XPower Technology has introduced its RPP architecture AE7100E chips, designed to challenge the dominance of large-scale chips in AI inference. These small, fingernail-sized chips can be deployed in groups, utilizing distributed computing to support models with up to 1.6 trillion parameters. The company claims this approach offers a lower token cost compared to traditional large-chip solutions. AI
IMPACT Offers a potential alternative for AI inference hardware, focusing on distributed computing for large models.
RANK_REASON Product launch of a specialized chip architecture.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →