Multi-OEM Strategies & More Key to Infrastructure in AI Era

As demand for AI in its all forms continues to build, organizations are running into roadblocks. For enterprises, existing infrastructure might not be up to the task, leaving many to reconsider how their tech stacks need to be optimized for a new era of computing demand.

We spoke with Chief Revenue Officer Mike Watkinson to learn more about how Future Tech is meeting the modern demand and supporting its clients for the needs of today and tomorrow.

Supplying AI-related compute is the name of the game for providers, customers

By this point in the year, most in the tech industry are familiar with the GPUs and compute required to run extensive AI programs. For the enterprise and large federal clients Future Tech supports, that demand is scaling at a rapid pace, but the preparation to support that demand remains in the early stages for many.

“I think everybody knows that this is coming. Enterprise IT is adopting GPUs today at about 30% across enterprises and it’s going to be somewhere around 80-85% in the next two years,” Watkinson said. “It’s here, it’s coming, and how do we prepare for it?”

Preparation, for Watkinson, involves guiding clients through the various components of true “AI readiness” that can scale as their needs grow.

“So we like to acknowledge where all the problems are. There’s certainly the power and cooling problem, but we like to share where they’re gonna run into some challenges around networking, with InfiniBand versus Ethernet. And let’s talk about your accessibility or your skills around those types of technologies,” Watkinson said. “You know, it’s always gonna start with the workload. So what are you trying to do?”

From there, Watkinson says, the Future Tech team also walks their clients through understanding the downstream implications of changes in infrastructure. Those typically include concerns in networking and object storage, plus ensuring data governance and security components are in place as the organization looks to leverage their data in AI deployments.

“There’s a real opportunity rather than building in silos that are highly underutilized. Now we can build them in more consolidated clusters and drive utilization and drive have more value,” said Watkinson, though he noted this has historically been a challenge due to chargeback and operational constraints.

The multi-OEM strategy Future Tech utilizes for its enterprise and government customers

Watkinson says Future Tech has found success with bringing together best-fit solutions from multiple OEMs into a bundled offering that addresses the needs of their federal and enterprise customers. Watkinson estimates that five to six technology vendors are now standard for one solution across infrastructure, networking, and related security and other needs.

“It’s a whole different group of partnerships in some ways for our customers. They may have had a degree of leverage with one OEM, but now there’s multiple OEMs and that leverage gets diluted,” said Watkinson. “Whereas as a partner across a large ecosystem, we have that leverage, right? We have those partnerships in place and we’re able to leverage those on behalf of our customers and sort of bring that ecosystem to the conversation, keeping everyone in their swim lanes.”

“That’s new value for us, really. I guess we’ve always been doing that. But more traditionally, it’s, you know, pick an OEM, let’s do an RFI and compare these OEMs for this solution. That was our value,” he continued. “Now it’s bringing this ecosystem together and creating a true solution and an outcome versus a transaction.” 

How evolving tech is shifting the way businesses think about their infrastructure

Watkinson says many large companies are still in the early stages of understanding and preparing for this new era of infrastructure. As he and his team continue to walk through everything outlined in this article, Watkinson is also focused on helping clients see the bigger picture when it comes to cost.

“The next milestone is like, we’re building these things, these architectures in entirely different ways because they’re so large and so scalable. And the traditional CapEx funding for the next five years isn’t going to cut it,” Watkinson said. “And so you have to think more in terms of OpEx, you have to think more in terms of how we’re going to maybe charge back for consolidated environments across parts of the organization.”

Tech partners like Future Tech are equipped and well-positioned to guide enterprises and other organizations through these decisions and new deployments. To Watkinson, operationalizing technology is how everyone will find success.

”We as a partner think a lot about how we’re going to, on day two, support the technologies across the organization. Again, the word that I use is sort of operationalize that for the customers,” Watkinson said. “So I think that’s another milestone in this … it’s finding the patterns in terms of the architectures, but then really trying to figure out how we’re going to operationalize that, charge back, do consumption models differently on-prem, and drive it across a larger ecosystem of partners.”