Intel (INTC) Deutsche Bank 2026 Technology Conference summary

Capital allocation and investment plans

• Raised $23 billion in equity to fund significant CapEx increases in 2025 and 2026, targeting advanced process nodes and substrate investments to meet rising demand.

• CapEx will support ramping Intel 18A and 14A nodes, with high-volume manufacturing for 14A committed for 2028 and risk production starting in 2027.

• Investments are also directed at expanding clean room and fab capacity in Ireland, Arizona, Oregon, and Ohio to address supply constraints and future growth.

• Advanced packaging (EMIB-T) is expected to ramp revenue in late 2027, become a run-rate business in 2028, and reach billions in revenue by 2029 with strong margins.

• CapEx efficiency is a focus, with ongoing efforts to optimize spending and drive operational improvements before finalizing next year’s budget.

Technology and product roadmap

• Intel 18A and 14A nodes are progressing ahead of internal milestones, with defect density and yields outperforming previous generations.

• Internal and external customer demand for 14A is strong, with internal teams now designing products on the node, boosting confidence in execution.

• Advanced packaging technology (EMIB-T) is highly differentiated, enabling larger reticle sizes and supporting AI needs, with margins projected at 40% gross and 30% operating.

• Data center CPU demand is surging due to AI-driven workloads, with double-digit unit growth expected through 2025 and beyond.

• Client product roadmap includes Panther Lake and Nova Lake, with the latter expected to address broader market needs and improve competitive positioning.

Market dynamics and demand trends

• Data center demand is shifting toward higher CPU density, driven by AI inference and agentic workloads, resulting in both unit and ASP per core growth.

• Supply remains constrained through at least 2025, with manufacturing expansion in Ireland and Arizona aimed at doubling output.

• Client demand is intentionally deprioritized in favor of data center, reallocating wafer supply to maximize growth opportunities.

• Edge and physical AI markets are expected to grow significantly, potentially matching client segment size, with new leadership focused on system-level solutions.

• Memory remains a bottleneck for AI workloads, with ongoing collaboration with major memory suppliers and product architecture adjustments to address challenges.