Intel today unveiled its Xeon 600 series, featuring up to 60 cores per socket, eight-channel DDR5 memory support at 4800 MT/s and 96 PCI-Express 5.0 lanes. The new chips include dedicated AI acceleration units that boost inference throughput by approximately 50% compared with the prior generation, while integrated networking enhancements deliver up to 2× the aggregate bandwidth for high-performance computing clusters. Intel expects the Xeon 600 line to address demand in high-end workstations and AI training rigs, targeting revenue growth of up to 15% in the data-center segment for fiscal year 2026.

Intel’s share price rose by nearly 26% in January after independent benchmark reports praised the Panther Lake client CPU, the company’s first product built on its new 18A process node. Test results demonstrated a 20% improvement in single-thread performance, a 30% uplift in battery life on mobile platforms and a 40% gain in integrated graphics throughput versus the prior generation. Despite lowering first-quarter sales guidance due to capacity shifts converting fabs to server-chip production, management cited strong preorders for both Panther Lake and the forthcoming Granite Rapids server processor line.

In the fourth quarter, Intel reported revenue of 13.67 billion dollars, surpassing consensus estimates by roughly 2%. Adjusted earnings reached 15 cents per share, well above the consensus of eight cents, even as total revenue declined 4% year-over-year. Growth in the network-and-edge business offset lower client PC volumes, and gross margin expanded by 120 basis points as advanced-node yields improved sequentially.

Looking into 2026 and 2027, Intel expects synchronized ramp-up of its 18A process volume alongside robust commitments for its 14A node from major OEMs. The company forecasts that leading-edge process technologies will contribute over 30% of foundry revenue by the end of 2027, marking a fundamental shift in its manufacturing credibility after nearly a decade of delays.