How Intel Fell from World Chip Leader to Takeover Target

Now Intel itself has become a takeover target, a signal that strategic missteps and the rise of artificial intelligence have affected the fate of America’s largest semiconductor company.

Qualcomm’s recent acquisition approach, reported Friday by The Wall Street Journal, reflects a vulnerability that is unmatched in Intel’s 56-year history. The problems began with manufacturing setbacks before Gelsinger took over. And they got worse as the CEO pursued a costly restructuring strategy that failed to anticipate how the explosion of interest in artificial intelligence would fundamentally shift demand toward the type of chips made by rival Nvidia.

“The last two or three years, the transition to AI was really the final nail in the coffin for them,” said Angelo Zino, a veteran industry analyst at CFRA Research. “They just didn’t have the capabilities.”

Even if Intel proves receptive, a deal with Qualcomm is far from certain for regulatory and other reasons. But the idea of ​​the smartphone giant acquiring Intel would have been almost unheard of not long ago.

Intel reigned for decades as the world’s most valuable semiconductor company, and its chips were nearly ubiquitous in personal computers and servers. In an industry where specialization was increasingly the norm, it was rare for a company to design and manufacture its own chips—and be a world leader in both.

When Gelsinger took over as CEO in early 2021, Intel had lost some of its potential as it fell behind Asian rivals in the race to produce the fastest chips with the smallest transistors.

Gelsinger, who had spent decades at Intel and served as its first chief technology officer, had a plan to restore the confidence the company had enjoyed under CEOs like Andy Grove and Paul Otellini.

The move would have caught up with Asian rivals Taiwan Semiconductor Manufacturing Co. and Samsung Electronics. He also planned to spend big money to expand Intel’s manufacturing operations and sell that capacity to pure-play chip design companies like Qualcomm — pushing into the so-called foundry business, where TSMC and Samsung dominate.

It was a costly and ambitious bet, but it seemed to have all the ingredients to succeed: a strong core business of making chips for personal computers and servers, plus a range of side businesses that could help fund the next stage of Intel’s growth.

Gelsinger quickly sought to use Intel’s financial resources to build out its contract chipmaking business, engaging in talks to buy GlobalFoundries for about $30 billion the summer after he took over. That deal fell through, but in an August 2021 interview, the CEO said Intel was still focused on acquisitions. “The industry is going to consolidate,” he told the Journal. “That trend is going to continue, and I expect we’ll be a consolidator.”

It eventually decided to acquire Tower Semiconductor, another chipmaker, for more than $5 billion, though that deal was canceled last year because Chinese regulators did not approve it.

Intel’s contract manufacturing business had a slow start toward Gelsinger’s goal of becoming the world’s second-largest by 2030. The company went through a cycle of several leaders and numerous potential customers that scaled back or exited after encountering technical bugs.

Nvidia’s growth

As costs rose at Intel, generative AI took off. The boom shifted demand from Intel’s central processing units toward Nvidia’s “graphics processing units,” whose different design is better suited to building and deploying the most advanced AI systems. While tech companies were begging for Nvidia’s rare AI chips, many Intel processors sat on shelves.

Gelsinger has been forced to cut costs to keep his restructuring efforts afloat. Intel has laid off thousands of people starting in 2022 and cut its dividend last year. That wasn’t enough. Last month, Gelsinger said he would cut 15,000 jobs, cut costs by $10 billion next year and eliminate the dividend.

“The AI ​​surge has been much more severe than I expected,” Gelsinger said at the time, calling the cuts “the hardest thing I’ve done in my career.”

Intel last week announced new moves that include tighter spending controls and further separating its design and manufacturing divisions. Gelsinger, however, has stopped short of selling or spinning off the manufacturing division, as some investors have urged.

“We have to fight for every inch and execute better than ever before,” Gelsinger told employees. “Because that’s the only way to silence our critics and deliver the results we know we can achieve.”

Analysts say the prospects for a positive turnaround for Intel are dwindling but still possible. Cost-cutting could help the company weather the difficulties, although a falling stock price has made it more vulnerable to takeover bids and investor activism.

By Thursday’s close, Intel shares were down nearly 70% from their early 2020 highs, when they hit their highest since the dot-com collapse. Nvidia shares have surged more than 18-fold in that time.

Intel shares ended Friday’s session up 3.3% after the Journal published a report on Qualcomm’s interest.

“It may be too late”

Stacy Rasgon, an analyst at Bernstein Research, said Intel’s future hinges on the success or failure of its next-generation chip technology, which is set to go into production next year, and that Intel hopes to outperform its rivals, at least technologically. A return to technological leadership could help improve profit margins and inspire customer confidence.

But Intel faces a fundamental problem that won’t go away: It doesn’t expect its core chip business to recover anytime soon because of continued high spending on AI chips.

“We can argue whether the strategy is right or wrong, but the problem is the core business doesn’t support that path,” Rasgon said. At that point, though, “it might be too late for them to stop.”

For Qualcomm, buying Intel could help it jump into new segments of the chip industry. Qualcomm specializes in chips for mobile phones — it supplies Apple’s iPhone, among other devices — and has been building a portfolio of automotive and internet-of-things chips for several years. Intel would add a large franchise of chips for personal computers and servers.

It’s unclear, however, whether Qualcomm would retain Intel’s manufacturing operations in any deal. Those operations make Intel’s business very different from Qualcomm’s, which outsources manufacturing. Manufacturing is incredibly complex and expensive. Intel poured $25.8 billion into capital expenditures last year, about 48% of its revenue. Qualcomm’s capital expenditures totaled $1.5 billion in its most recent fiscal year, just over 4% of its sales.

Write to Asa Fitch at [email protected]