Taalas

Google is developing a new server chip informally called 'Frozen v2' that will hardcode the architecture of its Gemini AI model directly into silicon, aiming to achieve 6 to 10 times greater efficienc…

Google is developing a custom chip codenamed 'Frozen v2' that hardwires the architecture of its Gemini AI model directly into silicon, aiming for 6 to 10 times greater efficiency than its latest TPUs,…

Nvidia dominates AI training hardware, but the inference market is wide open as companies like Cerebras, Groq, d-Matrix, Etched, and Taalas challenge Nvidia's GPU-based approach by designing chips opt…

Tensordyne is developing AI chips that use a logarithmic number system to multiply numbers more efficiently than conventional floating-point math, aiming to improve compute efficiency for AI workloads…