The 10 Biggest Intel News Stories Of 2023 (So Far)

Catch up with the most significant developments for Intel in the first half of 2023. The chipmaker’s biggest news stories so far this year include developments related to its fledgling GPU business, exits from noncore businesses, cuts to the company’s dividend and staff pay, and a change in how Intel runs its manufacturing business.

Pat Gelsinger is now more than two years into his tenure as Intel’s eighth CEO, and the pace of changes at the semiconductor giant has been almost as “relentless” as the company veteran’s desire to innovate and return the chipmaker to its former glory.

While Gelsinger has maintained that Intel is on track or ahead with his plan to surpass the chip manufacturing capabilities of Asian foundry giants TSMC and Samsung by 2025, he has had to lead the chipmaker under less-than-ideal conditions.

[Related: Oracle Bets On AMD, Ampere CPUs; Ellison Says ‘Intel x86 Architecture’ Is ‘Reaching Its Limit’]

The Santa Clara, Calif.-based company has not only had to deal with a slowing economy that weakened demand for its chips so far this year, but it has also had to contend with rivals like AMD and Nvidia that continue to push out competitive products. The chipmaker faces another threat in the form of Arm, the chip architecture powering most of the world’s smartphones that is now finding increased adoption in the PC and server markets by Apple, Amazon Web Services and others.

“While the macro is challenging, and [there are] plenty of headwinds out there, we also believe that our execution on our financials, a beat on top and bottom line, great execution on our process and product road maps, and here we are, two years into my tenure, and the journey to date has had some unexpected bumps in the road,” Gelsinger said during Intel’s first-quarter earnings call in April.

“We are also beginning to see clear points that increase my confidence that we have the right strategy, the right team, and we are executing on this transformation,” he added.

What follows are the 10 biggest Intel news stories of 2023 so far, which include developments related to the company’s fledgling GPU business, exits from noncore businesses, cuts to the company’s dividend and staff pay, and a change in how Intel runs its manufacturing business, among other topics.

No. 10: Intel Cuts Dividend After Announcing Staff Pay Cuts First

Intel in February announced that it had slashed its quarterly dividend to shareholders by nearly 66 percent to 12.5 cents, which started with the payment to investors in June.

The chipmaker said it was making the move as part of its cost-cutting efforts to save money for its ambitious comeback plan. But the move came after Intel employees asked management why it had not yet cut its dividend when the company told staff a few weeks earlier that it was cutting the base pay of senior employees and executives in addition to suspending merit-based pay raises, quarterly bonus programs and employee recognition programs.

“When we communicated the recent compensation and reward changes, two questions consistently rose to the top: When will we restore these programs and reward employees who remain with the company? And why aren’t we cutting the dividend?” Gelsinger wrote in a mid-February memo when he announced the dividend cut.

In the same memo, Gelsinger announced that the company planned to restore the salaries of executives and senior employees, along with other compensation and reward programs, this fall.

No. 9: Intel Co-Founder Gordon Moore Dies

Gordon Moore, one of Intel’s two co-founders and the pioneer behind “Moore’s law” that drove decades of semiconductor innovation, died on March 24 at the age of 94.

Moore’s death was announced by Intel and the Gordon and Betty Moore Foundation, who said Moore died at his home in Hawaii. He was born on Jan. 3, 1929, in San Francisco.

Moore founded Intel with Robert Noyce in 1968. Moore first served as Intel’s executive vice president from the company’s founding until 1975, when he took over as president. He then took over CEO responsibilities from Noyce and also became chairman of the board in 1979.

Moore stepped down as CEO in 1987 and continued as Intel’s chairman until 1997. He then held the title of chairman emeritus for the company until 2006.

Before helping found Intel, Moore was a co-founder of Fairchild Semiconductor, which developed the world’s first commercially viable integrated circuits.

When he was at Fairchild, Moore in 1965 said that the number of transistors in an integrated circuit would double every year. A decade later at Intel, Moore revised his forecast to say the number of transistors would double every two years. This notion became known as Moore’s law.

While some in the semiconductor industry, like Nvidia CEO Jensen Huang, have said that Moore’s law has come to an end because the reliable scaling of transistor densities has come to an end, Gelsinger has said the notion is being kept alive by new techniques in chip design and manufacturing, such as advanced chip packaging technologies.

No. 8: Intel Eyes Closer Ties With Arm

While Intel competes with companies designing Arm-based chips for PCs, servers and embedded systems, the chipmaker has signaled an increasing interest in working with the namesake company that develops the instruction set architecture.

As part of efforts to drum up business for its revitalized contract chip manufacturing business, Intel has made it clear that it will manufacture chips for customers working on Arm-based designs. In April, that commitment became clearer when the company’s contract manufacturing division, Intel Foundry Services, unveiled a “multigeneration agreement” with Arm that will “enable chip designers to build low-power compute system-on-chips” on its 18A manufacturing node.

Intel Foundry Services said this agreement will include design technology co-optimization efforts that will optimize Arm chip designs for Intel’s 18A process to improve power, performance, area and cost. The agreement will focus on Arm chip designs for mobile devices first, but Intel said the collaboration could eventually expand into automotive, IoT, data center, aerospace and government applications.

A few months after the agreement was unveiled, Reuters reported in mid-June that Intel was approached by Arm to become an anchor investor in the British chip designer’s initial public offering, which is expected to happen later this year. Gelsinger said early last year that the company would potentially be interested in joining a consortium to buy Arm.

No. 7: Intel Exits More Businesses, Sells Minority Stake In Another

Intel continued to exit more businesses as part of the semiconductor giant’s efforts to double down on its core chip business and cut costs elsewhere.

The business cuts this year began in late January when Gelsinger said the company will end future investments in the Tofino network switch chip product line that originated from Intel’s 2019 acquisition of Barefoot Networks.

In March, the company confirmed that it had exited its wireless wide-area network business for LTE- and 5G-connected PCs. According to analyst Ian Cutress, Intel was expected to transfer 5G technologies to Fibocom and MediaTek, which collaborated with Intel on 5G modems. As for Intel’s 4G modem portfolio, the company had begun the discontinuation process for the products.

Intel then disclosed in April that it had shut down its prebuilt server system business, the Data Center Solutions Group, and gave permission for Taiwan-based MiTAC, parent company of server vendor Tyan, to “manufacture and sell products based on our designs.”

The chipmaker also said in April that it had discontinued its Blockscale 1000 Series ASIC for cryptocurrency mining to “prioritize our investments” in its IDM 2.0 strategy.

Most recently, Intel announced in June that it had reached a deal to sell a roughly 20 percent stake in its IMS Nanofabrication business to private equity firm Bain Capital. The transaction values IMS at roughly $4.3 billion, and it’s expected to close in the third quarter. IMS develops multi-beam mask writing tools, which are required to created advanced EUV masks, a critical component of next-generation chips.

No. 6: Intel Plans To Run Fabs As Separate Business

In a major shakeup to Intel’s business model, the chipmaker said in mid-June that it will run its manufacturing operations as a separate business that will act like a third-party contract chipmaker, such as TSMC or Samsung, to the company’s product groups.

Intel said this move, which it calls an “internal foundry” model, is key to the transformation of the semiconductor giant’s integrated device manufacturing strategy, dubbed IDM 2.0. The IDM 2.0 strategy is an instrumental part of Gelsinger’s comeback plan, which seeks to put the company’s chip manufacturing capabilities ahead of TSMC and Samsung by 2025.

By breaking out its manufacturing operations into a stand-alone business, Intel’s product groups such as the Client Computing Group and the Data Center and AI Group will be on equal footing with customers of the company’s revitalized contract chip-making business, Intel Foundry Services.

“Intel will extend the use of market-based pricing to its internal business units, offering them the same certainty and stability as the company’s external customers,” Intel said.

The company added that its “manufacturing groups will face the same market dynamics as their external foundry counterparts and need to compete for volume through performance and price.”

“This includes Intel’s internal customers, who will have the flexibility over time to engage with third-party foundries,” according to Intel.

As a result of this change, the chipmaker said, it will begin presenting profit and loss statements for its manufacturing operations in the first quarter of 2024.

Intel said this internal foundry model will “play a significant role” in the company’s move to make billions of dollars in cuts through 2025. It will also lead to higher gross margins, according to the company.

No. 5: Intel Completes Installation Of Long-Delayed Aurora Exascale Supercomputer

Intel in mid-June said it had completed the installation of more than 10,000 compute blades in the U.S. Department of Energy’s long-delayed Aurora exascale supercomputer, a significant milestone meant to showcase the chipmaker’s advances in CPUs and GPUs for high-performance computing and AI.

The chipmaker completed the installation with Hewlett Packard Enterprise, saying each of the 10,624 server blades contain two Intel Xeon CPU Max Series and six Intel Data Center GPU Max Series processors. This amounts to a total of 21,248 Intel Xeon CPUs and 63,744 Intel data center GPUs installed in a space that is equivalent to two professional basketball courts, according to Intel.

Intel said Aurora, originally unveiled in 2015 before suffering a series of delays and a major architectural shift from Intel, is “expected to be the world’s first supercomputer to achieve a theoretical peak performance of more than 2 exaflops.” That amounts to more than 2 quintillion, or more than 2 billion billion, floating point operations per second.

Intel said it expects Aurora to hit this milestone later this year when Argonne submits performance data for Aurora to Top500, the organization that ranks the world’s fastest supercomputers. If verified by Top500, it would make the supercomputer the fastest in the world, surpassing the horsepower of the first publicly verified exascale system, AMD-powered Frontier at the DOE’s Oak Ridge National Lab.

No. 4: Intel Aims At Nvidia With Data Center GPU Launch

Intel launched its long-awaited response to Nvidia’s domination of the data center GPU market in January with the Intel Data Center GPU Max Series.

Focused on the convergence of high-performance computing and AI workloads, the company’s new GPU line has been installed in the U.S. Department of Energy’s Aurora exascale supercomputer, which could become the fastest in the world when it’s expected to go online later this year.

Each GPU consists of more than 100 billion transistors on a single package, and it sports up to 128 of Intel’s Xe cores, each of which come with Intel Xe Matrix Extensions for accelerating AI workloads as well as enabling vector and matrix math capabilities in a single device. The GPU also comes with up to 128 GB of HBM2e memory and 128 ray tracing units to simulate realistic lighting in real time.

In May, Intel said the Max Series GPU provides better performance than an Nvidia H100 PCIe card, which is a slower version of Nvidia’s H100 GPU, “by an average of 30 percent on diverse workloads.” The company also stated that ISV Ansys measured a 50 percent speed increase for the Max Series GPU over Nvidia’s H100 PCIe card on AI-accelerated HPC applications.

Intel had spent several years developing the Max Series GPU, which was previously known as Ponte Vecchio, as part of broader efforts to address diverse workloads with an expanded portfolio of chips.

No. 3: Intel GPU Lead Raja Koduri Leaves Company, Amid Other Exec Moves

Raja Koduri, the chief driver behind Intel’s discrete GPU efforts, left the company in March after spending more than five years leading the buildout of its graphics capabilities and other technologies.

Koduri, whose departure was among a handful of significant executive moves within Intel in the first half of the year, was most recently executive vice president and chief architect. He is now working on a new software company “focused on generative AI for gaming, media and entertainment,” according to Gelsinger. Koduri also joined the board of AI chip startup Tenstorrent, which is led by fellow AMD architect veteran Jim Keller, who was previously at Intel for two years.

Koduri joined Intel in 2017 as chief architect and senior vice president of what was the newly formed Core and Visual Computing Group at the time and general manager of edge computing solutions.

In mid-2021, Koduri moved to the new role as general manager of the Accelerated Computing Systems and Graphics Group, a new business unit formed by Gelsinger to focus on Intel’s GPUs. Roughly a year and a half later, Intel moved Koduri back to the chief architecture role and split the graphics division into two, moving the company’s client GPU efforts under the Client Computing Group and its data center GPU efforts under the Data Center and AI Group.

The company this year also saw the departure of Randhir Thakur, who was replaced by Stuart Pann as president of Intel Foundry Services; the resignation of Omar Ishrak as Intel chair, a role that was transferred to board director Frank D. Yeary; and a change in roles for Nick McKeown, who became chief innovation officer for the Network and Edge Group after serving as its general manager. Sachin Katti took over as the head of the Network and Edge Group after serving the role in an interim capacity.

No. 2: Lower PC, Data Center Demand Impacts Financials

A major slowdown in business for Intel that began last year became even more pronounced for the chipmaker in the first few months of 2023.

For Intel’s first quarter, reported by the company in late April, revenue declined by 36 percent year over year to $11.6 billion while its net income plummeted by 134 percent to a loss of $2.8 billion from a profit of $8.1 billion during the same period last year.

Behind these precipitous drops in revenue and profit were Intel’s three largest businesses—the Client Computing Group, the Data Center and AI Group, and the Network and Edge Group—which all reported revenue declines of 30 percent or more.

Intel blamed the Client Computing Group’s 38 percent decline in revenue on lower demand for PCs as well as OEMs focusing on clearing out existing inventory of Intel components. For the Data Center and AI Group, Intel said the division’s 39 percent decline in revenue was the result of lower demand for data centers and “competitive pressure.” The Network and Edge Group declined by 30 percent to due to customers continuing to work through existing chip inventories, according to the company.

Gelsinger said he expects a “modest recovery” in the second half of 2023, thanks in part to a ramp in production of the company’s next-generation client CPU, Meteor Lake, and its next-generation server CPU, Emerald Rapids, in the second half of the year.

No. 1: Intel Makes Changes to Data Center GPU, AI Chip Road Maps

Intel unveiled changes to road maps for its data center GPUs and AI chips in the first half of the year.

In March, Intel said that it no longer planned to release data center GPUs every year, instead opting to launch products in the category every two years.

The road map changes resulted in Intel canceling development for Rialto Bridge, which was previously expected to start sampling with customers in mid-2023. Rialto Bridge was supposed to serve as the successor to the first generation of the Intel Data Center GPU Max Series, which launched for high-performance computing and AI workloads in January and was previously known as Ponte Vecchio.

The chipmaker said it had also pushed back the launch of Falcon Shores, a next-generation accelerator chip in the Max Series, to 2025 from 2024. The company also changed expectations for Falcon Shores, saying it would be available in a GPU-only configuration in 2025 and that variants with CPU cores would come later. The company previously promised that Falcon Shores would bring together x86, Xe GPU cores and memory into one chip package that can fit into a single server socket.

Intel also unveiled changes to its other segment of data center GPUs, the Intel Data Center GPU Flex Series for video transcoding, cloud gaming, virtual desktop infrastructure and AI inference workloads. These changes resulted in the cancellation of Lancaster Sound, the successor to the first-generation Flex Series GPU that launched last year. The company will instead focus on Melville Sound, which it said would provide a “significant improvement” in performance, features and covered workloads.

Intel unveiled further changes to plans for Falcon Shores in May, saying that the upcoming chip would no longer include variants that combine GPU and CPU cores. The company also said that it would incorporate technology from its Habana Gaudi deep learning processors into Falcon Shores. This means that there won’t be a successor to the Habana Gaudi 3 chip, which hasn’t released yet, as Intel is merging the road maps for Habana Gaudi and the Max Series GPUs.