Computex Chronicles Part 5: Intel (INTC) Strikes Back

In a time of remarkable innovation and intense competition in the semiconductor market, it will be interesting to see how key players have refined their strategies and messaging at this year’s Computex show in Taiwan. In the face of trends like generative AI (GenAI) and competitors, big tech companies are working to position themselves in unique and differentiated ways.

For Intel (Nasdaq:International Trade Commission) CEO Pat Gelsinger said^Number Part of the CEO’s main keynote here was tasked with reassuring the industry that the longtime leader was back in full force. With his usual energy, Gelsinger laid out the company’s broad vision, detailing new server architectures (Xeon 6), AI accelerators (Gaudi 3) and a new Copilot+-compliant AI PC architecture (next-generation Core Ultra, code-named “Lunar Lake”). Along the way, he also highlighted the company’s commitment to transformation. The company is focused on its foundry chip manufacturing business for Intel’s own products and for other potential partners, many of whom also gave keynote speeches at Computex.

The overarching theme that connected all these points was a focus on speed of execution. Gelsinger gave several examples of how the company has accelerated schedules for various products, especially the Lunar Lake PC architecture, proving that Intel has overcome recent challenges and entered a new era of technology leadership and innovation. In reality, Intel is still several years away from regaining the undisputed leadership position it once had, and that Gelsinger has been striving for since taking over as CEO in 2021. But this year’s Computex keynote felt like a tipping point, with enough impressive products and technologies to make people believe the Intel of old is ready to fight back.

In the server and data center space, the company unveiled its Xeon 6 architecture, the first of which is being released just six months after its launch five years ago.^Number Next-gen Xeon. Like other major computing architectures, Intel’s Xeon offers both efficiency and performance cores, each designed for different types of workloads. The new Xeon 6 E core product, code-named “Sierra Forest,” has been officially launched here in Taiwan. It focuses on the growing power efficiency challenge of data centers. Intel illustrated the improvements by showing that a few servers with the new chips can match the performance of a rack of older Xeon-based servers. Intel also said that while power efficiency is the main focus of the Xeon 6 E core, the upcoming version with 288 cores will be able to handle larger workloads.

As the name suggests, the Xeon 6P core-based CPUs, codenamed “Granite Rapids,” are optimized for higher performance (and higher power) for the most demanding types of workloads. They were originally expected to launch in 2025, but Gelsinger commented that they are now expected to be available by the end of this year.

Intel’s Gaudi 3 (a GPU-based AI accelerator chip) has been announced and its performance capabilities discussed at previous events, but during the Computex keynote, the company announced it would be competing with rival Nvidia (NVDA) cards. Gaudi 2 is 1/3 the price and Gaudi 3 is 2/3 the price, highlighting what we believe to be better value for organizations looking for alternatives.

The biggest news from Gelsinger’s keynote was the official announcement and architecture details of the Lunar Lake SOC, which focuses on AI PCs.MSFT) Lunar Lake, which meets the requirements of the Copilot+ AI PC, features a new CPU architecture, a significantly enhanced Xe2 GPU, and a 48 TOPS NPU. Equally, or perhaps even more importantly, Intel emphasized the power efficiency of the overall SOC. As the company explained in a pre-event briefing, a chip’s power efficiency is not necessarily determined by its instruction set. In fact, Intel believes that the x86-based Lunar Lake portion can compete at a power level with Qualcomm’s Arm-based Snapdragon X-series products. Unfortunately, neither portion has officially shipped, so there are no benchmarks to prove this one way or the other yet, but this area will undoubtedly attract a lot of attention once both ship.

One of the reasons Intel is so confident about this technology is because of Luna Lake’s unique design.AAPLLike Intel did with its M-series SOCs for Macs, Intel’s Lunar Lake has memory directly on the chip, which has the benefits of lower power consumption and faster access to memory, and likely plays a big role in the chip’s overall performance and efficiency.

Besides the product details, the other big surprise about Lunar Lake is that its launch is scheduled for this fall. AMD will be releasing new Ryzen AI 300 series chips ( Part 2 of this series), Intel has significantly accelerated its release date to better compete with Qualcomm’s offering. As a result, we’ll see an incredible variety of Copilot+ AI PC systems available for the fall and holiday shopping season. In fact, several PC makers will be selling systems with all three chips. How they’ll cleverly categorize these different products without completely confusing potential buyers remains to be seen, but it’s shaping up to be an exciting time for the PC market later this year.

One challenge Intel may face is the availability of x86 and Intel NPU-specific versions of its Copilot+ software; Microsoft still needs to complete that work, and there’s no firm date yet for when that will be available. Interestingly, Intel is coming to market later than AMD, so it may be able to avoid Copilot+-enabled systems needing a software update before they can use the full AI capabilities, but we won’t know for sure for a few months.

Another important factor to remember in all these developments is that despite some recent challenges, Intel is still a giant in the computing (especially PC) industry with very broad and deep roots throughout the supplier ecosystem. This is something that Intel has emphasized to all of its partners in the Taiwanese ecosystem and will ultimately impact a number of things, including the number of systems available and the total number of systems sold.

Intel has a long history of setting long-term roadmaps for its key products, and during the keynote, Gelsinger did the same for future PC architectures. Next year’s Panther Lake chips in particular are going to be interesting because not only will they build on improvements from Lunar Lake, but they will also be the first parts to be manufactured on Intel’s 18A process technology. This will signal what Intel believes is a return to semiconductor process and performance leadership.

While the final story is yet to be told, the good news to come out of Intel’s keynote is that competitive confidence is clearly returning, inspiring the entire computing industry to level up.

Disclaimer: Some of the authors’ clients are vendors in the technology industry.

Disclosure: none.

Source: Author

Editor’s note: The summary bullet points for this article were selected by Seeking Alpha editors.