Verizon Plans to Back Generative AI Connectivity, Compute DemandVerizon Plans to Back Generative AI Connectivity, Compute Demand

Verizon just unveiled its strategy for tapping into the exploding demand for network capacity and computational power brought about by the growth of generative AI.

Dubbed Verizon AI Connect, the initiative consists of a suite of services that would tap into the telecom giant’s existing assets, including network connectivity, edge computing and space, which will be leveraged to satisfy the AI needs of hyperscalers, cloud providers and global enterprises. 

“Verizon's leadership in the enterprise-connectivity market is poised to become even more crucial as AI applications become more widespread among end users,” Verizon Business CEO Kyle Malady said on a fourth-quarter earnings call, during which the company unveiled the strategy. “We believe Verizon will serve as the essential link between these GenAI applications and the firms who utilize them.” 

In a separate announcement, Verizon noted that while current demand centers around training AI models, in the next five years, AI workloads are expected to migrate to real-time decision-making, or inference, creating an urgent need for low-latency connectivity, compute and security. 

“As AI shifts from training to deep deployment, the need for distributed computing will become increasingly important for real-time decisions and predictions,” Malady said. “We have thousands of distributed telco facilities, many of which already have power, space and cooling available for this compute at the edge.”

Related:Verizon, Nvidia Team to Power AI Workloads on 5G Private Networks

In addition to the existing telecom facilities, Verizon controls up to 200 acres of land, much of it already zoned for data centers, which can be developed as demand increases.

Malady noted Verizon’s geographically wide fiber infrastructure, which it has built up over the last 20 years and comprises its 71-city One Fiber buildout and long-haul network connecting the metros.

Verizon’s edge-to-cloud connectivity through its Converged Intelligent Edge Network would satisfy customer demand for increased bandwidth and network visibility, Malady said, adding that the company’s network is increasingly intelligent and programmable, features that are going to be in high demand going forward.

“Customers want to control where and when AI workloads are running,” he said. “And to that, we are building new capabilities to give users more insight and control via new tools. Essentially, we are allowing them to program their network resources to optimize their operations as they see fit.”

Verizon noted its recent partnership with Nvidia to integrate graphics processing unit (GPU)-based edge platforms with its 5G private networks and said Google and Meta have purchased network capacity to handle increased AI-driven demands.

Finally, Verizon announced a new strategic partnership with Vultr, a GPU-as-a-service and cloud-computing provider. Initially, Vultr will deploy its GPU-as-a-service infrastructure in one of Verizon's data centers and use its fiber network for distribution. Looking ahead, Verizon expects Vultr to expand its reach and eventually provide mutual customers with AI training and inference capabilities.

“GPUs play a crucial role in powering AI workloads, and rapidly growing demand for real-time inference is driving the need for AI infrastructure at the edge,” Vultr CEO J.J. Kardwell said. “With demand for data centers and GPU processing power outpacing supply, Verizon’s connectivity infrastructure is uniquely positioned to support our growth. Through Verizon AI Connect, we can extend our global cloud footprint and bring cutting-edge AI solutions to Verizon Business’ global customers.”