‘Red Hat’s biggest opportunity is with our channel, being able to make this safe, this (AI) journey,’ says Red Hat CEO Matt Hicks.
Red Hat CEO Matt Hicks sees solution providers as key to helping customers overcome the technological and change management humps to unlocking enterprise-grade open source and artificial intelligence platforms for customers.
“Red Hat’s biggest opportunity is with our channel, being able to make this safe, this (AI) journey,” Hicks told CRN in an interview. “If you’re a CEO, it’s the biggest bet and choice you’re going to make. And if you don’t know these things yourself, you have to have someone you can trust that can help put you on that journey.”
The CEO of the Raleigh, N.C.-based open-source enterprise tools vendor and IBM subsidiary caught up with CRN during the annual Red Hat Summit conference, coming a week after its parent’s IBM Think 2026 event. Red Hat Summit 2026 runs through Thursday in Atlanta.
[RELATED: IBM Think 2026 Showcases Agentic AI And Sovereign Cloud Strategy]
Red Hat CEO Matt Hicks On Partners
Bo Gebbie, president of Hamel, Minn.-based IBM and Red Hat solution provider Evolving Solutions—No. 162 on CRN’s 2025 Solution Provider 500—told CRN in an interview that Red Hat’s virtualization portfolio has been an enticing alternative for VMware users that do not like the company’s pricing changes.
Evolving Solutions has also seen customer demand for Red Hat’s automation products as a way to keep costs in check and remove human-influenced risk and security concerns, he said.
“As you think about people cost increasing, you think about AI workloads–the automation side is needed for all of that,” Gebbie said.
IBM and various divisions of its portfolio, including Confluent and Red Hat, have been investing in their partner programs to seize opportunities in AI, data streaming and hybrid cloud.
Red Hat, for example, is looking to add core sales training, core technical training, sales tools and other capabilities throughout the year to help partners with its AI portfolio, according to CRN’s 2026 Partner Program Guide.
During Red Hat Summit 2026, the vendor revealed a host of new product innovations that better arm solution providers around customer engagements ranging from AI to automation, including the upcoming Red Hat AI 3.4, closer integrations with chipmaker giant Nvidia and the upcoming Ansible Automation Platform 2.7.
Here’s more of what Hicks had to say about opportunities ahead for Red Hat solution providers.
What’s your message to Red Hat solution providers coming out of Summit 2026?
There is a ton of value for partners to add. (In software currency), whether it’s with Mythos (Anthropic’s unreleased frontier AI model that can autonomously find and exploit complex vulnerabilities) or other areas, I think the dynamic of the risk of running out-of-date software has really tipped the scales to where vulnerabilities can be found now.
The basics of just RHEL (Red Hat Enterprise Linux) currency, patching through Ansible (Red Hat’s automation platform) for your environments, the value of that just went up tremendously.
A lot of companies, they are not going to have, necessarily, the skills to be able to (have) awesome system integrator service capabilities.
(For MSPs), whether it is getting to RHEL 9 or 10 for quantum-safe encryption, that may be the next challenge. Or just staying on current versions or layers of security to avoid AI vulnerabilities, a ton of opportunity in that area.
For customers, it is a threat and risk. But as partners–we have all the tools to mitigate this risk and help customers get there. (That) can be transformative.
(In virtualization), you get this flywheel effect of–you are able to move to a consolidated platform. You run more efficiently. Your costs are lower. And you modernize.
That builds on that currency thing–if I have to make a change, and I can make that change to get in a better position, that’s great.
(With partners), you have so many different integrated options, both in the current state, as well as where you want to get to from storage vendors, backup, resiliency, VDI (virtual desktop infrastructure), all of the software you’re running on.
Partners, they provide the reach that get customers over this hump. We’ve had some awesome wins with partners.
Partners have available our assessment toolkits and our migration toolkits. We do not want them to have to build that from scratch. But we do need their help in being that last mile of getting customers over that hump.
What else should solution providers know about the continued virtualization opportunity?
It is a tremendous market opportunity. (And) we will always have services that we can backstop. We always want to be experts on our products. But partners, then, can add that capability of knowing customer environments, knowing the integrated ecosystem, and transforming.
(Red Hat virtualization solution providers see) multiple wins that they get out of what was just a frustrating opening position–’my virtualization costs just went up a ton, and I wasn’t planning on dealing with that.’ And now I’ve been able to turn that into a win.
If you’re moving to virtualization, you’re moving to OpenShift (Red Hat’s enterprise-grade hybrid cloud container orchestration platform). And you’re running that on bare metal–and it’s a different way to manage OpenShift.
If you want another win out of that, the perfect setup for running OpenShift on bare metal is running AI workloads. If you can get access to GPU (graphics processing unit) resources, you’re not going to want to virtualize those.
You can give the business virtual machines and AI capabilities. This is the first year we shared what we’ve been doing internally. We call it hybrid AI. But we’ve built our deep research agent, which 85 percent of that, now, runs on open-weight models at a significant volume.
We started with all frontier models. This actually works at the scale of our company. For partners, being able to say, ‘There is a hardware element there of what’s the best way to get GPUs?’ Is it run them yourself? Is it rent them? What is your environment that lets an OpenShift umbrella these, but be able to use this capability.
It’s a skillset a lot of customers don’t have directly yet.
How important are Red Hat solution providers in establishing AI use cases for customers?
The most important thing for me is the use case of–if I can help you through virt(ualization), what is the best place for you as a customer to start with AI.
In banking, it might be check scanning. That’s a well-known area to play in. If you’re a software company, it might be a deep research agent, like (ours), that can help really make you move faster.
We’re going to show you the power of both frontier models, but also, more importantly, the non-frontier models.
And then we’re going to show you a lot of layers (where) you can control this so it’s trusted for you.
(Partner success relies on), can they help you modernize and reduce this board-level risk that has been created by AI? Can you free up the operational funds for a customer to be able to invest in AI by solving their virtualization challenge? And then can you move them down that use case?
The minute you have that deep research agent win–which for us is like 200 agents at this point–you will never go back to thinking about your company the same way.
It’s really powerful. And for me, I don’t think a lot of customers can do all three of these themselves. That’s the opportunity with SI, service providers, MSPs.
What’s the right balance to host? It could be in the neocloud GPU agentside. It could just be in currency areas, ‘I can give you a better OpenShift stack faster.
It could be an arms and legs, ‘I can help you make this transition journey.’ It is the best time right now. There’s so much demand, both on the challenge side and opportunity, where partners can help shape that.
What are the obstacles to further Red Hat adoption?
When I talk to customers about it, it’s giving them the safest possible choice.
For a lot of customers, this technology, it’s quite alien to them. They don’t know what’s going to be. We live in this world, but they don’t necessarily know the nuances between (OpenAI AI model) GPT-5.5, (Anthropic AI model) Opus 4.7 and (Nvidia AI model) Nemotron (3) Super.
It sounds scary. If I make the wrong bet, am I going down the wrong area?
Then you have the virtualization challenge, which is, who else has gotten over this hump? Because it feels really risky if I only get 50 percent of it done. And partners, in many cases, for us, are that link.
We can provide the technology all day. And I can show that customers can get over this hump. And we are very good at virtualization. From the telco days on, we run this at massive scale. You’ll be fine.
And in AI, I can say, ‘Look at what we’ve been able to do with Nvidia’s open-weight models. With Nemotron and IBM Granite (the open AI models of Red Hat’s parent company). It’s all possible.
The reach is usually partners bringing that to life for a customer they know. We are able to cover some large customers ourselves. We can take them on that journey. The channel and partners are the ones that reach the rest of the world for us.
It’s having the confidence of where they add value to it, because it’s also fast moving technology for them. This is why we make the virtualization toolkits available to them and open sourced. Let them shape that to their own offering.
That’s why we really try to lean into the open-weight models’ accessibility, so they can work with that.
Red Hat’s biggest opportunity is with our channel, being able to make this safe, this journey.
If you’re a CEO, it’s the biggest bet and choice you’re going to make. And if you don’t know these things yourself, you have to have someone you can trust that can help put you on that journey.
There’ll be some customers where that’s us. There’ll be a lot more customers where that is through our partners.
How important is Red Hat and its solution providers for customers looking for model flexibility and vendor flexibility in the AI era?
I’ll pick Opus just as an example. Really capable frontier model. It works. This is incredible. I brought my context, my intelligence, to use and I can see it work.
And I can’t scale this to the level I want, because the dynamics don’t work on this model. And now I have the biggest win I have ever seen, and I don’t have the reach with that win.
Doing the inverse is equally challenging. So you know what I’m going to do? I’m going to start with the 3 billion parameter model. And I’m going to try to make it reason. That’s also going to fail, because that’s not a tool that’s suited for what Opus can do.
I usually tell customers, ‘You will have to use frontier models. I believe in this. But you won’t want to be beholden to anyone.’
Know what (Google AI model family) Gemini is good for and be able to use it for that. Know what Opus is good for. Know what GPT-5.5 (is good for). Don’t try to use it for volume because you’re going to get stuck in that area.
The other side is also true. Don’t mix up a 3 billion parameter model that we demoed in our lemonade stand with Opus. They’re very different things.
For partners, being able to shape that journey for a customer and say, ‘We are going to start you in frontier models. But don’t worry about being locked into this, because when we have that working, we’re going to go through this playbook that Red Hat knows.’
We will then start delayering this to figure out if It’s IBM Granite or a Nemotron Nano. You’re going to save a lot of money in that domain.
It may not work, and we’ll move up to a Nemotron Super or we’ll look at the (Google) Gemma models.
That journey of saying, ‘I can start you with something so we learn the art of the possible. And in learning the art of the possible, I will not let you get locked in.’ That nuance of a story is what customers need to hear.
Going back to the virtualization opportunity, how can solution providers use that as an opening to go deeper with customers?
We try to be pretty open with this, where virtualization is–don’t just do this as a cost play.
You are moving to a converged platform on OpenShift. And when you make that move, you can lift and shift everything you had and run it the same way you did after.
You won’t get a ton of benefit from that. You just get out of a commercially challenging situation.
But in that move, if you say, ‘Well, I now have core OpenShift cloud-native capabilities, and I can containerize some of these so I can take on new work. And I have that proximity to AI.’
Where you were constrained before on what virtualization could do, it is expanding it to really cool areas on both sides. I love that story because it makes the switching costs worthwhile.
Your yearly costs will be lower. But there’s still effort and retraining and re-skilling people. And so when you start seeing these compound benefits into adjacent areas with it, that makes it a really powerful journey on virtualization.
We (Red Hat products) run from space to factories to traditional data centers. Partners being able to see those opportunities and say, ‘Hey, I can just lift and shift every VM (virtual machine) you have. But have you thought about modernizing this one space? Because you’re going to get just a really asymmetric benefit out of a little bit of effort.’
That’s that nuance that if you know a customer, you know the environment, it can be really powerful.
Sounds like this might help with solution providers navigating customer change management around AI, introducing AI capabilities in familiar environments and context?
We announced the Redhat.com/skills location because, on one side, a lot of the software industry says, ‘Well, OpenAI and Anthropic feel like competition because they can write software. The world’s not sure how to frame this.’
I’m always an optimist. But let’s flip it to the other side–you have this semi-standard way to introduce your knowledge to them through skills. They can be the best teachers on the planet for our customers and partners as to how to get the most out of the Red Hat subscription.
And for us, it’s more of a natural position because we never really charge for the software anyway. The software is free. The customers own the licenses.
But we’re big and we have a lot of capabilities that customers and partners can leverage. And it’s sometimes hard to know how they make the most out of that. I love being able to introduce things to make partners more effective, to make customers more effective with our subscription offerings.
It’s how I’ve noticed I use the tools myself, even on commercial software that I buy. But being able to extend that when you’re not really competing on the bits and bytes, but on how people use the most of your service, that’s another powerful element.
For partners that are familiar with Red Hat and they’re comfortable with OpenAI and Anthropic skills-supported tools, using these to learn how they can do even more is new. It’s evolving really quickly. But it’s exciting to get to step into that space, too.
How about IBM solution providers interested in growing their Red Hat practices, and vice versa, and now solution providers that work with IBM’s latest acquisitions–what’s your message to those partners?
We always partnered closely with Hashi(Corp, which IBM purchased last year). We are like two sides of the sandwich. You have the procedural side with Ansible and you have the state side with Terraform (HashiCorp’s open-source infrastructure-as-code tool).
Vault (HashiCorp’s identity-based sensitive information management tool) has been super well-aligned with OpenShift.
Having those capabilities in the family and not worried about whether, on open source, we’ll compete with each other–it makes that partnership much stronger.
Confluent (the real-time data streaming vendor IBM bought in March) has been a phenomenal add to the IBM portfolio.
When we talk about hybrid and we talk about how you’re going to stitch edge environments together and how you’re going to get AI to respond, messaging technology, where Confluent really leads that area–it is the core glue of how you move data between these.
They both have a foot in open source. We speak the same language on it, whether it’s Terraform on one side or Kafka (the open-source data streaming platform Confluent is built on) on the other. And that’s been a natural synergy.
And how about in IBM’s product portfolio outside of the recent acquisitions?
If you look at IBM Sovereign Core (a new software foundation revealed during IBM’s Think 2026 conference earlier this month) or IBM Concert (an AI-powered operations platform now in public preview)–I talked about the hygiene, the matching and back to basics with RHEL. An enterprise has more to deal with than just RHEL.
We work really closely with IBM. We can have every aspect of our security subscriptions offerings, what we’re doing available in something like Concert. And then IBM Concert can umbrella your total security story for an enterprise.
Those have just been really nice alignments. We have those with other partners. But IBM has a vested interest in Red Hat success, so they’re going to tend to come to us first to make sure that they’re going to use every aspect of the Red Hat subscription to show up in Concert.
We’re going to make sure that we hold that balance correctly with Confluent and Kafka. So those have been really beneficial.
Red Hat, we are infrastructure plumbers. That’s the lane that we stay in. With a lot of these solutions, you will always need partner capabilities. And IBM, I think, has done a good job at being an exemplar of what it can be to build on OpenShift, RHEL, Red Hat capabilities.
What other markets are ripe for Red Hat solution provider expansion?
Red Hat Desktop with Podman Desktop (an open-source containerized software applications tool for local developer environments that features Red Hat as lead contributor)–that developer motion of, to run AI workloads, it’s otherworldly complicated infrastructure.
And so how do you start as a developer on something that will mimic prod(uction environments). That has always been a developer cycle challenge. But if you don’t get good at that, you’re kidding yourself that you’re going to build your own agents with it.
Getting partners really thinking about that full lifecycle, not just, ‘I am going to deploy GPU boxes and I’m going to expose them through Red Hat AIE (AI Enterprise) with Nvidia.’
But really thinking through the use case and the developer enablement of that–that’s the other area that, I’m biased obviously, but we add a ton of value of all of these layered security capabilities and our strengths in process control and SELinux (Security-Enhanced Linux, a kernel security module for mandatory access control).
We can start you off on a very good production-worthy path in development. If you don’t do that, you may be ending up in a whack-a-mole pretty late in the cycle, which could really challenge your AI experience.
We always believe in choice. We will support every tool out there from the platform perspective. But I do think that’s an area to be aware of.
It’s lower level than the IDE (integrated development environment). But what are those primitives of how you would start building agents on your laptop in a way that you are using some of those capabilities that are going to be part and parcel for how you secure agentic workloads in prod?
They’re very new. We use some of the same technologies. But then we also use very different technologies. And getting developers on a good path there is critical.
Should developers fear AI taking their jobs?
I grew up as a developer. (And) I love the craft of writing code. It was puzzle solving for me. I genuinely enjoyed it. I could stay up all night doing it. It brought me energy.
That part of the craft is gone at this point. As fun as it is to write algorithms, solve those puzzles, AI is just better at it than the vast majority of the developers on the planet.
If you’re stuck in that world, it is more disruption than good. If you’re like, ‘But I love doing this.’ I’m like, ‘I know. But I love doing it, and it’s still gone. That doesn’t mean I don’t love it. But AI is very good at it.’
If you switch to the engineering mindset, if you look at what we built internally, my developers had to deal with me contributing code into our systems, and their world shifted to, ‘How do you tell if my code is good or not good?’ Because I don’t have time. I used AI for all of it.
That mindset shift–I’m not writing code, but I am writing evaluations. And I am really thinking through how to do continuous integration at a different scale and level.
That’s not going to be for everyone. And I have talked to some engineers, that doesn’t drive energy for them.
But that is a very valid application of an engineering mindset that is critical. It is your biggest limiter of scale in AI. There’s plenty of opportunity there. The field will shift there.
I’ve gotten plenty of enjoyment working in both of those domains. Maybe not as much as the puzzle solving, but I do think the puzzle solving aspect is gone.
The skills are still equally valuable in this. But it does require a shift in thinking.
The good part is, it’s not just developers going through this. It requires a shift in thinking in CFOs and CEOs and COOs and general counsels. AI will require a different way to shape your role in every role in the company. But developers are included in that.
In 20-some years, they have not really been changed. (And) this does require a change. But I’m very optimistic on it. For people that have that mindset of, ‘I can solve problems. And I can shape AI to solve them, they will be more valuable than they were a year ago. Not replaced by the technology.’
