- Vultr CMO Kevin Cochrane says data sovereignty is driving demand from enterprises
- The company is offering a playful $50,000 prize for the developer that builds an AI app that can drive a robot to deliver coffee to CMO Kevin Cochrane
- And the company is reportedly pursuing $1B in new funding
Data sovereignty is rewriting the rules for cloud infrastructure — and Vultr wants to be the provider enterprises turn to when they decide they need to control their most sensitive workloads.
Given the need for enterprises to lock down sensitive data, Vultr CMO Kevin Cochrane said he doesn't understand how enterprises can run sensitive AI workloads on OpenAI or another frontier AI model.
"I don't get it," Cochrane told Fierce at the Nvidia GTC conference in San Jose last week, arguing that doing so puts sensitive enterprise data at risk. "That's a crazy world," he said.
Vultr has been beating this drum for a while. Cochrane told Fierce early last year that sovereign cloud requirements are forcing enterprises beyond the major platforms, and in January he spoke about the rise of alternative hyperscalers like Vultr to meet enterprise needs. Vultr, of course, also has a vested interest in convincing enterprises to consider neoclouds as a viable — and safe — option for their workloads.
Still, Cochrane's take is interesting. He believes frontier model providers like OpenAI and Anthropic will end up serving consumers and personal productivity users. In contrast, enterprises with mission-critical workloads will deploy their own secure clusters, optimized storage and data pipelines — and they'll do it on infrastructure they control.
To help meet that expected demand, Vultr is reportedly pursuing at least $1 billion in additional funding as reported by The Information this week. Vultr did not respond to a request for comment from Fierce.
Rubin validates the full-stack thesis
Nvidia's Vera Rubin GPU platform announcement dominated GTC, but for Cochrane the headline was Nvidia's CPU. Vultr sees the introduction as a validation of Vultr's strategy, emphasizing the importance of the CPU for AI workloads.
"The CPU doesn't go away," Cochrane said. "The CPU becomes more important than ever, because the CPU is what orchestrates and saturates the GPU."
Nvidia's new Vera processor finally integrates a proper, enterprise-grade CPU into the Nvidia stack, unlike the earlier Grace chip, which Cochrane described as "a little helper thing." Vultr has already upgraded its default infrastructure stack to run Nvidia Dynamo for GPU memory management and utilization optimization, and the Rubin platform deepens that integration.
12,000 sign-ups and a robot barista
One fun tidbit from the conversation: Vultr is launching a nine month-long worldwide hackathon series based on NemoClaw, Nvidia's enterprise- and telco-hardened version of OpenClaw, the popular open-source agentic AI stack.
As part of the hackathon, Cochrane issued a coffee challenge — a $50,000 prize for a developer who can build an app that monitors his calendar, finds a 15-minute break, pre-orders a grande Americano from the nearest Starbucks, and dispatches a robot to collect the coffee so it arrives exactly on time.
Why does this matter? Because Nvidia CEO Jensen Huang pushed OpenClaw hard at GTC. He said the software, released in November, has already grown faster than Linux, Kubernetes or HTML. OpenClaw is "essentially the operating system of agentic computers," and every company needs an OpenClaw strategy, Huang said.
Some 12,000 developers signed up for NemoClaw on Vultr the first day after its launch less than two weeks ago, Cochrane said. In June, the company plans to debut a showcase highlighting applications built on the platform. We wonder if Cochrane's coffee-focused assistant will be one of them.
Full-stack enterprise AI integration is also the subject of an upcoming Fierce Network Research report and webinar, sponsored by Vultr. Sign up for the webinar and to receive a free copy of the report when it becomes available.