Analysis: Intel Seeks To ‘Stabilize’ Waning Server CPU Share, But It Faces AMD And Other Challenges

While interim Intel co-CEO Michelle Johnston Holthaus says ‘this year is all about improving’ Xeon’s competitive position ‘as we fight to close the gap to competition,’ the semiconductor giant is facing challenges in the form of AMD, Arm and a delay for a key product.

After failing to achieve its modest revenue goal for AI accelerator chips last year, Intel’s ambition to compete with Nvidia was dealt another blow when it announced last week that it would cancel its next-generation Falcon Shores AI chip that was due in late 2025.

With Intel now turning its focus to a successor AI chip called Jaguar Shores set for release further into the future, the semiconductor giant is facing a greater sense of urgency to protect its data center business by defending its dominant but waning server CPU market share against AMD and companies developing Arm-based alternatives.

Michelle Johnston Holthaus (pictured), interim Intel co-CEO and CEO of Intel Products, said during Intel’s earnings call last week that “this year is all about improving” the competitive position of its Xeon CPU products “as we fight to close the gap to competition.”

The company’s Xeon 6 CPUs with performance cores, code-named Granite Rapids, “has been a good first step” after they debuted late last year, according to Holthaus.

Intel is also “making good progress” with Clearwater Forest, its next-gen Xeon processors with efficient cores that will debut in the first half of 2026 as the first server product to use the Intel 18A manufacturing process, she added. (Intel 18A is the last node in former Intel CEO Pat Gelsinger’s comeback plan to surpass Asian foundry rivals TSMC and Samsung.)

Holthaus said all this work provides a “strong foundation on which to build,” noting that most of the world’s data center workloads “still run on Intel silicon” along with the company’s “strong ecosystem, especially within enterprise.”

“We are going to leverage these strengths as we work to stabilize our market share in 2025,” she said.

It’s an important fight for Intel. In November, CPU-tracking firm Mercury Research reported that AMD’s share in the x86 server CPU segment reached 24.2 percent against Intel, which means that the rival is two points from reaching the record it hit in 2006. This, along with AMD’s early success in the AI accelerator chip market, allowed the competitor to make nearly as much data center revenue as Intel did last year.

Intel is also facing a threat in the form of Arm, whose instruction set architecture that has dominated smartphone chips is now the basis for custom server chips designed by Amazon Web Services, Microsoft Azure and Google Cloud for their own infrastructure. In fact, AWS said in December that more than half of the new CPU capacity it brought online over the last two years was based on its Arm-based Graviton chips.

How Intel Plans To Fight Back In The Data Center

Two major ways Holthaus said Intel plans to win back server CPU share is “by re-engaging the x86 ecosystem” with the recently formed x86 ecosystem advisory group—which has the backing of Microsoft, Dell Technologies, Google, AMD and other major tech vendors—and by building semi-custom and custom solutions, like it is doing for AWS.

Holthaus also noted that Intel “has a leading position as the host CPU for AI servers, and we see a significant opportunity for CPU-based inference on prem and at the edge as AI-infused applications proliferate.”

Another element that will help Intel compete is competitive pricing. In response to a question from a Wall Street analyst on whether competitive pricing will weigh down product margins throughout the year, Holthaus said Intel needs to win every deal it can. (Intel reportedly cut the prices of its Granite Rapids CPUs by as much as 30 percent last month.)

“We have to stem the tide of share loss in data center, and so we will be fighting for every socket in that business. And the way I look at it is we need to be aggressive, we need to win share, and we need to show our customers that they can win with us,” Holthaus said.

Intel Global Channel Chief Dave Guzzi told CRN in an exclusive interview in January that while the company has cut its direct partner coverage by an unspecified amount, it has increased funding for partner incentive and investment programs, which includes money used for winning customer deals when Intel is faced with a competitive bid.

Intel’s Challenges: Arm, AMD And A Delay

While Holthaus touted Intel’s Xeon 6 P-core Granite Rapid CPUs as a positive step in becoming more competitive again, she suggested the other side of the Xeon 6 lineup featuring E-cores, known as Sierra Forest, isn’t getting the traction it had hoped for.

“What we’ve seen is that’s more of a niche market, and we haven’t seen volume materialize there as fast as we expected,” she said.

Whereas the Xeon 6 processors with P-cores continued Intel’s legacy of delivering high-performance CPU cores in servers, the E-core chips are optimized for efficiency, allowing the company to deliver a higher number of cores with lower energy consumption.

With Intel pushing the E-core processors for cloud-native and telecom workloads, the chips have been viewed as Intel’s response to the high-density, energy-efficient custom Arm-based processors that are allowing Intel’s largest cloud customers—AWS, Microsoft Azure and Google Cloud—to lower their reliance on the semiconductor giant.

Intel has also been faced with Ampere Computing, a chip design startup founded by former Intel executive Renee James that provides its own Arm-based processors with high core counts and low energy consumption for cloud-native workloads.

But while Ampere originally planned to go public through an initial public offering, Bloomberg reported Wednesday that Japanese investment giant SoftBank is in advanced talks to acquire the company, whose primary backer has been Oracle.

On the other hand, Qualcomm, which has begun selling custom Arm-based chips for PCs, has signaled that it plans to enter the server market with the recent hiring of Intel Xeon chief architect Sailesh Kottapalli, among other developments.

Then there’s AMD, which beat Intel in releasing energy-efficient x86 server CPUs with high core counts for cloud-native workloads in 2023. The chip designer then released a follow-up to its cloud-native server chips last year.

During AMD’s earnings call on Tuesday, the company’s CFO, Jean Hu, didn’t seem fazed that Intel planned to fight hard this year, saying that the chip designer “always” assumes that the server CPU market is “very competitive.” She added that AMD has the “best lineup” and the “best” total cost of ownership with its fourth- and fifth-gen EPYC CPUs.

“So overall, we are actually quite confident about [continuing] to drive the server CPU business,” which is growing in unit sales, average selling prices and market share, she said.

In the same call, AMD Chair and CEO Lisa Su said the company’s server business “has been performing extremely well” between cloud and enterprise customers.

According to Su, the company has seen “some strong traction” in the enterprise thanks to go-to-market investments it made last year—which the leader of AMD’s server CPU business detailed in a CRN interview a few months ago.

“The enterprise sales cycle is often a six- to nine-month sales cycle, but as we’ve invested more resources in it throughout 2024, we’ve seen that convert into a significant number of new [proofs of concept] that are now converting into volume deployments,” she said.

The other challenge Intel is facing is an internal one.

On the company’s earnings call, Holthaus acknowledged that Intel’s plan to release its next-gen Xeon E-core product, Clearwater Forest, in the first half of 2026 amounts to a delay from the original 2025 timeframe the chipmaker had given.

Clearwater Forest has been seen as a strategically important product in large part because of how it will be the first server CPU to use the Intel 18A node, which the company has repeatedly said will allow it to surpass the advanced chip-making capabilities of TSMC and Samsung. Several months before Gelsinger abruptly left the company as CEO, he said the product would allow Intel to “accelerate share gains.”

Despite the delay, Holthaus said she still has high hopes for Clearwater Forest.

“It does have some complicated packaging expectations that moved it to 2026, but we expect that to be a good product and continue to close the gap as well. But this is going to be a journey. It’s not a destination,” she said.