Just like Justin Timberlake brought sexy back in 2006, Pat Gelsinger is bringing the geek back to Intel 15 years later.

And with the geek’s return, the recently appointed Intel chief promised to make developers a priority — something he admits “we haven’t done a great job of recently.”

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

This philosophy was at the core of this week’s Innovation event, which offered yet another glimpse at the future of Gelsinger’s Intel.

Gelsinger Takes Control

Since Intel ousted its former CEO Bob Swan in February, Gelsinger has wasted little time bringing his vision to bear.

The engineer-turned-CEO knows the company well, having spent nearly three decades at Intel, where he served as CTO, helped architect the infamous 486 processor, and led 14 different microprocessor programs.

Less than a week into his tenure as CEO, Gelsinger addressed some of the most glaring criticisms leveled at the company and outlined four simple — but not insignificant — goalposts. These included restoring Intel’s position as a market leader, executing “flawlessly” on commitments, focusing on rapid innovation, and attracting and enabling top-tier engineers.

“We need to bring back some of the Groveian disciplines for direct, transparent, and data-driven decisions and accountability,” he wrote at the time, referring to famed Intel co-founder and CEO Andy Grove.

However, this was only the start of Gelsinger’s grand vision.

Gelsinger Cracks the Seal on Intel’s Foundries

A month later Gelsinger announced one of the largest manufacturing initiatives in the company’s history when he formed Intel Foundry Services and opened the company’s long-private fabs to contract manufacturing.

And to support this new initiative, Intel pledged a $20-billion investment to construct two new, leading-edge foundries in Arizona. The company broke ground on the facilities, which are expected to come online in 2024, late last month.

Gelsinger says that by opening its foundries and competing directly with the likes of Taiwan Semiconductor Manufacturing Co. (TSMC), Samsung Electronics, and Global Foundries, Intel stands to realize a $100-billion market opportunity by 2025 and rebalance the distribution of global semiconductor capacity.

Intel invested another $3.5 billion in early May to modernize its Rio Rancho, New Mexico chip fab to produce semiconductors using the chipmaker’s Foveros package technology.

Meanwhile, Gelsinger has spent the last several months scoping out locations for additional foundry projects destined for Europe. This spring, Gelsinger met with European Commissioner Thierry Breton to discuss the possibility of bringing foundries to the region.

Gelsinger again signaled his interest in a European expansion last month by dedicating an unspecified amount of capacity at its Kildare, Ireland, foundry to support the growing demand for automotive silicon.

The company is now weighing two more facilities in Europe with plans to invest as much as $95 billion in the region over the next 10 years.

Intel Gives Executive Staff a Hard Shake

Another dynamic on full display at Intel Innovation was the company’s latest executive shakeup.

Former VMware CTO Greg Lavender joined Gelsinger at his round table along with Nick McKeown, head of Intel’s Network and Edge group; Sandra Rivera, head of the Datacenter and AI group; and Gregory Bryant of the company’s Client Computing Group.

Gelsinger poached Lavender early this summer to fill the seat left by former Intel CTO Michael Mayberry who retired in late 2020. In addition to filling the role of CTO, Lavender oversees the company’s newly formed Software and Advanced Technology Group.

Lavender’s appointment came amid the departure of long-time executive Navin Shenoy and a major restructure of the company’s Data Platform Group (DPG). From the ashes of the DPG, Gelsinger formed the Datacenter and AI Group and Network and Edge Groups, led by Rivera and McKeown respectively.

Rivera previously led the company’s Network Platforms Group, while McKeown was the co-founder and chief scientist at Barefoot Networks prior to its acquisition by Intel in 2019.

Intel also announced a third business unit called the Accelerated Computing Systems and Graphics Group under the direction of Raja Koduri, former GM of architecture, graphics, and software at Intel.

The new group will focus on expanding Intel’s influence in the high-performance computing (HPC) space, an area currently dominated by rival chipmaker Nvidia.

However, Intel is expected to see more changes to its executive staff later next year following CFO George Davis’ decision to leave the company in the first half of 2022.

Goodbye Nanometers, Hello Angstroms

With his executive staff aligned, Gelsinger moved on to his next task: reshaping the perception surrounding the company’s troubled process technology.

In July, the chipmaker abandoned its decades-old tradition of naming process technologies based on transistor gate widths — a practice Gelsinger called a meaningless bar for comparison.

According to Gelsinger, ever since the inception of strained silicon in 1997 and later the development of fin field-effect transistors (FinFET) in 2011, this method no longer reflects the relative performance of one chip versus another.

Moving forward, Intel’s 10-nanometer Enhanced Super FinFET manufacturing process — slated for use in the company’s recently announced Alder Lake and now delayed Sapphire Rapids processor families — was renamed Intel 7.

Meanwhile, the company’s long-delayed 7-nanometer process now bears the Intel 4 moniker. The process was the source of considerable upset early last year after the company announced it had been delayed by 12 months.

According to Intel, the new naming scheme better reflects the relative performance compared to competing process nodes. And while Intel didn’t name them directly, the implication is Intel 7 and Intel 4 will compete with 7-nanometer and 5-nanometer nodes from TSMC and Samsung Electronics.

Intel also offered a glimpse at the company’s long-term process road map, which will see the chipmaker return to a measurement-based naming scheme with the launch of Intel 20A in 2024. Intel 20A will be the company’s first chip to use a backside power delivery network and adopt gate-all-around transistors, which Intel calls RibbonFET.

The chipmaker will follow the process a year later with a revision called Intel 18A. This is the same technology that IBM used to achieve a 2-nanometer process node earlier this year. However, it remains unclear whether Intel is using the same technology in its 20A process.

A Long Road Ahead

Despite Intel’s recent developments, the road ahead remains fraught with challenges.

While Intel aims to best TSMC and Samsung on process technology, both foundries are still outspending Intel by a wide margin. TSMC recently announced $100 billion in capex spending over the next three years to bolster its fab capacity, while Samsung plans to spend nearly $200 billion during the same period.

To date, Intel has announced just under $24 billion in new spending on fabs. However, the company could receive some assistance under the recently proposed CHIPs for America bill, which lays out $52 billion in subsidies and tax incentives to support domestic semiconductor expansion.

Beyond the foundry space, Intel also faces some of its stiffest competition in nearly a decade. Long-time rival AMD has experienced explosive growth in recent quarters as Intel’s leadership in the consumer and enterprise markets slipped.

AMD posted its fifth straight quarter of greater than 50% revenue growth this week, and the chipmaker is now targeting full-year revenues of $16 billion, up 64% from the prior year.

And while Intel is a much larger company with annual revenues in the range of $77 billion, the company struggled in recent quarters to maintain positive growth.

However, AMD isn’t the only threat to Intel’s dominance. The chipmaker also faces competition from a new cohort of Arm chipmakers including Marvell, Ampere Computing, and even Apple.

Ampere’s latest Altra Max chips offer core counts more than three times that of Intel’s 40-core processors and twice that of AMD's top-tier chips.

Intel's upcoming Sapphire Rapids Xeon Scalable refresh is expected to bring higher core counts along with a slew of next-generation technologies like DDR5, high-bandwidth memory, PCIe Gen. 5.0, and compute express link. However, the release has already been delayed to early 2022 to provide additional time for validation.