Equinix, whose stock is up 33% YTD, pushing its market cap to $100B, inks a deal with Nvidia to let customers run their AI models on the Together AI platform
Long before the era of the hyperscalers and neoclouds, and decades ahead of the AI data center boom, there was Equinix.
Context & Ripple Effects
Nvidia’s data-center GPU leadership was already central to the AI infrastructure buildout, while operators such as Applied Digital were pairing leased data-center capacity with Nvidia-chip cloud services in a capacity-and-Nvidia compute model.
Together AI entered 2026 in reported fundraising talks after rapid revenue growth, according to The Information’s sources; Equinix gives that platform a route into its enterprise customer base rather than leaving AI compute distribution to cloud specialists alone.
First-order effects
- Equinix customers gain a path to run AI models on Together AI using Nvidia infrastructure, extending Equinix from colocation and connectivity into the inference-service layer.
- Together AI gains enterprise distribution through Equinix, while Nvidia adds another channel for workloads built around its AI compute platform.
Second-order effects
- Data-center operators competing for enterprise AI workloads face pressure to bundle compute access, model platforms and network connectivity rather than sell space and interconnection as separate services.
- Together AI’s platform becomes a more direct alternative for enterprises weighing specialized AI infrastructure providers against hyperscaler-led AI services.
Third-order effects
- If comparable partnerships proliferate, AI infrastructure will be organized less around standalone facilities and more around integrated exchanges that combine capacity, accelerators, networking and model software.
- Nvidia’s position in the stack gives infrastructure providers an incentive to differentiate through distribution and service integration, even as they depend on the same underlying compute ecosystem.
The trend: Enterprise AI infrastructure is shifting toward platformized offerings that package Nvidia compute, model access and data-center connectivity into a single service path.