Source: Content from nextplatform, etc.

Intel Chief Executive Officer Pat Gelsinger was forced to step down after less than four years at the helm, handing control to two deputies as the faltering U.S. chipmaking giant searches for a permanent successor.

Gelsinger resigned on Dec. 1, according to people familiar with the matter. He resigned at a board meeting last week, where directors decided that Gelsinger's expensive, ambitious plan to turn around Intel wasn't working and that changes weren't happening fast enough. The board told Gelsinger he could choose to retire or be removed, and he chose to resign, according to sources.

His departure came before he had completed a four-year roadmap to restore the company's lead in making the fastest and smallest computer chips, a title that had been taken away by Taiwan Semiconductor Manufacturing Co., which makes chips for Intel rivals such as Nvidia.

Under Gelsinger's leadership, Intel - founded in 1968 and for decades the cornerstone of Silicon Valley's global chip dominance -- has shrunk to a market value more than 30 times smaller than that of Nvidia, the leader in artificial intelligence chips.

In 2021, Gelsinger inherited a company fraught with challenges that he made worse. Intel had big ambitions for manufacturing and artificial intelligence capabilities among major customers, but under his leadership, Intel ultimately lost or canceled contracts and failed to deliver promised products, according to an October Reuters special report. He made optimistic forecasts for future AI chip deals that exceeded Intel's own expectations, leading the company to withdraw its recent revenue forecast about a month ago.

Bloomberg previously reported on Gelsinger's retirement.

Gelsinger, 63, assured investors and U.S. officials who are funding Intel's transformation that his manufacturing plans remain on track. But the full results won't be known until next year, when the company plans to put its flagship laptop chips back into its own factories.

The company's shares fell 0.5%. The company's shares have lost more than half their value this year, and last month Nvidia overtook it as a blue-chip Dow Jones Industrial Average stock. Rival Advanced Micro Devices rose 3.6% and the Philadelphia Semiconductor Index rose 2.6%.

The company then named Chief Financial Officer David Zinsner and senior executive Michelle Johnston Holthaus as interim co-CEOs while the board began searching for a new CEO. Less than a week ago, U.S. officials had just given Intel $7.86 billion in subsidies.

The board has set up a committee to search for Gelsinger's successor.

"While we have made significant progress in restoring manufacturing competitiveness and building our capabilities to become a world-class foundry, we know the company has much more work to do and we are committed to restoring investor confidence," Frank Yeary, the independent chairman of the board, said in a press release.

Intel communications chief Karen Kahn also plans to leave, according to two people familiar with the matter. Gelsinger announced his turnaround plan in July 2021, when the company was already reeling from years of manufacturing missteps, and then began investing heavily. The company began building a new $20 billion plant in Ohio and hired 132,000 employees, more than Intel had when it was the largest player in the chip industry.

But those expenditures coincided with the post-pandemic collapse of the laptop and PC market, which left Intel's gross margins well below historical norms and depressed its stock price, sparking acquisition interest in the company.

The expenditures ultimately forced Gelsinger to propose a series of layoffs and potential asset sales and spinoffs.

"The stock is down more than 60% during his tenure, so this shouldn't be a huge surprise," said Ryan Detrick, chief market strategist at investment advisory firm Carson Group. "We need new leadership to turn things around, and it's safe to say that any of his major strategic decisions are in danger of being overruled, including the move to focus on becoming a contract manufacturer."

Gelsinger has also failed to launch a working AI chip to challenge Nvidia, which is starting to move toward a $3 trillion company by supporting services such as ChatGPT.

"At the end of the day, you need cutting-edge products, innovation and execution, and that's all we haven't seen under Pat Gelsinger," said Hans Mosesmann, an analyst at Rosenblatt Securities.

The centerpiece of Gelsinger’s transformation plan is for Intel to become a major player in contract manufacturing for other companies, a business model known in the chip industry as “foundry.” Intel has announced some foundry customers, such as Microsoft and Amazon, but none of them can supply Intel factories with the large volumes of chips needed to ensure profitability.

The huge spending and lack of substantial progress in the company's foundry business led to tensions on the board, and Lip-Bu Tan, who had led the chip industry into trouble, left because he disagreed with Kissinger's strategy.

Kissinger leaves, then what?

Intel's prodigal son, now former CEO, was one of the few executives in the semiconductor industry to be groomed by company co-founders Gordon Moore, Robert Noyce, and Andy Grove. Gelsinger's departure from Intel does not bode well for the company in any way. Gelsinger returned to Intel as CEO at a time when Intel was facing its worst crisis since leaving the DRAM memory business in the late 1980s.

This crisis could be worse, which is why Gelsinger left VMware to take the helm of Intel in January 2021. At the time, in What Gelsinger Could Do to Turn Intel Around, we outlined the situation in 1987 when Grove became CEO, Intel’s exit from the DRAM business, and Gelsinger took the helm of Intel. We won't repeat all those ideas here.

But what we have to do is compare it to IBM's self-proclaimed "near-death experience" in the late 1980s and early 1990s, when the mainframe market collapsed and RISC/Unix systems rose in the data center, and client/server computing moved a lot of application work from data center systems to the desktop. John Akers, a mainframe expert and Big Blue insider at the time, didn't read the market forces correctly.

The comparison between IBM and Intel or Akers and Gelsinger is not pretty.

In the late 1980s, Big Blue was at its peak, with more than 400,000 employees, half of them in the United States. We were entering the data center market ourselves, and it was an exciting time, with maybe 25 different chip architectures in systems and maybe 40 operating systems with some market penetration. But inevitably, consolidation was coming, and IBM was the biggest and richest target for competitors.

In 1991, IBM laid off about 30,000 people and took a $3.4 billion charge, with a $2.8 billion loss for the year. In 1992, IBM's revenues shrank slightly, with a loss of nearly $5 billion, and the number of employees decreased by nearly 43,000. This brought IBM's global workforce to just over 300,000, a reduction of nearly a quarter in two years. In 1993, revenues fell a few percentage points again, and IBM laid off another 45,000 people, incurred restructuring charges of $8.9 billion, and lost $8.1 billion on sales of $62.7 billion. Akers' fate was sealed, and he "retired" in April of that year, and the board invited American Express executive Louis Gerstner to replace him. In his first year in office (1994), IBM's operating income increased by two points to $64 billion, and it did achieve a profit of $3 billion, but Gerstner continued to cut jobs, and the global headcount decreased by 36,000 to 220,000.

During this time, there were persistent rumors that IBM would break up into "Baby Blues," spinning off the PC, mainframe, X86 servers, other proprietary servers (AS/400), Unix servers, middleware and database software, printers, and services businesses into separate companies and letting them operate freely in the market. Gerstner decided - correctly, in our view - that it would be wiser to keep IBM together, pivot heavily toward services, and let its software run on other systems than to break up the company. IBM acquired Lotus’ middleware, sold low-end printers, and corrected its financials with massive stock buybacks that paid off in EPS growth for IBMers and Wall Street. (We call that burning the furniture in the house to keep warm.)

We can debate whether this strategy worked. But while IBM subsequently sold off DRAM memory, disk drives, PCs, high-end printers, X86 servers, and various services businesses, many of the businesses that were there are still there today. The mainframe and Power server businesses are relatively healthy and profitable, and IBM owns Red Hat and the soon-to-be-acquired HashiCorp. This is one possible equilibrium that Big Blue could engineer.

Future of Intel

The question we face (and the question facing Intel's board) is, can we say the same thing about Intel in ten or twenty years?

This brings us back to Gelsinger, who is as definitely Intel Inside as Akers is True Blue. So far, he has only laid off 15% of Intel's workforce, and the transformation process has lasted as long as Akers tried to do, ultimately resulting in a nearly 50% reduction in IBM's workforce.

If Gelsinger is unable to fulfill his CEO fiduciary responsibilities due to health issues, then Intel must make that clear in the announcement of Gelsinger's "retirement" because it is so relevant to the company's current state and the strategy that Gelsinger laid out several years ago. In our opinion, Gelsinger is not tired, let alone retiring, which means we can only speculate that Gelsinger's removal is a repudiation of Intel's IDM 2.0 strategy, in which Intel will revitalize its struggling foundry business while revitalizing the chip design business it relies on.

Perhaps Gelsinger was unable to make the dramatic layoffs and strategic changes that Akers had to make to get Intel back to financial health as quickly as possible. Both men cared too much about the companies they ran, and they had to bear the brunt of this shame personally. Running companies like IBM and Intel is extremely difficult, and the economy and markets change rapidly. Look at how Nvidia quickly went from interesting HPC accelerators to the largest and most profitable systems business since IBM's System/360 mainframes in the 1960s and 1970s.

Despite Intel's missteps over the past decade, it remains an innovator in chip packaging and transistor design, and given its good manufacturing processes, we still think Intel can build compelling X86 processors for PCs and servers.

It's important to remember that Intel still has two-thirds of the X86 server market, and the X86 server decline that plagued Intel and AMD—which we think is partly due to a shift in spending focus on AI systems—is largely over.

But the problem is that X86 servers are becoming less and less relevant in the datacenter, outside of being the legacy platform for Windows Server applications, which are themselves the last proprietary operating system on the planet driven by volume economics. X86 servers have been slowly shrinking over the past decade as Linux has risen and the variety and number of server CPU platforms based on the Arm architecture that can run Linux has also increased. Hyperscale server and cloud builders and many HPC centers around the world have their own Arm processors. Nvidia has one, and we think AMD may need to give up its own Arm server chips at some point. Together, these organizations represent a growing base of server CPU acquisitions, and as we've said before, we can foresee a day when AMD, Intel, and the Arm Group all have large and relatively equal shares of the server CPU market.

In today's world, Intel can't extract the huge profits from X86 CPUs that it did in the 2010s, when it essentially had a monopoly on datacenter computing. (Just as Nvidia is doing today with data center GPU accelerator computing, and benefiting greatly from it.)

As we said four years ago, Intel has to decide whether to be a foundry or a chip design company. It may not be able to be both and still be Intel. Perhaps it does need to separate the two companies, rather than keep them together, so that they can go their own ways and either survive or die.

We think it will be easier for Intel product groups to turn to TSMC as a foundry than for Intel foundry to undercut TSMC as a second source of chips. Intel and TSMC forecasts suggest this - calculations we did in April showed that it won't be long before TSMC's AI chip manufacturing business is larger than Intel's entire foundry business.

The answer may be that TSMC acquired Intel's foundry business to gain a foothold in the United States - a real foothold, and find a place to continue operating after China invades Taiwan. (This is becoming more likely as time goes on, and it could happen soon with the Trump administration starting in a few weeks and the trade war with China expected to escalate.) In this doomsday scenario, TSMC would keep Intel's current processes, provide jobs for Intel employees, and provide future foundry for Intel products (we'll call it Intel if it comes true) as TSMC builds capacity in the U.S.

IBM tried to do this in 2014 when it paid GlobalFoundries to acquire IBM's microelectronics division, which etches Power and z mainframe chips, among other things. (That plan didn't work out very well, and IBM sued GlobalFoundries three years ago and chose Samsung as its foundry partner instead, but don't let that scare you.)

Intel's client PC business is relatively healthy, with annual revenues of about $30 billion and operating profits of about $10 billion, which is why Michelle Johnston Holthaus (who was the head of sales for the PC business, promoted to general manager of marketing and communications in 2017, and was personally appointed by Gelsinger to general manager of the Client Computing Group in January 2022) is now CEO of Intel Products and co-CEO of the entire Intel Corporation, a position she holds alongside CFO Dave Zinsner. Frank Yeary, who was the independent chairman of Intel's board of directors, is now executive chairman, and Naga Chandrasekaran, who was named general manager of Intel's foundry business in July, continues in that role.

The statement released by Intel doesn't indicate a change in Intel's strategy, but rather a change in leadership, but it also indicates that Gelsinger is retiring, and we all think he was asked to leave, because the turnaround wasn't happening fast enough.

"While we have made significant progress in restoring manufacturing competitiveness and our ability to build a world-class foundry, we know the company has much more work to do and are committed to restoring investor confidence," Yeary said in the statement. "As a board, we know first and foremost that we must put the product groups at the center of all we do. Our customers are asking us to do that, and we will deliver for them. With MJ's permanent promotion to Intel's product CEO and her interim role as Intel's co-CEO, we are ensuring the product groups will have the resources they need to deliver for our customers. Ultimately, restoring process leadership is core to product leadership, and we will continue to focus on that mission while driving efficiency and profitability."

Yeary went on to say that Intel will simplify its product portfolio and ramp up its foundry capabilities. There is no indication that Intel will sell its foundry business. As far as we know, Gelsinger wants to break up Intel, and the current board doesn't - just as Akers wanted, but Gerstner shot down the idea.

We don't know. We also don't know how the Trump administration will use executive orders and national security to force American tech giants to use American-owned foundries — whether it's Intel, TSMC, any group of private equity firms, or Uncle Sam himself. The US government could federalize Intel foundries to achieve its own ends without affecting the US Treasury balance sheet. Make no mistake, the semiconductor business and national security are at risk because Intel dropped the ball on advanced manufacturing processes a decade ago.

It will be very interesting to see who Intel hires as CEO, when, and what happens next.

View more at EASELINK