‘(O)ne of our largest oil and gas customers moved over to Lenovo from Dell at our recommendation, and we sold them a blend of high-end workstation notebooks, mid-range, desktop stuff, and we also manage their data centers,’ John Rohrer, AmSys vice president of enterprise sales, tells CRN.
Texas born-and-bred IT systems integrator AmSys Innovative Solutions was founded in 2003 and now boasts roughly $350 million in annual revenue generated across 11 divisions.
Based in Houston, the company has partnered with tech giants Lenovo and Nvidia to build complex AI deployments for enterprises, municipalities and universities as large institutions take their first steps into AI.
“What really makes us different is we are not your classic reseller,” John Rohrer, vice president of enterprise sales and AI practice leader for AmSys, told CRN. “We’re really what you would call a master systems integrator, where we have 11 different business units that focus on anything from energy to hospitality to telco to field services to staffing to my group, which focuses on AI to go support our smart buildings and integrated technology group.”
He said the firm was historically a managed services company that offered hosting, data center and solutions focused around smart buildings, smart campus, and technology for contractors, healthcare, and energy. Now, leveraging its deep partnership with Lenovo, it is moving its customers into the AI era, delivering new PC deployments and infrastructure hardware, he said.
“So we’ve seen it run the gamut, for example, one of our largest oil and gas customers, moved over to Lenovo from Dell at our recommendation, and we sold them a blend of high-end workstation notebooks, mid-range, desktop stuff, and we also manage their data centers.”
Rohrer said AmSys’ education customers have also been early movers into AI, with institutions adapting their curriculum for this wave of technology. Rohrer singled out AmSys’ work with Prairie View A&M University, a historically Black university in Prairie View, Texas.
“Prairie View actually engaged us to provide modernization on how they bring AI to the university at every level, from the networking, Wi-Fi access and wayfinding to curriculum development for their doctoral nursing program and helping to enable students to get through that program faster,” he told CRN.
He said the university is building a custom large language model that incorporates the school’s digital library to help students navigate the educational road map towards a doctoral nursing degree. Additionally, Rohrer said the system gives students a digital concierge that can help in providing general information. CRN has reached out to the school about their AI initiatives but had not heard back by the time this story was published.
“We’re using Nvidia blueprints that we design our software on top of to run on the GPUs in the Lenovo servers. And so we’re starting off basically with around four to six servers that have eight to 10 GPUs to run the digital experience platform we’ve created,” Rohrer told CRN. “It’s a small block of servers, GPU and networking to go create that building block for them to grow on. So as they add workloads and use cases, we can just scale up and scale out as needed.”
In addition to education, AmSys also works with large enterprises in communications, municipal planning, and the energy sector, incorporating AI into smart designs for customers around the world, he said.
“We’ve got folks that are getting into the biggest hospitals in the world doing crazy pipeline leak detection systems in Kuwait, working in Saudi to go build vision AI platforms for the pilgrimage to Hajj,” he said. “So we have a global footprint.”
He said projects like that would not be possible without a partner like Lenovo, which saw massive growth in the sale of its infrastructure products during the most recent fiscal year, with revenue up 63 percent year over year to $14.5 billion as customers sought out the company’s high-end Neptune servers to meet the demands of AI workloads.
Lenovo’s annual revenue grew 21 percent year over year to $69.07 billion as sales of its PCs, infrastructure and services surged by double digits in each category for “one of the best years in company history,” Lenovo President of North American Sales Ryan McCurdy previously told CRN.
AmSys – which was named Lenovo’s partner of the year in 2025 – is a Lenovo 360 Gold tier partner. He said their collaboration with the Beijing and Morrisville, N.C.-based vendor has been critical to AmSys’ success.
“What’s really kind of driven our growth with Lenovo, is because they let us go do our thing and then support us in that effort and in this the ecosystem,” he said. “We can come in and say, ‘Look, here’s what we’re seeing in the market. Here’s where we need help.’”
Here’s more of what Rohrer had to say.
In the PC market, in terms of the refresh, do you see that coming? Is that actually bearing fruit?
So the way we’re seeing it is we’re going into clients that are trying to modernize and trying to get smart about leveraging AI as well as the Windows 10 end of life.
So they’re trying to be smart about how they build in the capability and functionality within a notebook. And what we’ve seen, we’ve actually been pushing clients to the P14s, which is the entry level workstation class notebook, as well as the P3 Ultra, which is a smaller, tiny level workstation, because we’re getting a lot of requests around Nvidia GPUs, which is one of our other partners.
Our collaboration with Nvidia on the data center side has led into our focus to help clients on the PC side. And so we’re basically promoting X1 Carbon, the premium PCs, because we want these things to actually last three years, and we want their users to like them for three years .
So we’re getting the latest and greatest from Lenovo, and they’re doing a great job of providing us demo systems. We’ve got a petting zoo in our office. I just got the new X9 and this thing is amazing. So they’re doing a great job of kind of helping guide us on what the systems look like. And then our customers, we’re just providing a platform that we then manage and secure for them.
Lenovo was one of the first that I saw to put Llama model on a device so that a user could run it, whether or not they were connected to the internet. Are you seeing demand for that? And have you seen any users demand models on a PC or on a workstation?
How we approach it is, we have our own product line and experience platform that uses large language models and visual search summarization and all the tool sets from Nvidia.
So we provide that as a kind of endpoint as well as an experience solution in the corporation or hospital or stadium or whatever it is. So we’re approaching it from that angle.
But I do see some clients that are wanting to build it out at the local endpoint level. So instead of investing in data center equipment, some of them are actually just buying a more beefy PC so they can do it at the endpoint.
Out of 10 customers we talked to, there’s probably one of those that’s kind of looking to do that. But, you know, we’re approaching it holistically, because we realize that not everybody wants to go to an open, large language model solution that’s in the cloud. Some of them are wanting to build their own or kind of build a hybrid solution.
Can you tell me about the Nvidia solution that you folks have? I’d love to hear a little bit about that.
We’ve been partnered with Nvidia for a couple years now and working closely with Lenovo on the Lenovo Hybrid AI 285 solution, and really the AI factory platform story that’s out there.
So we built some AmSys building blocks that provide customized, tailored, generative AI and RAG solution models so that when we’re going into universities, for example, they want to create curriculum specific to their students and then also support operational efficiencies within the organization and for the faculty.
So we’re creating customized language models, along with vision AI solutions across campus to go create these model solutions that are tailored to either the higher education institution, or the hospital, or the company itself.
It’s all customized, secured and sovereign and prominent to their environment, versus having to go out to the cloud. So we’re building those solutions on prem, and our data scientists are working with them to go build those neural networks and neural brains that all run the models that’s all based on their library of information.
That sounds very sophisticated, but also what a huge opportunity. Is it?
It’s been a journey, and we’ve built these platforms kind of custom tailored. So it’s a little bit more of an on-prem, services-led, software approach, versus just buying something off the shelf.
But what we found is that a lot of these universities and large organizations are buying racks and racks of GPU clusters. But those are primarily for research and for high-end kind of machine learning inferencing.
But at the end of the day, a lot of companies and organizations, whether you’re state, local government or higher education institution or a school, or an airport, you’re trying to solve a business outcome or an outcome for your students or faculty.
So we’re really addressing that part of the market, where we come in and start with workshops and help them understand what AI is and then figure out what are the process-based outcomes they want to achieve.
You’ve seen a lot of information from us around collaboration and collective and smart cities, because within these cities, you’ve got hospitals, you’ve got universities and sports and entertainment, sports tech that are all collaborating, trying to figure out where to start with AI.
So we really joined forces with Nvidia and Lenovo to create these building blocks, leveraging that technology foundational layer. And then AmSys has a practice around digital twins vision AI for computer vision, leveraging cameras and IoT devices, and then also the language modeling to create digital concierge or agentic AI or avatars based on what they’re trying to do, because the world of AI is not off the shelf.
There’s a lot of components from networking and from servers and from endpoints that have to deliver that solution. And so we’ve optimized that, whether it’s at the endpoint, understanding what kind of CPU or GPU they need, or the server. Not saying you got to go buy $5 million of GTX Nvidia products, but we can start with something smaller as a building block to get you going.
You’re on the ground. What are you seeing around agentic AI?
They talk about Wayne Gretzky is coming to skate to where the puck is going, right? Well, the puck is moving so fast. With agentic AI, that is definitely here, and we’re really working with ecosystem partners from Nvidia and ISVs, with Lenovo and some others to really build a foundational layer of starting with where is your data? How do we classify that data? What do you want to accomplish with that data? And then let’s use agentic AI to solve problems where we can automate or streamline to create efficiencies in your organization.
So we look at AI not to replace people, but just to help enable process and automation optimization. So whether it’s insurance companies or universities, trying to enhance contract management, procurement and business affairs, ultimately, we’re focusing on people that are trying to do more with less, which seems like everybody.
They’re not trying to replace people, they’re just trying to help the people that are in their job be quicker to deliver a business outcome.
So agentic is here. We’re working with a couple partners on it, and we have our own data science team that’s collaborating to create those workflows, and I actually just met with a partner yesterday to talk about their solution and what they’re doing with agentic.
And we have this digital experience platform framework that we feed off of. This is kind of our base level foundation toolset, and then everything kind of feeds off of that, so that you create this library across your organization that you can then go create pillars for departments as needed.
Lenovo saw a massive boom in server revenue last year, up 63 percent from the previous year. This is obviously a very competitive market. From a partner point of view, how are they doing when it comes to standing up their devices against their rivals?
At AmSys, we’ve worked with them all, so I can just tell you that we’ve shifted a focus and alignment with Lenovo for a few different reasons.
Number one is they have achieved hypergrowth. A lot of that’s probably been through their cloud team. I’m working closely with (Lenovo channel chief) Rob Cato. He’s been a key driving force, along with the rest of the channel team. … Those folks in the AI and HPC team have been incredible.
They have really kind of been a great partner, where we can come in and say, ‘Look, here’s what we’re seeing in the market. Here’s where we need help.’
And I think they’re leading the space in liquid cooling. The amount of breadth and depth they have in their portfolio is huge, and they’ve been doing it for a long time.
And then introducing us to folks like DreamWorks to promote what DreamWorks is doing in liquid cooling with Lenovo to give you that real understanding of what’s happening in the data center. And if you just take out the fans, you’re saving 30 percent in power and some of the nuances enabling liquid cooling. So that’s one area.
The second area, Lenovo is an incredible technology company. So huge R&D. We’re in Raleigh all the time, and they’re just really invested on the partnership part of it.
So they give us all the resources, all the demo gear we need. They work closely with us on some of our HPC clients to pull the right levers, turn the right knobs, whether we’re trying to test a crazy AMD processor or a high-end database workload or maybe an Intel CPU that’s just come out that we’re not familiar with. And we need to figure out how that works in the HPC world. They go do all those things for us.
We just haven’t had that kind of attention from some of the others. So that’s what’s been a driving force for our partnership.
And then it’s really about how we go to market together and then provide best cost to our customers too. So as we’re building these high-end solutions, how they work together, with the Nvidia GPU, with the high-end CPU and a ton of RAM that we’re trying to put in there. Those things get really hot. And we want to make sure everything’s going to work and it’s designed right. And then also we get a really good price.
So the teams across the board have done a great job of helping us kind of get to the number we need for our customers. So, I think their growth is coming from their ability to be aggressive in the market and go acquire new logos, because at the end of the day, their market share is a lot lower than the others. So how can we be tip of the spear to go provide these unique, differentiated solutions, leveraging their platforms, and then get to the number we need to get to so that we can price it effectively to the market once you kind of add in the distribution and the partner layer?
So they’re hungry. They’re aggressive. They’ve got great technology. And they’re coming out with really cool solutions around liquid-cooled platforms that we need to start talking about.
You’ve got such a great view of the market. Who’s doing AI right? Who is seeing return on investment?
It’s funny because AI is everywhere, but people are still figuring it out. It reminds me of kind of when VMware first came out, everybody was like, ‘Oh my gosh, you know, is this going to kill my server market share? What is this going to do?’
So, as an industrial revolution, I think what folks are really trying to do with AI is how do I create ROI. So let’s take universities, for example. So I was just at an event in Dallas with Dallas College, where they are the first to kind of build an AI bachelor’s degree. And, we’re talking to different universities, community colleges and state colleges, about kind of starting that journey.
There’s a lot of folks that have bought very large research solutions, and there’s a story around sovereign AI. And so creating an AI factory is great, but how do we kind of start with building blocks, to start small, to give people access at the curriculum level, to give students access and faculty access to learn what is AI and what does it mean to their organization? How do they train? How do they have vocational alignment and workforce development to understand what tools they’re going to need to take to the market so that when they graduate, whether it’s from high school or college, how can they get a job?
Because what we’re seeing is the biggest challenge is actually getting a job right, and AI is changing that paradigm. So for me, it’s the folks that really understand how to leverage AI as a way to make their organization more efficient versus just applying a big box that does inferencing and spits out GPT-like information.
