您好,欢迎来到一带一路数据库!

全库
全文
  • 全文
  • 标题
  • 作者/机构
  • 关键词
  • 主题词
  • 摘要
高级检索
热词推荐: 能源基础设施

人工智能芯片在新技术加持下落地更多应用场景

文章摘要

2021年,传统冯·诺依曼架构的人工智能芯片算力继续提升,工艺普遍进入7纳米。神经形态架构、存算一体、芯片粒等新技术也加持人工智能芯片发展,陆续有产品问世。英伟达仍占据着人工智能云端训练芯片市场的霸主地位,初创企业努力寻求更多落地空间。人工智能芯片企业仍获资本持续关注,投资总金额刷新纪录。人工智能芯片下一步发展重点仍将在技术、应用、生态、融资四个方面,有明确市场和稳固生态的人工智能芯片,将继续获得资本的加持,推进技术发展和企业壮大。未来,多种技术将持续推进,并在通用性和专用性间寻求平衡,获得最大限度的发展。国内人工智能芯片企业将在巨大的市场需求中获得更多发展机会。

Abstract

In 2021,the arithmetic power of traditional von Neumann architecture AI chips will continue to increase and the process generally enters 7nm. New technologies such as neuromorphic architectures,storage and computation integration,and chiplets are also supporting the development of AI chips,with products coming out. Nvidia continues to dominate the AI cloud training chip market,with startups struggling to find more room for implementation. Artificial intelligence chip companies continue to receive capital attention,with the total amount of investment setting new records. The next development of AI chips will still focus on four aspects:technology,application,ecology and financing. AI chips with a clear market and a solid ecology will continue to receive capital boosts to advance technology development and enterprise growth. In the future,a variety of technologies will continue to advance and seek a balance between generality and speciality to gain maximum development. Domestic AI chip companies will have more development opportunities in the huge market demand.

作者简介
张瑶:张瑶,国家工业信息安全发展研究中心工程师,主要跟踪国内外人工智能、智能语音、计算机视觉等多个领域的产业发展动向,在人工智能产业发展、政策规划及安全、伦理、就业相关领域具有丰富的研究经验。