OpenAI is reportedly developing a new chip, codenamed "Jalapeño," designed for inference tasks. This chip is expected to feature 128 individual chips and achieve 1.7 exaFLOPS of processing power, along with 27 TB of HBM memory. This development aims to give OpenAI a competitive edge against rivals like Nvidia and potentially accelerate their AI capabilities. AI
IMPACT This development could signal a significant advancement in AI inference hardware, potentially leading to faster and more efficient AI model deployment.
RANK_REASON The item discusses a specific hardware development by a major AI lab, but it is not a direct release announcement from the lab itself.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →