, kissinger issued an internal letter and announced changes:

first, as intel 18a (equivalent to the 1.8nm process) is about to be launched, we must build on the momentum in the fab and improve the capital efficiency of this part of the business.suspended chip factory project in poland. in addition, intel will establish an operating committee, including independent directors, to manage the subsidiary and continue to focus on improving transparency, optimization and accountability across the business.

second, continue to take urgent action to build a more competitive cost structure and achieve the $10 billion in savings we announced last month.the measure is to lay off 15,000 people. intel has approximately 125,000 employees worldwide. they also plan to reduce or withdraw from about two-thirds of global real estate and sell equity in subsidiaries to make them public.

third, as we advance our ai strategy, we must refocus on our strong x86 franchise while streamlining our product portfolio to serve intel customers and partners.

this reform measure looks familiar. simply put, it means reducing costs and increasing efficiency while betting on the future.

in the 1980s and 1990s, inqualcommwhen chip companies such as nvidia and nvidia abandoned "wafer fabs" and adopted the "offshore production" model to define the design and manufacturing of chips, intel still insisted on this cumbersome approach. at the time, it was the dominant player in servers and pcs, and it had cost advantages to design and produce its own products, but as it declined, the fabs became a heavy burden.

this is reflected in the latest financial report for the second quarter of 2024: intel's second quarter net profit turned from profit to loss, and the foundry lost a significant us$1.6 billion, offsetting its original net profit of us$1.5 billion.

according to a document intel previously submitted to the u.s. securities and exchange commission, intel's foundry lost as much as $7 billion last year. the company expects the business to "achieve operating breakeven by the end of 2030," and gelsinger hopes that this date will be in 2027. year.

but the main unit volume of intel's chip foundry comes from intel itself. some external market customers are not confident about giving the most advanced design chips to intel's foundries. after all, the products of these external customers are in direct competition with intel.

but intel will not directly abandon the foundry, because this is the most important focus of the recent "chip and science act" implemented by the united states, and it is also the company with the most subsidies in the united states. it represents the success of the chip industry's "american production, american manufacturing" . it is understood that the latest subsidy is us$3 billion, which is not related to the us$8.5 billion that intel may receive under the chip and science act to support its commercial factories in four states in the united states.

in other words, the united states directly subsidizes intel's foundry production, and the cumulative amount may exceed us$11 billion, but building a wafer fab - purchasing euv equipment, training engineers, etc., may be a loss-making deal.

multiple ceos, each laying out minefields

when a company encounters problems, replacing the ceo may be the easiest thing to do.

but more and more company cases show that simply replacing the ceo without reforming the company's culture and team spirit may not be a wise move. whenever intel lags behind in a certain business, the outside world will think that it is time to change the ceo - but often, every ceo will lay the groundwork for intel.

the foreshadowing was laid in the 1990s.

before kissinger, intel had seven ceos, three of whom were founders. one of the founders, andy grove, is known for his "disruptive innovation".

in the 1990s, as intel ceo, he seriously considered converting intel's main chips from x86 to risc architecture because x86 was complex, large, and energy-intensive. but he ultimately decided not to do so. the reason is that although risc is more efficient, the conversion cost is very high: intel has a monopoly on x86 and has the largest profit. if you switch to risc, you have to give up the previous profits and technical barriers, and you need to start from scratch.

for a long time, the computer industry was designed around the x86 architecture, and intel arguably dominated the entire ecosystem.the x86 architecture dominated personal computers not because it was the best, but because ibm's first personal computer used it.

paul otellini served as intel ceo from 2005 to 2013. during this period, intel consolidated its dominance in the industry, but he also missed out on the mobile internet because he refused to produce chips for apple's mobile phones. however, he secured an order for apple's mac notebook computer, which made up for some of the shortcomings.

brad krzanich served as intel's ceo from 2013 to 2018. during this period, tsmc has advanced to an advanced process almost every two years: 16nm in 2013, and then it has been continuously "upgraded" to advanced process wafers such as 10nm and 7nm. in 2020, tsmc began to develop 5-nanometer process wafers. however, after intel's foundry broke through the 14nm node in 2014, it stopped moving forward and called it "14nm+, 14nm++" to the outside world, giving it the nickname "toothpaste factory". since then, intel has created its own chip manufacturing system.

bob swan served as intel's ceo from 2018 to 2021. he chose to give up the research and development of 5g baseband chips, which caused intel to perform poorly in the mobile phone industry. in addition, in june 2020, during his tenure, apple officially announced that it would abandon intel processors and switch to arm-based chips. five months later, apple officially released the m1 chip and announced that it would complete the replacement of all x86 architecture chips by 2023, ending the cooperation between the two parties since 2005.

the elephant's turns should be uniform and coordinated.

gerstner, the legendary ceo of ibm, summarized in "who says elephants can't dance":big is important because scale is leverage, and depth and breadth can accommodate larger investments, greater risks, and longer-term investments in the future.

gerstner said that this is not a question of whether the elephant beats the ant. it is a question of whether an elephant can dance. if the elephant can dance, then the ant must leave the stage.

in early september, founders in silicon valley started a discussion about the "founder model" after brian chesky, co-founder and ceo of the homestay giant airbnb, complained in a lecture.

as airbnb expanded, brian received a lot of advice along the lines of “hire good people and give them the space to get the job done.” but the results were unsatisfactory. "the less hands-on i am, the more involved i am in a problem," brian said. "when i get involved in an issue, it's like 10 times more work."

paul graham, the godfather of silicon valley investment and co-founder of y combinator, proposed the concept of "founder model". he said that compared to professional managers who only control the company through subordinates, founders who are hands-on are more conducive to the development of the company.

the article quickly spread across silicon valley. tesla founder musk retweeted this article, saying it was "worth reading." tobias lütke, founder of the e-commerce giant shopify, posted that he had the same experience and called for "companies in all industries to adopt the founder model." as a result, discussions between the founder model and the professional manager model have started in silicon valley.

this path is not for intel: founder robert noyce died on june 3, 1990, andy grove died on march 21, 2016, and gordon moore will die in march 2023 died on the 24th.

but the core of the "founder model" is not the return of the founder's physical body, but the return of the company's sense of purpose, acumen to the market, and the control of the helm's voice.this is not easy for the 56-year-old intel.

“in today’s highly competitive and ever-changing world, few, if any, large companies can pursue a complete corporate spin-off strategy. because relying on a corporate spin-off strategy to promote the necessary major changes in the company comes at a cost too expensive and too slow," gerstner concluded.

from the perspective of gerstner's remarks, it seems reasonable for intel to be cautious about completely separating its foundry and other businesses. the company needs to be tempered rather than given strong medicine.

references:

"chip wars: the battle for the world's most critical technology", zhejiang people's publishing house, chris miller

news hotline & submission email: [email protected]