Q4 2023 Earnings Summary
- Intel expects stabilizing market share and year-on-year growth in their Data Center business, driven by the ramp-up of Gen 4, Gen 5, Granite Rapids, and Sierra Forest products, and is well-positioned to benefit from AI inferencing on-premises, leveraging its enterprise strength.
- Intel's client business is anticipated to grow, with stable market share and improved product lines, including the introduction of AI PCs (AIPCs), expected to bring a multi-year cycle of growth. The AIPC launch is described as the most exciting category-defining moment since Wi-Fi was introduced.
- Intel's internal foundry model is expected to drive $4–5 billion in savings, improving gross margins towards the 60% goal, and aiding in regaining process leadership with the Intel 18A node. Customers and IP providers are affirming the competitiveness of their process technology, particularly in backside power enhancements.
- Intel anticipates significant start-up costs in 2024 due to their aggressive '5 nodes in 4 years' plan, which will be a huge headwind on margins.
- The company is guiding to the low end of seasonality in Q1 for core businesses, suggesting weaker demand than expected despite abating cyclical pressures.
- Persistent capacity utilization challenges are expected until new products ramp up, potentially not until 2026, which may impact profitability.
-
Gross Margin Outlook
Q: Will gross margins improve year-over-year?
A: Management expects a 60% fall-through on revenue growth, improving gross margins from last year's 43.6%. While Q1 starts at 44.5%, they anticipate margins to improve throughout the year despite quarter-to-quarter volatility. Factors include revenue growth, better factory utilization, achieving leadership in nodes and products, and savings from the internal foundry model, which is expected to deliver $4–5 billion in savings. -
Data Center Business Outlook
Q: What's the outlook for the data center business in Q1 and beyond?
A: Excluding FPGA, the data center decline in Q1 is seasonal. Management sees strength from server customers, particularly in enterprise. They expect year-on-year growth, with market share stabilizing and momentum building from ramping Gen 4, Gen 5, Granite Rapids, and Sierra Forest. They anticipate the AI surge will result in more on-premise AI inferencing, benefiting Intel. -
Manufacturing Node Leadership
Q: How confident are you in achieving leadership with 18A nodes?
A: Intel is confident in the progress of 18A, with Clearwater Forest taping out, indicating health of the node. They are ahead on backside power, offering substantial gains to customers. While they use external foundries to complement and manage capital, they are focused on maximizing internal capacity as part of their Smart Capital strategy. -
AI Market Impact
Q: How will AI investments affect your CPU business growth?
A: While high-end cloud customers focus on AI training with accelerators, Intel expects the market to shift towards AI inferencing, where their products are more substantial. The enterprise market will deploy AI on-premise, leveraging Intel's strengths. They need to participate more in accelerators, with Gaudi3 gaining excitement. The message for 2024 is "inferencing AI everywhere," areas where Intel has a stronger footprint. -
Client Market Outlook
Q: What's driving confidence in client market growth?
A: Intel expects the client market to be a few points larger than last year's 270 million units, with stable market share and a better product line. Tailwinds like AIPC ramping, Windows 10 EOS, and strong positions in gaming and commercial segments contribute to confidence. Excitement around AIPC is high, marking the most significant category-defining moment since WiFi's introduction. -
Foundry Business Prospects
Q: How significant is your foundry business, and what's expected ahead?
A: The decline in IFS from Q4 to Q1 is due to the natural ending of traditional packaging volume and equipment business fluctuations. Intel is focusing on advanced packaging and has secured four 18A customers. Lifetime deal value is a key metric, representing external foundry business growth. Internal business will drive factory build-out, and more updates will be provided at the IFS Direct Connect event on February 21. -
Savings from Internal Foundry Model
Q: How will the internal foundry model contribute to savings?
A: The internal foundry model is expected to deliver $4–5 billion in savings. Teams are focusing on test times, sample activity, loadings, and capital investments. This new approach is revealing significant savings opportunities and will help drive towards 60% gross margins. -
Sierra Forest and Granite Rapids Mix
Q: What is the long-term mix between Sierra Forest and Granite Rapids?
A: Both products are on Intel 3. Sierra Forest targets cloud providers needing high core counts for bulk workloads, focusing on TCO. Granite Rapids offers performance and features. Most enterprise customers are expected to stay with P-core products like Granite Rapids, but there's a steady rise anticipated for Sierra Forest in the cloud segment. -
Manufacturing Technology Challenges
Q: How are you managing technological challenges in 18A development?
A: Intel is carefully managing risks by introducing EUV with Intel 4 and 3, thereby de-risking it before 18A. They ran an internal node using Intel 3 with backside power to de-risk backside power before 18A, which combines backside power and gate-all-around transistors. Backside power on 18A has been yielding well, with focus now on the new transistor structure. High NA EUV is not part of 18A, avoiding additional risk. Overall, they are confident in delivering 5 nodes in 4 years.
Research analysts covering INTEL.