Hardware for your AI

We want you to be as creative as you wish, so we’re not finicky on hardware. We’ve listed your options.

The A-Z on Computing power

Cloud
  • Best pick for proof of concepts and occasional model trainings
  • Ideal for projects that require scale and flexibility
  • Lower startup, energy and maintenance costs
  • We partner with Microsoft Azure and Google Cloud
On-prem
  • Best for live applications that need high speed, performance and reliability
  • Ideal for in-line industrial applications
  • Use With NVIDIA, IBM or Robovision box

TRAINING AND INFERENCE

A GPU (graphics processing unit) speeds up the AI process as it allows for processing operations to occur in parallel versus a CPU (central processing unit) where processing occurs sequentially. Which GPU to use is dependent on both the phase of your AI journey (training vs inference) as well as the performance requirements of your applications and use cases.

Training
  • NVIDIA
  • NVIDIA Tesla V100 cards
  • NVIDIA DGX-station
  • NVIDIA DGX-1
  • NVIDIA DGX-2
  • IBM Power AC922

(See all NVIDIA, all IBM)

Inference
  • Jetson AGX Xavier™
  • Robovision box
  • NVIDIA Xavier
  • NVIDIA TX2
  • NVIDIA T4
  • NVIDIA GTX, RTX series
  • IBM Power IC922

(See all NVIDIA, all IBM)

Our storage partners

To make the best use of your computing power, you’ll be happy to have a storage partner who’s able to deliver the data as fast as your training server demands it.

Robots with brains

Mechanical robots was the first step. Now we’re augmenting robots – which means we’re giving robots the ability to reason and to infer like you and I. With brains and automatic hand-eye coordination.

  • Denso Orin2
  • ROS (open source)
  • KUKA KRC5
  • Fanuc Karel

AI robotic Hand-eye coordination

We help you maximize your potential