Products face multi-front competition
Intel's product portfolio is under siege from AMD (CPUs), ARM (client), NVIDIA (AI accelerators), and Broadcom (networking). The question is whether Panther Lake on 18A can stem the share losses.
CCG under dual pressure
Intel's client business faces both AMD Ryzen (taking desktop share to 42.6%) and ARM-based processors (Qualcomm Snapdragon X threatening the Wintel paradigm). Panther Lake on 18A is Intel's best response in years.
AMD EPYC share gains are structural, not cyclical
AMD's server CPU share has grown from near-zero to 39%+ over 5 years, driven by genuine product superiority (more cores, better efficiency). Intel's recovery depends on 18A-based Xeon recapturing performance leadership.
Gaudi is technically competitive but commercially irrelevant
Gaudi 3 beats H100 on price-performance for inference, but ~$500M in annual revenue vs NVIDIA's $250B+ run rate makes it a rounding error. CUDA ecosystem lock-in is the real barrier, not hardware specs.
NEX retained after spin-off review
Lip-Bu Tan reversed the NEX spin-off plan in late 2025, concluding that networking capabilities complement Intel's AI data center strategy. The segment is no longer reported separately.
Altera: $8B in destroyed value
Intel acquired Altera for $16.7B and sold the majority stake for $8.75B — a textbook case of semiconductor M&A value destruction. The FPGA-for-data-center thesis collapsed as GPUs captured the AI acceleration market.