Introducing IPU-M2000

Second generation IPU systems for AI infrastructure at scale

Watch Video

OUR PRODUCTS

Graphcore IPU-Machine

The new core building block for AI infrastructure, designed for both training and inference. Available standalone to get started and with built-in networking for supercomputing scale, the IPU-M2000 packs 1 PetaFlop of AI compute in a slim 1U blade compatible with standard datacenter infrastructure. Available to pre-order today.

IPU-Machine: M2000

Graphcore IPU-POD

Building blocks and reference designs for AI supercomputing scale that every enterprise can benefit from. Graphcore IPU-POD₆₄ incorporates transformative innovations in compute and networking in an integrated AI infrastructure design built on Graphcore IPU-M2000 systems. Available today in our private Graphcloud™ preview.

IPU-POD

Graphcore IPU-Server

A range of industry standard OEM server systems for both training and inference, based on Graphcore dual-IPU C2 PCIe Cards with MK1 Colossus™ GC2 IPUs are available to order today from Dell and Inspur. Get started with development and experimentation then ramp to full scale production with 8xC2 card IPU-Server₁₆ or 2xC2 card IPU-Server₄.

IPU-Server₁₆ IPU-Server₄

Poplar® Software

Poplar is our easy to use, production-ready software co-designed with the IPU to simplify development and deployment so you can quickly be productive. The Poplar SDK supports both our first and second generation silicon and systems. PopLibs libraries are open source and available on GitHub with self-serve documentation, tutorials and guides on our developer portal.

Poplar Software

IPU Cloud And Server Solutions

Azure IPU Preview

The Graphcore IPU preview on Microsoft Azure is open for customers focused on developing new breakthroughs in machine intelligence.

Dell IPU-Servers

Dell offers 8-way and 2-way C2 PCIe IPU-Servers for enterprise data centre AI compute. Available to buy from Dell today.

Buy now

Cirrascale IPU Cloud

Cirrascale offers an IPU bare-metal cloud service and Dell IPU servers to buy - ideal for evaluation, prototyping, pilots and production AI workloads.