英伟达豪掷超65亿美元押注光子技术，加码AI基础设施「数据传输瓶颈」
BlockBeats 消息，5 月 30 日，英伟达自今年 3 月以来已累计承诺至少 65 亿美元投资光子学与光网络相关企业，以解决 AI 数据中心面临的数据传输和能耗瓶颈问题。
据报道，英伟达分别向 Lumentum、Coherent 和 Marvell 投资 20 亿美元，并向康宁投入 5 亿美元用于光连接系统研发，同时参与光学初创公司 Ayar Labs 5 亿美元 E 轮融资。
光子技术利用光信号替代传统铜线电信号传输数据，可显著提升带宽并降低能耗，被视为下一代 AI 数据中心核心基础设施。随着 AI 模型规模持续扩大，数千块 GPU 之间的数据交换需求激增，网络互连正成为继算力芯片之后的新瓶颈。
英伟达 CEO 黄仁勋今年 3 月在 GTC 大会上表示，公司正加速将硅光子技术应用于网络平台和 GPU 互连系统，相关产能需求已远超当前全球供应能力。
分析人士指出，AI 机柜系统对高速光互连需求正快速增长，光模块、激光器、光纤等产业链有望持续受益。受英伟达投资推动，Lumentum 年内股价已上涨逾 130%，Coherent 和 Marvell 股价也显著走强。
---------------------------------
点击下方原文链接，加入动察 Beating · 飞书 AI 新闻渠道，7×24 小时不间断监测全球 AI 热点与新闻。
Disclaimer: OKX Orbit content is provided for informational purposes only. Learn more
Replies
Related Flash News
Hyperliquid's crude oil contract trading volume reached nearly 30% of its benchmark futures
The NVIDIA Cosmos 3 Physical AI World Model is officially available for download today, with two versions now available on HuggingFace
NVIDIA has secured Anthropic's first batch of users to expand its AI data center layout
The AI investment boom has elevated Serenity to the pedestal, with X platform subscribers only 4,000 shy of Musk
NVIDIA launched the Vera CPU designed specifically for the era of intelligent agents
"New Stock God" Serenity reviews ARM's big win: from $134 to $354, the $15 billion revenue target is now fully reasonable
NVIDIA CEO Jensen Huang has released the new AI model, Nemotron 3 Ultra
Intel plans to launch new AI chips by the end of the year
Loeb refutes the bubble theory: investing in AI is definitely not about flushing toilets with money; we have only just touched the surface of the hype
Zhiyuan Robotics: The 2-billion-parameter model GE 2.0 topped the international review rankings
