In just shy of a year, Pat Gelsinger, Intel’s favorite son and CEO, managed to turn the faltering chipmaker’s fortunes around in more ways than one.

The former VMware chief’s return to Intel late last winter came at a precarious moment for the company, which was under intense pressure not only from rival chipmakers, but its own investors. Two weeks prior to the announcement, activist hedge fund Third Point LLC threatened to install new members to Intel’s board if the company didn’t take steps to shore up its competitive standing.

The chipmaker a few months earlier was rocked by news of a manufacturing defect in its upcoming 7-nanometer manufacturing process, setting it back by a year.

Faced with these challenges, Gelsinger, who previously spent 30 years at Intel and led development of 14 different microprocessor programs, wasn’t satisfied with getting the company’s manufacturing capabilities back on track. Instead, the engineer turned executive set out to reinvent the company that jumpstarted his career. Here are some of his biggest moves of 2021.

Gelsinger Goes Big, Boosts Chip Production

Arguably Gelsinger’s biggest and boldest move came a month into his tenure as CEO when he announced Intel would open its fabrication facilities to contract manufacturing and construct two leading edge chip fabs in Arizona at a cost of $20 billion. The facilities broke ground in September.

The move, Gelsinger boasted, represented an opportunity to both realize a $100 billion market by 2025 and rebalance the distribution of semiconductor capacity, which is heavily centered in the Asia-Pacific region.

The foundry expansion was, however, only the beginning. The chipmaker in May announced a $3.5 billion upgrade to its Rio Rancho, New Mexico foundry to produce chips using its Foveros packaging technology.

And earlier this month, Intel announced a $7 billion expansion in Malaysia, which will see the vendor expand its packaging footprint in the region to address the growing demand for semiconductors.

Finally, Gelsinger is actively seeking support for an expansion in the European market, where it expects to invest as much as $95 billion over the next decade. The company currently plans to add two foundries in the region pending financial backing from European leaders. Each foundry is expected to cost around $10 billion.

Gelsinger Turns Intel on its Head

Gelsinger’s next order of business came early this summer when he gave his executive staff a hard shake and poached former VMware CTO Greg Lavender to fill the seat left by former CTO Michael Mayberry who retired late last year.

In addition to his duties as CTO, Lavender now oversees the newly formed Software and Advanced Technology Group, which according to Intel, is tasked with driving Intel’s technical innovation and research programs, including Intel Labs.

The announcement came amid the exit of longtime executive Navin Shenoy, who previously led the company’s lucrative Data Platform Group (DPG). Following Shenoy’s departure, the chipmaker broke DPG into two business units — the Datacenter and AI Group and Network and Edge Group — and reshuffled its executive staff.

Intel executive Sandra Rivera, who previously led the company’s Network Platforms Group, now leads the Datacenter and AI Group. Meanwhile, Nick McKeown, the co-founder and chief scientist at Barefoot Networks, which Intel acquired in 2019, now heads up the company’s Networking and Edge Group.

The shakeup also saw the formation of a new business unit called the Accelerated Computing Systems and Graphics Group — AXG for short — which is focused on expanding Intel’s presence in the high-performance computing space, an area currently dominated by rival chipmakers Nvidia and AMD.

Leading the group is Raja Koduri, who most recently served as Intel’s GM of architecture, graphics, and software. Intel reentered the GPU market last year with a series of products targeting cloud game streaming and consumer devices.

Gelsinger Says Bye to Nanometers, Welcomes Anstrom Era

Gelsinger returned to the spotlight a month later to announce Intel would abandon its decades-old tradition of naming process technologies based on transistor gate widths — a practice he called a meaningless bar for comparison.

“Originally, process nodes were named after the physical length of these transistor gates,” he said at a press conference in July.

He explained that, since the inception of strained silicon in 1997 and later the development of fin field-effect transistors (FinFET) in 2011, this method no longer reflected the relative performance of one chip versus another.

Because of this, Intel rebranded its process technology to better reflect its relative performance and transistor density compared to competing offerings. With the change, the chipmaker’s 10-nanometer process became Intel 7.

“This is equivalent to a full node of performance gain,” said Ann Kelleher, SVP and GM of technology development at Intel, inviting a comparison to TSMC’s 7-nanometer process used by AMD.

Intel’s ill-fated 7-nanometer process, meanwhile, will bear the Intel 4 moniker when it officially launches in 2023.

The chipmaker’s aversion to measurement-based naming schemes will, however, be short lived. The company’s first “angstrom-era” chip, dubbed Intel 20A, will hit the market in 2024. An angstrom is a unit of measurement that is one-tenth of a nanometer.

Intel 20A will also be the chipmaker's first to feature a backside power delivery and adopt a gate-all-around transistor that vertically stacks transistor gates within a channel, allowing for greater compute densities for a given die size.

IBM used the same technology to achieve a 2-nanometer process node earlier this year, which coincidentally is also equivalent in size to 20 angstroms and is slated for release in 2024.

Gelsinger Champions Developer Focus

Not all of Gelsinger’s efforts have been focused on technology. In a speech at Intel’s Innovation virtual event this fall, Gelsinger embraced a customer-first philosophy and promised to make developers a priority.

“Let me be clear, Intel’s commitment to you is a developer-first approach,” Gelsinger said in his keynote speech “We can and we will do better… Through the developer community, we are entering a renewed era of innovation.”

Gelsinger reinforced that message in a LinkedIn post wherein he reaffirmed Intel’s commitment to building an open ecosystem of technologies.

“Innovation thrives in an open, democratized environment where people can connect, communicate, and respond together to new stimuli,” he wrote. “Now more than ever, the challenges of the world demand innovation. They demand transparency. They are solved in the open.”

Many of the technologies we take for granted today — personal computing, the internet, and the networks that connect society — wouldn’t have been possible were it not for open collaboration, often between rivals, he said.

In support of this mission, Gelsinger pledged to strengthen the company’s “deep legacy in open platforms,” invest in open software stacks like oneAPI, and empower collaboration among developers, academia, and the broader ecosystem to further research in areas like neuromorphic and quantum computing.

Delays Shorten Intel's Next-Gen Processor Lead

Intel’s ambitious new trajectory hasn’t been without its challenges. One of the biggest blows came this summer when Intel was forced to delay the launch of its next-generation Xeon Scalable processors until the first quarter of 2022.

The reason for the delay, according to Lisa Spelman, VP and GM of Intel’s Xeon and Memory Group, was to provide original equipment manufacturing and system integrators additional validation time prior to production.

Intel’s Sapphire Rapids processor family is the company’s first with support for DDR5, high-bandwidth memory, PCIe Gen. 5.0, and the Compute Express Link (CXL) standards.

It’s also Intel’s first processor to embrace a chiplet architecture, which enables multiple smaller processor dies to be packaged together on a single substrate. The approach is reminiscent of the one used in AMD’s Threadripper and Ryzen processor families to increase core counts well in excess of Intel’s highest performance chips to date.

Given the complexity and the sheer number of new features, the delay isn’t all that surprising or out of the ordinary for Intel. The delay, however, does shorten what would have been a substantial lead over rival AMD, which isn’t expected to launch comparable server processors until next fall.