Intel 25 years ago refused to acquire Nvidia, ‘reversely absorbed’ by Qualcomm, ‘decline’ due to government support
Semiconductor chipmaker Nvidia has been the leading stock in the Dow Jones Industrial Average since the 8th, replacing Intel, which had been the leader for 56 years, and has been reversely absorbed by Qualcomm, citing the huge federal subsidy as the cause of its decline.
S&P Dow Jones Indices said in a statement on the 1st that the change will be made “before the start of trading next Friday” “to ensure more representative exposure to the semiconductor industry.”
Nvidia laid the early foundation for the AI revolution, building custom chips for machine learning tasks and building a community of AI programmers who want to develop the technology on the company’s hardware, and leaving Intel, which received billions of dollars in federal funding as a beneficiary of the CHIPS and Science Act, out. Nvidia announced a 10-for-1 stock split in June, making it easier for individual investors to buy the stock without diluting the company's value, and its stock price has continued to rise. Intel, which makes chips that serve as the brains of most computers, brought up the idea of acquiring Nvidia at its board meeting in 2005 as its top earner, but Intel's board, complacent with its technological superiority, rejected it, and now it is a target for a takeover bid from Qualcomm. Chipmaker Qualcomm CEO Cristiano Amon said on Sept. 20 that the company had approached rival Intel about the possibility of acquiring the Silicon Valley giant.
A Qualcomm board member told the Times that the company had not yet made a formal offer to Intel and that the hurdles to a deal remain steep.
The deal, the Times said, “is likely to face significant regulatory scrutiny given the size and national security implications of the two chip companies,” adding that “it is unclear whether regulators would allow Qualcomm to acquire Intel without taking on its struggling foundry business, and it is equally unclear whether Qualcomm would take on that complex business.”
The deal costs Intel, whose stock has fallen nearly 40 percent over the past year, has a market capitalization of $93 billion, while Qualcomm, whose stock has risen 55 percent, has a market capitalization of $169 billion.
Intel’s board of directors On the resistance to the Nvidia acquisition 25 years ago, “Paul Otellini, then Intel’s CEO, presented the incredible idea of acquiring Nvidia, a Silicon Valley startup famous for its chips used in computer graphics, at a price tag of $20 billion to the board, but it was rejected,” a board member testified on the 24th of last month. “Some Intel executives believed that the basic design of the graphics chip could eventually perform important new tasks in the data center, and this was the approach that would eventually dominate AI systems,” the board said. Nvidia now accounts for the majority of AI chip sales and is the world’s second most valuable company, slightly behind Apple, at $3.32 trillion as of the trading hours of the day.
In October, Intel reported its largest quarterly loss in its 56-year history, a “$16.6 billion operating loss in the third quarter,” which included a $15.9 billion loss that reflected a downward valuation of the company’s assets in its accounting and restructuring costs related to the layoff of more than 15,000 employees. That's a combined $2.8 billion.
Intel CEO Pat Gelsinger ordered a restructuring and began laying off employees in August as profits dwindled, citing the cost of catching up with competitors in manufacturing technology and little success in the booming AI chip market.
Intel's stock has plunged 60% since Gelsinger took over as CEO in February 2021, and with a market value of less than $100 billion, it is being mentioned as a candidate for a reverse acquisition or breakup by Nvidia.
The stock value of Nvidia and Intel is already more than 10 times higher. The New York Times, which interviewed more than 20 former Intel executives, board members, and industry analysts about Intel’s downfall, said, “There was clearly mismanagement,” and defined it as “a byproduct of decades of success and high profits dating back to the 1980s when Intel’s chips and Microsoft’s software became the twin engines of the fast-growing personal computer industry.”
The New York Times attributed the failure of evolution to internal management, saying, “That culture was charged and focused on the franchise of personal computers and, later, data centers,” and “Intel executives half-jokingly described themselves as ‘the largest single-celled organism on Earth.’”
Intel’s downfall was a “corporate spirit” that worked against the technology company as it tried and failed repeatedly to become a leader in artificial intelligence chips, and after several years of projects, Intel management lost patience and technology shortages suddenly halted them.
At Intel, investments in new chip designs were always a secondary priority for the company’s costly flagship products. The New York Times diagnosed that “new chip designs were put on the back burner as Intel focused on protecting and extending the chip generation based on the PC era blueprint known as the x86 architecture.”
“That technology was Intel’s crown jewel, proprietary and highly profitable, and they were going to do everything they could to keep it,” James D. Plummer, a former Intel board member and professor of electrical engineering at Stanford University, told the New York Times.
Intel’s executives and leaders occasionally acknowledged this problem before its downfall.
Former Intel CEO Craig Barrett once likened the x86 chip business to a creosote bush. Nevertheless, Intel never really changed its technology or management direction because it had been so profitable for so long. Twenty-five years ago, when Intel considered acquiring Nvidia, the small company was mostly used in niche, specialized chips for computer gamers, but Nvidia later began to apply its chips to other types of numerical calculations, such as oil and gas exploration, and entered the industrial market.
In terms of chip capabilities, while Intel's microprocessor chips excelled at executing calculations quickly, Nvidia's chips divided up the work and distributed it across hundreds or thousands of processors that worked in parallel, delivering superior performance in graphics and, 25 years later, in artificial intelligence.
Intel's board rejected the Nvidia acquisition 25 years ago, and the board focused its efforts on an internal project codenamed Larrabee to stay ahead of its competitors in graphics.
The New York Times reported that Intel's 'Larrabee Project' was "led by Gelsinger, who joined Intel in 1979, and steadily rose through the ranks to become a senior executive," and that "it took four years and hundreds of millions of dollars, but Intel was arrogantly confident that it could change the field, and Intel's leaders sometimes acknowledged this problem."
"Pat Gelsinger is very focused on the manufacturing side," Robert Burgelman, a professor at Stanford Graduate School of Business, told the New York Times. "But they missed A.I., and it's catching up with them now."
Otellini, who proposed to Intel's board of directors to acquire Nvidia 25 years ago and was rejected, died in 2017.
Intel halted the 'Larrabee' project a few months after Gelsinger announced in 2009 that he would move to EMC, a data storage equipment maker, to become president and chief operating officer. Some of the Larrabee technology was used in specialized chips for scientific supercomputing, and Intel’s graphics push was scaled back.
Nvidia continued to invest in chip design as well as critical software for years to allow programmers to write broader software applications on the hardware.
Intel continued to stumble in the AI market in the years that followed, and in 2016 it paid $400 million for Nervana Systems, one of the newer companies in AI chips.
Intel CEO Naveen Rao was named head of Intel’s new AI products division.
Rao’s engineering team was able to introduce two new chips, “one of which was used by Facebook,” Rao told the Times. Intel's rift erupted in December 2019, when Intel's board acquired another AI chip startup, Habana Labs, for $2 billion over Rao's objections. The Times reported that “the deal came just before Rao’s team was ready to finalize a new chip,” and that Rao, who resigned shortly after the acquisition and is now vice president of AI at software company Databricks, had “a product ready to ship, and we shot it and then we bought Labs for $2 billion.”
Intel has lagged behind in its efforts at AI, having developed several now-discontinued graphics-style chips and taken years to deliver a reliable chip from its Habana Labs line. Intel’s latest version, Gaudi 3, is attracting interest from some companies, including startup Inflection AI, as a low-cost alternative to Nvidia.
Intel, long a leader in chipmaking technology, has been given billions of dollars in federal funding under the CHIPS and Science Act, a strategy that has allowed the Biden administration to catch up with Asian rivals from South Korea and China as part of a “make chips in America” policy, opening the way for Intel to finally recover.
The Times reported that “Intel recently designed new chips, including an AI chip for notebook PCs, with federal funding, but the fact that the new chips are being manufactured not at Intel factories but at TWMC (Taiwan Semiconductor Manufacturing Company) in Taiwan is a measure of Intel’s problems,” the report said. “The decision to exploit the company’s more advanced production technology is a measure of Intel’s interests in chips.”
In March 2024, the Biden administration announced that it would provide Intel with $8.5 billion in grants and up to $11 billion in loans to build the plant, and Intel would be eligible for $25 billion in tax credits from the government.
Intel was also selected as the sole partner for a $3 billion military chip manufacturing program.
Around this time, Intel filed paperwork in Ohio indicating that it would delay construction of the plant by two to three years.
Just before this, Commerce Secretary Raymond had urged executives from Apple, Amazon, Nvidia, AMD, Marvell Technologies, and eight other tech companies to consider ordering chips from Intel’s U.S. manufacturing plants, but they declined, citing a “lack of sophisticated technology.”
The Times reported that eight people familiar with the request said Intel’s chipmaking technology was not as sophisticated as that of Taiwan Semiconductor Manufacturing Company (TSMC), the world’s leading chipmaker and major supplier. Intel's request to delay construction of its factories led to a failure to secure customers, and on August 1, Intel reported dismal quarterly results, with profits falling and the company suspending its decades-old shareholder dividend, sending its stock price plunging 26%, its worst day in 56 years.
SK Chairman Choi Tae-won said in his keynote speech at the 'SK AI Summit 2024' on the 4th, "Jensen Huang, CEO of Nvidia, emphasizes 'quickly, quickly' as if he is Korean. When I met him recently, he asked me to bring forward the HBM4 supply schedule by six months. When I asked Kwak No-jung, CEO of SK Hynix, if it would be possible, he said, 'I'll give it a try. '" He officially announced the commercialization of the 4th generation HBM4 by the 'command economy system' in the first half of next year.
Samsung Electronics is a representative 'real estate loan-dominated manufacturing company' that has been unusually packaged as a 'national stock with no stock price decline' for technology stocks, and has provided a 'hunting ground for ants' to foreigners in the '100,000 electronics' packaging strategy guaranteed by the Korean government and securities companies, and is following in the footsteps of Intel's decline in technology companies due to its dependence on government support.
SK Hynix accelerated the ‘stock price increase speed war’ by ‘command economy’ in the HBM competition system that requires infinite investment costs, and also got involved in Intel’s follow-up procedures.
SK Hynix's HBM yield is at the level of 48%, and it is expected to be a hunting ground for ants by securities companies and the media, whose approach to yields depends on unilateral announcements from management.
,