Nvidia backs Upscale AI's heterogeneous compute networking push
Upscale AI, a startup with Nvidia as a strategic investor and technology supplier, has unveiled Token Fabric, a full-stack AI networking platform meant to let data centers run accelerators from Nvidia, AMD, Intel and Chinese vendors side by side. Its in-rack scale-up layer uses Upscale's own SkyFabriX switch chip and supports the UALink and ESUN open interconnect standards, while the cross-cluster scale-out backbone is built on Nvidia's Spectrum-X Ethernet, paired with the SkyOS network OS and SkyCMD operations tooling. The report says phase one is slated for Q4 2026 with the full platform rolling out in stages during 2027, and that Upscale AI raised a $190 million extension in June 2026 led by Premji Invest with Nvidia, Temasek and Salesforce participating, bringing total funding to $500 million at a $2 billion valuation.
If the approach holds, Nvidia could keep control of the cluster's outbound network even as customers mix chips, making interconnect a standalone standards and business battleground.
过去两年，AI产业的关键词几乎被算力垄断：GPU、HBM、先进封装、超节点。但越往前走，越多人意识到，算力不是孤岛。芯片再强，如果连不起来，也只是一座座昂贵的孤岛。
当大模型从千亿参数走向万亿参数，训练集群从千卡走向万卡、十万卡，网络互联技术正从后台配套，变成决定AI集群能跑多快、能扩多大的核心变量。
就在这个节点，一笔看似矛盾的投资引发行业关注：近日，一家叫Upscale AI的公司发布了一套名为Token Fabric的全栈AI网络平台，用途是让数据中心里同时运行不同供应商的AI芯片。
这件事本身不算稀奇，稀奇的是它的投资方名单中，英伟达赫然在列。一家靠整套自有互连把加速卡绑在一起卖的公司，掏钱支持了帮客户混用别家芯片的生意。
英伟达为何投资“挖自己墙角”的公司？
众所周知，行业所有人几乎都默认英伟达靠封闭NVLink+全栈绑定垄断AI集群，绝不允许多芯片混用生态。但Token Fabric可兼容英伟达、AMD、英特尔、国产AI芯片等不同厂商的AI加速器，而英伟达是其核心战略投资方与技术供给方。
#Nvidia#Upscale AI#Token Fabric#Spectrum-X#SkyFabriX#UALink#ESUN#Mellanox#NVLink Fusion#AMD#Marvell#Premji Invest