At JPMorgan’s annual conference, Intel CEO Chen Liwu disclosed that the Intel 18A process technology is now ready for high-volume manufacturing of Panther Lake processors, with yields climbing by roughly 7% each month—surpassing internal projections. Concurrently, the 0.5 PDK (Process Design Kit) for Intel 14A has been made available, while the 0.9 PDK is slated for release in October. The team is already laying the groundwork for even more advanced nodes, including 10A and 7A.

As AI workloads pivot from training-centric tasks to inference-driven operations, system configurations are shifting dramatically. The traditional CPU-to-GPU ratio of 1:8 is being rebalanced toward 1:1, with some scenarios even favoring a 4:1 CPU-to-GPU setup. To meet evolving demands, Intel is aggressively scaling its ASIC (Application-Specific Integrated Circuit) business, offering tailored hardware solutions for specialized AI applications.