Q&A with Isaac Douglas from servers.com: Industry Insights and Trends in Hosting Infrastructure

In a market dominated by AI hype, decisions around infrastructure are becoming more versatile. Businesses that want performance, cost-efficiency, and control can’t ignore the shift toward hybrid models and flexible server strategies.

Isaac Douglas, Chief Revenue Officer at servers.com, has seen this evolution firsthand. With more than 15 years in the industry, he spoke with us and offered a grounded take on why bare metal is making a comeback, how to stay agile amid shifting hardware trends, and why infrastructure providers should focus on doing one thing well rather than chasing every layer of the stack.

This transcript has been edited for brevity. Here’s Isaac.

A: It’s a great question and one I’d love to answer because I’m a firm believer in the benefits of hybrid infrastructure. It’s been a big component of what I’ve believed in for computing in the cloud for a long time.

It’s what we built back when I was at Multiplay, which feels like many, many years ago now.

But it’s also what we tell our customers here: I don’t think anyone should put 100% of their infrastructure with one supplier. That includes us — which sounds a bit counterintuitive for a CRO to say, I know — but I genuinely believe it’s what’s best for clients and for their businesses.

Quite selfishly, if it’s good for their business, it’s probably good for us too. If they’re successful, they’ll need to consume more compute, and that’s good news for us. So it’s a thesis rooted in what I believe is right. And if it works, we benefit as well.

The way we do it today is this: Customers can come to us and purchase a long-term, fixed, stable portion of their infrastructure. They commit to that hardware over a longer period, which is one of the most cost-effective ways to consume cloud in an OPEX model.

“I don’t think anyone should put 100% of their infrastructure with one supplier.”

Then, on top of that, they can add a flexible layer — bare metal that’s rented by the hour or month — so they get that next level of agility. And then, we’re very open about this: we actually recommend using hyperscale cloud providers too. They’ve got a brilliant product when it comes to spinning up large amounts of compute quickly.

Especially if you’re running in containers, it’s a very straightforward path to scaling — up or down — depending on your needs. It works great for predictable, short-term spikes. For the really bursty events — like Black Friday, or if you’re a betting site, maybe Super Bowl weekend — that’s exactly the kind of use case where public cloud makes perfect sense.

A: I think there’s always got to be a level of agility. If you look back 10 years in the hosting market — especially in the bare metal space — Intel was ruling the roost when it came to CPUs. And I’m sure Team Blue will have their day in the sun again. But AMD has been doing a fantastic job for a while now, and we’ve got to be ready for that.

If a vendor’s falling behind compared to someone else’s offering, you’ve got to be prepared to swivel on a dime. That’s not to say you throw those relationships away — you’ve still got to nurture them — but at the end of the day, the market wants what it wants: the most compute for the least amount of dollars, within the right margins for uptime and performance.

“If a vendor’s falling behind compared to someone else’s offering, you’ve got to be prepared to swivel on a dime.”

So it’s about keeping on top of that changing landscape and understanding how people want to consume. Servers haven’t really changed in 20 years — they’re still computers plugged into power, cooling, and the internet. But what’s inside is evolving: density, GPU configurations, maybe multiple GPUs if you’re doing GPU-powered workloads. The core idea is the same, just different nuts and bolts.

And you’ve got to be able to read that market. You can’t just build a product in a silo and then change it in a silo. You’ve got to speak to your old customers and your new ones.

“You can’t just build a product in a silo and then change it in a silo. You’ve got to speak to your old customers and your new ones.”

For us, we focus on six key industries: video gaming, iGaming, fintech, adtech, streaming, and SaaS. We really hone in on those and take customer feedback seriously. We build and adapt based on that. It’s got to be a loop.

A: Internally, yes. We’re definitely using LLMs to some extent. We have a very security-focused operation here. Our data centers hold all our customers’ important information. So we don’t want to be exporting any of that data into someone else’s cloud. Not for our sake, and definitely not for theirs.

So we’ve developed some LLM tools in-house, mainly to support our development teams and engineers. But from a customer-facing perspective, that’s not our game.

“But we’re not going to sell you an AI-as-a-service product. That’s just not the layer of the stack we want to be in. We want to be masters of the infrastructure piece, not the platform.”

We’re here to solve the infrastructure problem — so you can go do the sexy bit on top. And if that sexy bit happens to be AI, or integrating a language model into your app, then great: We’ll help you find the right infrastructure solution to power it.

But we’re not going to sell you an AI-as-a-service product. That’s just not the layer of the stack we want to be in. We want to be masters of the infrastructure piece, not the platform.

Isaac: Well, obviously, I want to say servers.com is the best of the best. But really, I think it’s an interesting time in the hosting world.

I’ve been in the space for about 15 years now. LinkedIn probably knows better than I do. It feels like a long time. I remember when AWS first came out — it was revolutionary. And now they’ve kind of become what IBM used to be: “No one ever got fired for buying IBM.” Today, no one gets fired for choosing AWS.

But we’re starting to see a new wave of buyers, new decision-makers, who don’t even know that something like bare metal exists. Or if they do, they think of it like it was 10 years ago — slow provisioning, long contracts, all that.

“You don’t need to put all your eggs in the hyperscale basket.”

There’s another path. You don’t have to go straight to AWS, GCP, or Azure — and those are great products, no doubt. But you don’t need to put all your eggs in the hyperscale basket.

You can get high performance, real support, and very competitive pricing with a mix of infrastructure. That’s what I want more people to understand.

Isaac Douglas is the Chief Revenue Officer at servers.com and has over 15 years of experience helping businesses build, deploy, and select infrastructure solutions that drive growth and deliver long-term value.