Tag: #process-node8 results
Intel's 14A node exceeds 18A in maturity, yield, and performance at equivalent stages. Tesla confirms 14A for Terafab AI chip project, praised by Musk. 0.5 PDK available, with 0.9 version imminent for design wins.
NVIDIA unveiled the Feynman GPU architecture roadmap at GTC, targeting a 2028 launch with custom HBM chips and 3D die stacking. The GPU core will pioneer TSMC's 1.6nm A16 process, but limited capacity forces mixing with 3nm nodes.
Intel released the Xeon 6+ processor family, codenamed Clearwater Forest, marking the data center debut of its 18A manufacturing process. With up to 288 high-efficiency cores, it prioritizes energy efficiency and system integration for cloud and telecom workloads over peak clock speeds.
Samsung plans to mass-produce 1nm chips by 2030 amid intensifying competition. Its foundry business shows turnaround signs post-2026, with 2nm GAA process yield reaching 60%, nearing the 70% target. This advances Samsung's semiconductor leadership.
Intel disclosed key data for 18A-P process node at VLSI 2026 symposium via paper T1.2. It delivers >9% performance gain at same power versus standard 18A. Alternatively, >18% power reduction at equivalent performance.
TSMC is advancing beyond 2nm with sub-1nm chip trial production targeted for 2029. Apple could benefit from these next-gen semiconductors in future devices. This signals continued scaling in chip technology.
Samsung delays 1.4nm commercialization, so Exynos 2800 will use 2nm GAA process. This chip is a milestone as Samsung's first mobile processor with a fully in-house GPU. Design expected to finish this year.
Analyst teardown of Panther Lake reveals Intel 18A process details, including 36nm M0 spacing for HP library and 76nm GAA gate spacing for logic. Power Via back-power delivery is limited to logic chips due to spacing constraints, absent in SRAM. Future 14A node expected to resolve SRAM limitations with BSCON tech.