Intel Corp Earnings - Q1 2025 Analysis

Positives

  • Revenue in Q1 was $12.7 billion, at the high end of guidance range, driven by better-than-expected Xeon sales.
  • Non-GAAP gross margin was 39.2%, approximately 3 percentage points above guidance on stronger-than-expected demand for Raptor Lake, combined with improved cost for Meteor Lake.
  • Intel Foundry delivered revenue of $4.7 billion, up 8% sequentially on pull-ins of Intel 7 wafers and increased advanced packaging services.
  • The company saw higher-than-expected growth in Q1 driven by a few hyperscalers.
  • The company's spending was down $400 million sequentially and $700 million YoY as it continues to focus on optimizing its cost structure.

Q&A Highlights - Q1 2025

  • Analyst asked about balancing fixing the roadmap with filling the Foundry and ensuring unit volumes.

    Lip-Bu Tan explained that the approach is to flatten the organization to a number of layers, allowing for a focus on the right products and delivering customer solutions on time. He also stated that Intel 18A is important for Panther Lake and that the company is working to improve yield and reliability to meet internal customer needs.

  • Analyst asked about gross margin for March, specifically regarding COGS headwinds from the way grants are accounted for.

    David A. Zinsner explained that the gross margin for Q1 was better than expected, with the biggest benefit coming from a handily beat on the top line. He also mentioned that the company had more volume from Raptor Lake versus Lunar Lake, which was mix beneficial to gross margins. However, he noted that the tariff dynamics do put additional pressure on gross margins through the year.

  • Analyst asked about the company's assessment of Panther Lake and Clearwater Forest, the 18A product portfolio, and their leadership position.

    The company is focused on execution and delivery on time, performance, and industry trends. Panther Lake is a priority for the company, and they are working on getting the first SKU out by the end of the year. Clearwater Forest is a derivative product based on Granite Rapids, and the company has prioritized Panther Lake over Clearwater Forest.

  • Analyst asked about the comments made about Intel 7 being constrained for the foreseeable future.

    Intel 7 is constrained due to better performance on Raptor Lake and Intel 7 parts. The company plans to leverage the network to produce more wafers and make every wafer less expensive. The company wants to see AI PC products gain more traction in the market and is optimistic that happens in the year.

  • Analyst asked about Intel's AI strategy, specifically focusing on new and emerging workloads like reasoning models, Agentic AI, and Physical AI.

    LipBu Tan, Intel's CEO, responded that the company's plan is to build purpose-built silicon and software to optimize for these new workloads, and that they are exploring new architectures and disruptive platforms to achieve this goal. He also mentioned that edge and inference are important areas for AI, and that Intel is looking to build an efficient architecture for these workloads.

  • Analyst asked about Intel's CapEx plan for the next two years, specifically regarding rationalizing the footprint.

    David A. Zinsner, Intel's CFO, responded that the company has $50 billion of assets under construction, which are still in bubble wrap, and that they are taking a more aggressive approach to driving better return on their existing investments. He also mentioned that they plan to leverage their assets under construction in the next year as well, and that they are aiming for a long-term capital intensity of 25%.

  • Analyst asked about the company's target of bringing everything in-house and the decision about internal versus external manufacturing for Nova Lake.

    The company's goal of bringing everything in-house, getting to a 70% mix remains steadfast, and there's no change to that POR. The company has optimized the process node at the SKU level for Nova Lake, and will build more wafers on Intel process than on Panther Lake. The company will continue to balance internal and external manufacturing of wafers to maintain a reasonable capital intensity relative to the business.

  • Analyst asked about the company's strategy for learning with what they've already had, particularly inside the data center.

    The company has spent time relooking at their AI strategy, portfolio of assets, and how to come to market to compete in the marketplace. They will start to talk externally about their AI roadmap and what they'll do in between then, and will dive in with customers on their workload roadmap to determine what needs to change. They know they need a robust portfolio to compete in the data center market, and will be aggressive in their approach.

  • Analyst asked about the viability of Intel's IDM structure, specifically the Foundry business, and whether it is sufficient to meet customer needs.

    LipBu Tan stated that the Foundry business needs to deliver products that meet the needs of both internal customers and Intel customers, while also maintaining simplicity and online schedules. He also mentioned that TSMC is a good partner and that Intel is working to collaborate with them to create a win-win situation. Additionally, Intel is working to drive efficiency and yield in the Intel 18A and 14A processes, and is looking to use its Foundry capacity footprint wisely.

  • Analyst asked about Intel's strategy for competing in the cloud infrastructure AI build-out, specifically in terms of rack scale networking and the pieces that Intel has internally.

    LipBu Tan stated that the rack scale approach is important and that Intel is working to develop full-scale hardware and software solutions to support it. He also mentioned that Intel's XPU, CPU, and GPU approach will be used to drive this development. Michelle Johnston Holthaus added that Intel is seeing great success with its IPU products and is expecting to double its revenue from 2024 to 2025. She also mentioned that Intel is the only Foundry with an optics-based Foundry option, and that customers are interested in building AI infrastructure with x86, which Intel already has a custom design win for.