Saving Intel: How the Former Chip Giant Is Fighting for Survival

This was clearly not the heroic reception that Pat Gelsinger might have hoped for.In March 2025, Pat Gelsinger was appointed CEO of Intel, only seven months after he had resigned from the company’s board. For some time, Intel had been in turmoil: a sudden retirement of its former CEO, and two senior executives serving as interim leaders.Now, Gelsinger was delivering his first formal address to Intel employees worldwide, online. Employees were hungry for frankness and honesty — and if they felt his remarks lacked sincerity, Intel’s corporate culture encouraged them to say so.“At our company, you can ask anything without worrying about consequences,” a retired employee who worked at Intel for 30 years told Fortune.That day, Gelsinger “was immediately asked: ‘Why did you resign? And are you back now to save Intel?’”He replied that he had some personal matters to take care of at the time, but the audience was unsatisfied.The veteran employee recalled that criticism broke out around him.“People were disappointed,” he said.If that moment had already made Gelsinger sweat, then by August this year the pressure skyrocketed.Arkansas Senator Tom Cotton accused Gelsinger of controlling Chinese companies and holding shares in hundreds of Chinese tech firms.President Donald Trump soon posted on Truth Social, declaring that Gelsinger had “serious conflicts of interest and must resign.”Four days later, Trump met with him.The situation turned around after the meeting.

The two men seemed to find common ground, and shortly thereafter reached an almost unprecedented agreement: Intel would transfer 9.9% of its shares to the U.S. federal government, while the government would provide Intel with $8.9 billion in funding.By late September, NVIDIA — the most valuable company on Earth — agreed to invest $5 billion in the chipmaker.Under the agreement, Intel would manufacture a new series of chips combining technology from both companies, and NVIDIA would purchase some of those chips.Former Intel CEO Craig Barrett told Fortune that this model of customers injecting capital into a supplier might be exactly what a cash-starved Intel needs.“Customers are the only real source of cash,” he said. “They have huge reserves. If eight of them each invested $5 billion, Intel would be saved.”Beyond NVIDIA, such customers might include Apple, Broadcom, Google, Qualcomm, and other firms seeking a second supplier of high-performance chips.At present, only Taiwan’s TSMC can supply chips at this level; it is the largest chip manufacturer in the world.The once-glorious Intel seems to have found a credible path to revival.Depending on execution, this opportunity could become a historic turnaround — or a cautionary case study about mismanagement and government intervention.But either way, its significance goes even further.Saving Intel matters not only to its stakeholders, but to U.S. national security.In the digital age, a nation’s security depends on reliable access to advanced semiconductors.

Yet nearly all cutting-edge chips today are produced by Taiwan’s TSMC, with a smaller portion coming from South Korea’s Samsung.The U.S. Department of Defense secretly purchases vast quantities of such chips because it cannot obtain them domestically.Many might be surprised to learn that even the F-35 fighter jet — the most advanced fighter in the world — depends on TSMC chips.In 2021, leadership at the U.S. National Artificial Intelligence Security Council wrote:“We do not want to exaggerate the danger, but given that most advanced chips come from a single factory, separated from our main strategic competitor by only 110 miles of water, we must reassess the resilience and security of our supply chain.”There are many manufacturers worldwide capable of producing commodity chips used in cars, televisions, refrigerators, and countless other products.As John Neuffer, CEO of the Semiconductor Industry Association, likes to say:If electricity flows through something, it probably has a chip.But the chips Intel produces are far more complex and are used across PCs, workstations, servers, cloud computing, certain AI workloads, and more.Intel is also pushing toward mass production of its advanced “18A” node, and crucially, is attempting to develop an even more advanced “14A” node.Intel operates fabs in Arizona, Oregon, Ireland, and Israel, plus assembly plants worldwide.But only TSMC, Samsung, and Intel can currently produce truly leading-edge chips.These are precisely the chips needed by the U.S. Department of Defense and by hyperscale AI companies.Intel has not actually produced such chips since 2017, and has steadily fallen behind.

Even so, it still has a chance to return.It possesses multiple fabs, several massive facilities under construction, state-of-the-art equipment, and most of the expertise required to compete.It may take years, and mass production of 14A may not arrive until 2027 at the earliest, but compared with the decades it would take any other company to build such capability, Intel remains uniquely positioned.How did America’s once-dominant chipmaker and Silicon Valley pioneer end up here?Because it missed wave after wave of technological change.In the 1990s, Intel dominated the PC chip era.Later, Steve Jobs asked Intel to build chips for the iPhone — then merely a concept.Intel declined.To this day, Intel has never successfully built a graphics processing unit (GPU), a device originally designed for gaming and now central to NVIDIA’s revolutionary AI chips, all made by TSMC.By the 2010s, under several missteps by successive CEOs, Intel entered decline. Revenue peaked in 2021 and has since fallen sharply.Wall Street expects further declines this year.Desperate for leadership, the board brought back Pat Gelsinger, a veteran who had served Intel for 30 years before leaving.“When he came back, people were energized,” a former project manager told Fortune. “We thought: finally, we’re getting back on track.”Gelsinger proposed an ambitious plan: build new fabs and return to leading-edge production.He pushed Congress to pass the CHIPS and Science Act.“I think this is the most important industrial policy legislation since World War II,” he told Fortune in 2024.The law provided subsidies to companies producing chips in the U.S.Foreign firms such as TSMC and Samsung received billions to build U.S.

facilities.Intel, however, received more than anyone.The CHIPS Act is explicitly about national security.As then-Commerce Secretary Gina Raimondo explained:“We cannot rely too heavily on any single region for the most critical hardware of the 21st century.”But the law does not fund any project fully.Some Intel directors worried about Gelsinger’s massive spending — two megafabs in Arizona, another in Ohio, several in Europe.Eventually investors lost faith.Intel’s stock collapsed under his tenure, falling 66% by the time the company announced his retirement in December 2024.When Thomas (Cheng Liwu in Chinese text, here referenced as the successor) left Intel’s board in August 2024, he gave no reasons publicly.But when he became CEO, one of his central tasks was to control the scale of Gelsinger’s projects.“There will be no more blank checks,” he later told employees.Today, no one inside Intel denies that the company needs saving.“To catch up and participate in the AI transformation, Intel has a huge amount of work to do,” said Sachin Katti, Intel’s CTO and head of AI.A Stanford professor who joined in 2022, Katti identified one of his main missions:“Revive the engineering and technical culture that made Intel successful. Intel must regain the vitality it has lost.”But turning a downward spiral upward is not easy.Intel has long attracted the brightest engineers in the world.When the company declines, it loses appeal to the top 1% of talent.Intel is now reframing weakness as opportunity.“Intel is not at its peak. If you join, you are here to restore it to glory. If you love challenge and technology — that’s who I want,” Katti said.Intel faces another pit: its finances.In December 2024, S&P downgraded Intel to BBB, outlook stable.

In August this year, Fitch downgraded it to BBB, outlook negative.Institutional Shareholder Services ranks Intel in the bottom 5% for overall financial quality among 28,000 companies.Bennett Stewart, a corporate finance expert, was blunt about the Gelsinger years:“They tried to spend their way out of trouble. It failed completely.”Under the new CEO, the tone shifted.“We are more financially conservative,” CFO Dave Zinsner said.The company began selling assets to pay down debt.Shortly after taking office, the CEO agreed to sell 51% of Intel’s Altera programmable chip business for $3.5 billion.Intel also sold $1 billion of Mobileye stock.SoftBank plans to purchase $2 billion in new Intel shares, providing more liquidity.Cost reductions followed.The Ohio mega-fab project was delayed.Factory plans in Germany and Poland were cancelled.In July, Intel announced it would cut 15,000 jobs, halving its level-11 management layer.All of this is preparation for returning to advanced chip production.But it is far from enough.Virtually everyone agrees: Intel needs massive external capital.NVIDIA’s investment is a start, but Barrett estimates the company may need around $40 billion.Gelsinger, now a partner at Playground Global, has proposed a U.S. sovereign wealth fund, noting that China subsidizes its strategic tech firms.But Barrett and David Yoffie argue that customers are the most likely source.“$5 billion is not a huge investment for these companies,” Yoffie said. “Spread over years, it’s less than what they already spend on TSMC.”To make participation easier, Intel even reorganized its manufacturing division into a subsidiary.There remains one final question.“Success depends on whether they can execute the entire process,” Bernstein analyst Stacy Rasgon said.“Process” is industry jargon for the vast chain of steps required to produce chips.“If customers cannot believe Intel can deliver reliably, on spec, and at acceptable cost — who would bet their future on Intel?”

Intel must prove it can mass-produce leading-edge chips on time and at scale.“This is the bare minimum,” Rasgon said.Intel also must overcome a damaged reputation.As Rasgon noted:“The past decade has demonstrated that betting against Intel was the correct decision.”CFO Zinsner said some customers believe investing in Intel “might be right,” but they are focused on the 14A technology.Specifically:“What adjustments must Intel make to meet customer needs?”Ultimately, Intel will manufacture chips for customers to test.In the ideal scenario, Intel reaches breakeven on 14A by the end of 2027.Intel believes it can.But it also admits the project might never be economically viable.In its latest SEC filing, Intel stated:If it cannot find even a single major customer and cannot meet key milestones, then “development and production of Intel 14A and subsequent nodes would not be economically viable,” and “we may temporarily or permanently cease work on these nodes.”If that happens, Yoffie warned:“There may never again be a U.S. company producing leading-edge processors in the U.S.”Intel’s future also sits at the heart of national security.William Megginson at the University of Oklahoma studies government ownership in private companies.Asked about U.S. government investment in Intel, he echoed most economists:“It is almost always a bad idea. History shows poor outcomes.”Douglas Redick questioned:“Are we now in an era where the U.S. government picks national champions?”But many foreign policy experts argue that a national champion is exactly what the U.S. needs.Researchers at CSIS wrote:“Supporting Intel is not about protecting a company. It is about ensuring the industrial foundation necessary for national security in a geopolitical competition. Otherwise, America’s future depends on importing advanced chips, which is strategically unacceptable.”As a business, Intel might survive even without 14A.But for the United States, abandoning leading-edge semiconductors would be a dangerous path.