Earnings Call Highlights
Intel's Q1 2026 call was notably bullish in tone, with management projecting a confident, almost triumphant posture after years of existential concern — Lip-Bu Tan explicitly contrasted the 'survival' narrative of a year ago with today's supply-constrained growth story. The dominant theme was the CPU's resurgence as the indispensable orchestration layer for AI inference and agentic workloads, a structural argument management made repeatedly and with conviction. Analysts were largely receptive, probing for competitive durability and TAM sizing rather than challenging the thesis, suggesting the sell-side is beginning to internalize the CPU-in-AI narrative. The call also surfaced meaningful foundry and advanced packaging signals, including the Terafab/SpaceX/xAI/Tesla partnership and accelerating 18A yield improvements, which management framed as inflection points rather than incremental progress. Overall sentiment was constructively positive, with the key tension being whether gross margin expansion can keep pace with the supply ramp.
- Management disclosed that the CPU-to-GPU deployment ratio in AI infrastructure is shifting materially — from roughly 1:8 in training environments toward 1:4 in inference, and potentially at parity or better in agentic and multi-agent workloads, a direct structural tailwind for Xeon demand that Intel says is already showing up in order patterns.
- Intel's ASIC business — purpose-built silicon optimized for specific customer workloads — is already running at a revenue run rate north of $1 billion annually, with ASIC revenue up more than 30% sequentially and nearly doubling year-over-year; Lip-Bu Tan called it a 'fast-growing' opportunity over the next five years given Intel's unique combination of CPU IP, advanced packaging, and process technology.
- Intel 18A yields are tracking ahead of internal projections and are on pace to hit year-end targets by mid-2026; advanced packaging backlog is growing with per-customer demand coming in at 'billions of dollars per year' rather than the hundreds of millions management initially expected, with revenue conversion beginning in 2027.
- The Terafab partnership with SpaceX, xAI, and Tesla was framed not as a standard foundry engagement but as a joint effort to 'refactor silicon process technology' and find unconventional ways to improve manufacturing economics — Lip-Bu declined to confirm whether it would be a typical foundry arrangement or a process licensing deal, signaling the relationship is still being structured.
- Intel's collective AI-driven businesses now represent 60% of total revenue and grew 40% year-over-year, with DCAI at $5.1 billion up 22% YoY; Xeon 6 was selected as the host CPU for NVIDIA's DGX Rubin NVL8 systems, reinforcing Intel's position as the default orchestration layer even in GPU-heavy AI infrastructure deployments.