The fall of Intel is a cautionary tale for Silicon Valley

Even in a world where Elon Musk announces company decisions on Twitter before telling his employees, it’s not every day that the CEO of a Fortune 500 company starts tweeting bible verses like this:

Granted, Intel CEO Pat Gelsinger has always been a proudly religious man, but his recent choice of scripture carries a tinge of extra weight given the fact that the company had to slash more than 15,000 jobs this month, or 13% of its workforce, while halting its dividend payments for the first time in 20 years. 

Viewed through that lens, this verse suddenly feels a lot less soothing: 

Now it reads like a rallying cry both for him and the employees who remain. 

For more than three decades, Intel was the worldwide leader in both the design and manufacturing of computer chips. Its rise between the 1970s and 2000s coincided with Silicon Valley bursting into mainstream consciousness.

The South Bay, where Intel is based — an unassuming region about an hour south of San Francisco — was responsible for pumping out the technology that powered the world’s personal computing revolution. As recently as June, Intel reported about 116,500 employees. 

But the good times clearly didn’t last forever. In recent years, one of Silicon Valley’s Mount Rushmore companies found itself in a precarious position that the current crop of tech billionaires are maniacally trying to avoid: getting left behind.

With demand tilting first to smartphones and now to artificial intelligence — two tech waves that Intel missed out on while at the top of the industry — the company is struggling to find a business line in which it can still succeed, let alone win. 

Now, in an effort to survive, it’s putting its chips (ha!) on one race where it’s already falling behind: building other companies’ designs, not just its own. 

Cutting labor costs apparently wasn’t enough. This week, in an effort to boost its cash reserves, Intel also sold off its 1.18 million-share stake in British chip company Arm Holdings, whose chips are used in virtually all modern smartphones. 

“Technologists know decline inevitably comes for all,” said Vivek Agarwal, a former Intel employee, who worked in both engineering and product management over his 30 years at the company. “So, they try to introduce new innovations as their current products are maturing with the hope that they can capitalize on the next hit.”

“It’s not like Intel didn’t have the technology either,” he said of AI and smartphones. “But they missed the boat from a timing and marketing perspective.” 

Intel was founded by two of Fairchild Semiconductor’s chief scientists, Gordon E. Moore (known for Moore’s Law) and Robert Noyce, in 1968. The company went on to create the first commercially viable microprocessor three years later, which essentially launched the personal computer industry. 

During the 1990s, Intel’s partnership with Microsoft garnered it the nickname “Wintel.” Whenever orders swelled for the latter’s new generation of Windows products, so did Intel’s. 

Its capabilities as a company that can both design and manufacture its own chips is a rarity even today. 

This virtuous cycle tracks with the rise of Nvidia, the current dominant AI chipmaker, Agarwal explained, which has seen its market valuation jump to more than $3 trillion, since its graphic-processing chips have not yet been matched in terms of production speed and capacity. 

Unlike Intel, however, Nvidia doesn’t manufacture its own chips.   

“Intel became the market leader in microprocessors,” said Agarwal, now a professor at San Jose State University’s Graduate School of Business. “But that came with some unintended downsides.” 

Chief among them was a misreading of where the technological headwinds were heading. “Intel assumed that the PC market would keep growing,” Agarwal said. Even though it had the capabilities to produce products for emerging smartphone and AI industries, it was unable to penetrate either market and eventually ceded ground to better-positioned competitors. 

For example, in 2006, Intel, Microsoft and Samsung partnered to launch the “Ultra Mobile PC” but marketed the tablet-like device as a miniature laptop rather than a cellphone. The iPhone was released only a year later. 

Seven years ago, Intel also reportedly had the chance to invest in OpenAI while it was still a little-known nonprofit, but Gelsinger’s predecessor, Bob Swan, decided against the deal since he did not think generative AI models would make it to market anytime soon.

“Putting down the right bets while staying persistent enough to see them through is a delicate dance,” Agarwal said, adding that he believed Intel’s decision making slowed. “You can only sell so many PCs,” he said. “Intel hit a wall.”

Meanwhile, Intel also began losing ground to the Taiwan Semiconductor Manufacturing Company, which leapfrogged it as the world leader in chip manufacturing in 2017 because it was able to make a greater variety of chips on spec for clients including Apple and Nvidia, while advancing its own computer chip innovations (Intel’s specialty) faster. 

All of this eventually came to head this year as profit margins kept shrinking while employee counts grew, said current Intel CEO Pat Gelsinger in a note to employees announcing the layoffs. 

“Clearly market conditions, some were good and some not so good, and you have to adjust the financial envelope appropriately,” he told The Wall Street Journal. “The AI surge was much more acute than I expected, and you have to adjust to those things.” 

Since sales of its own products were flattening, Gelsinger has doubled down on Intel’s factories around the world, with an eye toward transforming the company into a premier “foundry” for all types of chip designs, not just its own. 

Intel remains the biggest fish in the domestic pond in chip manufacturing facilities. According to market research firm TechInsights, the company’s current fabrication plants account for about 41% of the country’s 300 mm wafer production capacity — the type of chip production most commonly used around the world that involves growing and thinly slicing crystalline silicon. 

Despite having everything under one roof, Intel has a lot of ground to make up in terms of manufacturing other people’s designs, writes Dylan Patel, founder of semiconductor research group SemiAnalysis. 

“Nvidia, [Advanced Micro Devices], Broadcom, TSMC, and the hyperscalers are all beating Intel at each of their core competencies,” Patel said. “Beating or partnering with these firms is not going to be easy.” 

Gelsinger said as much in an interview with The Journal last year when discussing Intel’s pivot in strategy. “Foundry is a service business,” he said. “That isn’t the culture that Intel’s had.”

In the manufacturing-as-service space, Intel is still behind TSMC and Samsung. Gelsinger’s turnaround strategy at Intel revolves around closing the gap between those two by the end of the decade. After falling behind Taiwan over the past decade, Gelsinger also pledged that Intel will achieve five advances in its own chip technology within the next four years. 

Working in Intel’s favor is the U.S. federal government’s recent commitment to invest in domestic chip production amid growing tensions between China and Taiwan. In 2022, the White House signed the CHIPS Act into law, which outlined $39 billion in grants funneled to American chipmakers to help with the cost of building new facilities.

Intel was the greatest beneficiary of the legislation, landing more than $8.5 billion, which it used to help fund new factories in Arizona and Ohio. 

“There is still trust in the Intel brand,” said Bob Staedler, who has observed various tech booms in the region as principal of land-use consultancy Silicon Valley Synergy. “It’s like Volkswagen. It might not be top of mind anymore, but it will get the job done.” 

Even so, the new era at Intel is not off to a good start. A pivotal deal for SoftBank to partner with Intel to produce AI chips to compete with Nvidia reportedly collapsed after the chipmaker struggled to meet the Japanese conglomerate’s requirements.

When reached for comment, an Intel spokesperson pointed to an Aug. 6 press release announcing that the foundry business was on track to start production on AI and CPU products for select clients in 2025.  

“On paper, the new strategy makes sense since Intel is known as a [fabrication] house,” Agarwal said. “But what’s the guarantee that it can deliver on its promises?”