About Computer Modules

Who we are

Computer Modules Inc. was founded in San Diego in 1978 as a reseller of computer parts and systems. The company also trades as DVEO, a name many customers in broadcast and video know from decades of encoder and streaming hardware. Over forty years the catalog has moved with the industry, from boards and modules to servers and storage, and today the business is focused on one thing: AI compute hardware for people who need to run real workloads, not demos.

That means GPU servers, AI workstations, data center GPUs, cluster networking and the components that go around them. We sell to AI companies, research groups, universities, studios, engineering firms and the IT teams that support them. Most of what we ship is configured to order and tested before it reaches you.

Who buys from us

Over four decades the company and its DVEO division have supplied hardware to aerospace and defense programs, government laboratories, technology and telecom companies, and most of the broadcast networks in the Americas. The same purchasing, testing and export processes now serve buyers of AI compute.

How we source hardware

We are an independent reseller first. We work with several original equipment manufacturers and authorized distributors, and we hold registrations and partner status with NVIDIA's channel programs. We do not name those partners on this site, because the system you receive is specified as a generic class first, an 8 GPU HGX B300 server for example, and then sourced on the manufacturer platform that best fits your lead time, power budget and support preference at the time of the order.

For some classes, mainly PCIe GPU servers and workstations, we also offer our own white label configurations, assembled in San Diego from manufacturer chassis and components under our own part numbers. The quotation says which it is. HGX systems and rack scale products always come from the manufacturer as complete systems.

Working with more than one source matters for GPUs in particular. Allocation of data center GPUs moves between manufacturers from quarter to quarter. Having several routes to the same NVIDIA platform means we can usually find a system that lands when you need it, and we tell you plainly when we cannot.

What we do for buyers

• Specification. We turn a workload description into a bill of materials: GPU, CPU, memory, NVMe, fabric, power and cooling, sized with the rules in our sizing guide.
• Sourcing. We quote within one business day and manage allocation on GPUs that need it.
• Testing and acceptance. Every system, whether it comes from the manufacturer or from our own white label line, is run through the acceptance checks described on our services page before delivery, so you receive measured figures, not datasheet figures.
• Logistics. We ship from San Diego to anywhere in the United States and Canada, with a customs broker and export paperwork prepared before pickup.
• Export and end user registration. Data center GPUs require NVIDIA end user registration and US export compliance. We collect what is needed with your quotation and handle the filings. The plain English version is on our export compliance page.

Where to find us

Our office is at 7950 Silverton Avenue, Suite 103, San Diego, California 92126. Sales is open Monday to Friday, 8am to 5pm Pacific. Details are on the contact page.