Enterprise AI strategies extend beyond model selection and experimentation. Bringing AI into production requires infrastructure, deployment processes and operational frameworks that support applications after development ends. As organizations evaluate where AI can create business value, attention often shifts toward the systems and workflows that allow new capabilities to move from concept to implementation.
Open-source technologies, agentic AI and platform engineering practices continue to influence how enterprises build and deploy AI applications. Rather than assembling environments from scratch for every project, technology teams are exploring ways to create consistent foundations that support development, deployment and ongoing operations. This approach can help reduce complexity, support work across technical and business teams and provide a more repeatable path for bringing new AI services into production.
AI readiness reaches beyond the technology itself. Data, infrastructure, deployment requirements and business needs all contribute to production readiness. Preparing for production means creating environments that support scalability, operational consistency and broader deployment as adoption expands. For many organizations, the focus has broadened from what AI can do to how AI can be deployed, managed and supported across the enterprise.
To learn more about Vultr, visit https://www.vultr.com.
Steve Saunders:
There is an awful lot of noise around cloud and AI right now. What's actually happening in the market?
Kevin Cochrane:
The market has definitely shifted, and enterprises needed to get serious about agentic AI specifically. This was the NemoClaw launch. And secondly, tied to that is the market now moved from building frontier models to taking advantage of open source. What we see here at Vultr is a massive acceleration in the market, where the buying interest is moving from small startups that are test marketing and pioneering new AI services to actual enterprise buyers looking to understand how they could build and deploy real systems. But it's finally crystallized post-GTC.
Steve Saunders:
I get the impression that there are a lot of organizations that are really struggling operationally with AI deployment. Do you agree with that?
Kevin Cochrane:
Honestly, we've been here before, right? With the adoption of cloud, IT organizations struggled on how to retool and re-skill so they can start building and deploying workloads in new ways, right? IOCS teams need to be starting to get very serious about putting in place a platform engineering team and platform engineering practices so that their platform engineering teams can enable front-end developers and business analysts to work with the business to build and deploy new AI-driven solutions. So that's what we're specializing here at Vultr, which is helping platform engineering team take all of the skills of all of the subset teams within an IOCS team, and then inculcate that into our platform so you can have a rapid rollout, on-demand, worldwide, with all of the pre-built infrastructure stacks that you need for any AI-native workflow.
Steve Saunders:
Kevin, how does that work in practice though?
Kevin Cochrane:
Part of it actually gets back to GTC and NemoClaw. Really, what your platform engineering team wants to do is it basically wants to get from each of your virtualization engineers, each of your storage engineers, your network engineers, they basically want to model the skills and pre-build the artifacts, i.e. the templates that can be pulled forward into what we call a composable stack. So that means pre-building all of your containers, pre-building all of your CI/CD pipelines, pre-building all of your data pipelines, pre-building all of your Helm charts, pre-building all of your Terraform templates so that any front-end developer, through the platform engineering team, can simply spin up their own NemoClaw and simply say, "Spin me up a stack that will enable me to build a new commerce experience for my channel partners." And then, all the developer has to do is then just go sit with the business and say, "Okay. Well, I'm ready to go. I have everything I need. What are your requirements? What specifically do you want this new commerce experience to look like?" And everything else is already prebuilt. Everything else is ready to go.
Steve Saunders:
You're building the scaffold that people can hang their business off, and they essentially come in and get to do the aesthetics on the front of it.
Kevin Cochrane:
All of that infrastructure complexity, which is necessary because we're going through rapid innovation and we're building lots of net-new stuff, right? That's awesome. But it does leave a gap, which has been the gap for enterprise adoption, right? So we've got to close that gap. And closing that gap means enabling platform engineers to build these pre-composed stacks, right? So what we're doing here at Vultr is we're building this marketplace, where anyone can take this composable stack and then they can share it with others.
Steve Saunders:
There is no single company, there isn't even any single defined ecosystem that can do that. You need an army of 300,000 people to do that globally, and that's what you're doing.
Kevin Cochrane:
100%. And we're not doing it alone. We're doing it with partners. So for example, we're a member of the Cloud Data Compute Foundation. We want to bring together everyone in the community, and this is the power of open source communities. The gap between what cloud infrastructure does and the business needs has never been wider. Vultr closes the gap.