Robuta

Supermicro Ultra servers are now available from Graphcore Elite Partners as part of the IPU-POD product range, helping innovators to build AI compute systems...

Graphcore secures additional $150 million in new capital in Series D2 funding round.

Graphcore IPUs are shown to accelerate neural network training for cosmology applications in a new Université de Paris research paper.

Learn more about the Graphcore Poplar® SDK and get started programming IPU systems.

Graphcore announces Series E funding round of $222m from Ontario Teachers' Pension Fund, Schroders and Fidelity International as well as existing investors

A new paper by a group of Graphcore researchers will be published and presented during the upcoming ICML 2022 Hardware-aware efficient training workshop.

National University of Singapore researchers are using mixture of experts spatio-temporal graph convolutional networks to deliver journey time predictions.

In a new podcast, Graphcore speaks with Weights & Biases about hardware's role in advancing machine intelligence and the challenges of AI scale-out.

Graphcore is launching a partnership with AI platform company Pienso to make its powerful text analysis solution available on IPUs in the cloud.

Graphcore's C600 PCIe card delivers up to 560 TFLOPS (FP8) / 280 TFLOPS (FP16).

Graphcore expands its channel partner network in North America with new and extending partnerships to help meet growing demand for IPU compute systems.

Start building your AI product or service on Graphcore IPUs today - creating an API couldn't be simpler.

Graphcore establishes regional hub in Singapore to serve customers in Southeast Asia, supported by new partners NetWeb, Cxrus and INT2.

Participants selected to join the Graphcore Academic Programme will receive free access to IPU compute in the cloud.

New performance results for Graphcore MK2 IPU systems with orders of magnitude improvements for some models, outperforming Nvidia A100 DGX across the board.

Graphcore IPUs will deliver high-performance AI computation as part of a new US supercomputer designed to support cutting-edge scientific research projects.

The new partnership will allow Graphcore to accelerate IPU technology adoption in China by building a network of specialist resellers and channel partners.

Jul 24, 2025 - Aleph Alpha and our partner Graphcore are unveiling a significant advance in AI compute efficiency, with the sparsification of a 13bn parameter model down to...

Aug 12, 2024 - Monika Grubizna, PopVision Team Lead at Graphcore, shares her journey from graphic design to software engineering, including challenges and insights.

The new IPU-POD16 Direct Attach lets innovators explore new AI approaches, with TCO advantage and higher performance for training and inference workloads.

G-Core Labs announces the availability of Graphcore IPUs in its European-based cloud service for advanced AI compute.

Graphcore and Aleph Alpha have successfully sparsified a chatbot model from 13bn parameters to just 2.6bn.

Graphcore Series C $50M funding. Graphcore & Sequoia Capital partner for growth in AI processor market

Graphcore is open sourcing our Poplar Graph Libraries (PopLibs) to allow the entire developer community to modify and hopefully enhance these powerful tools....

First Graphcore Bow Pod system installed at EPCC, home of the UK's most powerful supercomputer.

The new PyTorch Lightning integration enables developers to run any PyTorch models on the IPU with minimal code changes and optimal performance.

PaddlePaddle, the popular deep learning platform developed by Baidu, now fully supports the Graphcore IPU for both training and inference.

Graphcore's new Bow Pod systems deliver 40% higher performance and 16% better power efficiency thanks to the word's first Wafer-on-Wafer processor.

Graphcore is delighted to announce that its Elite Partners are now able to offer Lenovo servers as part of the Bow Pod product range.

This major new partnership will see Atos and Graphcore bring AI compute to innovators around the world.

Graphcore have dramatically accelerated a weather forecasting model on the IPU-POD, helping to predict extreme weather as the climate crisis intensifies.

Graphcore researchers are presenting a paper at NeurIPS exploring a new application of FP8-AMAX for improved computational efficiency.

Han Zhao, software team lead at Graphcore China, provides an explanation and code walkthrough of running BERT-base on the IPU using PaddlePaddle, Baidu's...

Graphcore has open-sourced Argon Streams AV1, a comprehensive verification tool supporting next-generation video compression and delivery over the web.

Graphcore IPU systems can now be used to carry out bulk semantic embedding using Aleph Alpha’s ground-breaking Luminous-Explore model.

Storage technology specialists DDN, Pure Storage, Vast Data and WekaIO join Graphcore’s Elite Partner Program.

Graphcore is joining the Ultra Ethernet Consortium to help shape the technical and strategic direction of next-generation Ethernet.

Graphcore and Hugging Face have significantly expanded the range of modalities and tasks available in Hugging Face Optimum, an open-source library for...

The official Graphcore Research blog.

Researchers at the Oxford-Man Institute of Quantitative Finance have accelerated advanced price prediction model training using the Graphcore IPU.

Dell Technologies Capital invests in Graphcore Series A funding for AI hardware and machine learning processors

PyTorch Geometric (PyG) is the leading framework for developing Graph Neural Networks. In this tutorial we walk through how to use PyG on Graphcore IPUs.

Dave Lacey discusses the IPU's Poplar software architecture on the Practical AI podcast, from PyTorch and TensorFlow right down to direct IPU programming.

May 28, 2025 - Dive into the evolution of engineering to tech leadership with Graphcore. Explore the intersection of AI innovation and executive vision.

AI for Simulations is a rapidly growing class of applications - traditionally undertaken by HPC - where Graphcore IPUs are showing standout performance.

Graphcore CTO & Co-Founder Simon Knowles talks future directions for AI and how to train tomorrow's superhuman machine learning models efficiently at...

The leading framework for building and running Graph Neural Networks (GNNs) now works with the leading GNN compute platform — Graphcore IPUs.

In their search to improve performance of option pricing models, quantitative researchers from London's Man Group evaluated IPUs with impressive results.

Graphcore opens a new engineering centre in Poland. The Gdansk site will be home to Graphcore's largest software development team outside of the UK.

Graphcore has built a new type of processor for machine intelligence to accelerate machine learning and AI applications for a world of intelligent machines

In this guest blog, UCL Research Software Engineer Mosè Giordano explains his implementation of Julia on the IPU, along with a demo Paperspace notebook.

The Hartree Centre will use Graphcore’s IPU technology for its research to deliver clean and commercially viable fusion energy in collaboration with UKAEA.

Graphcore announces plans for a 10 ExaFLOPS AI supercomputer called The Good Computer.

In this post, we explore the implementation of temporal GNNs on a new architecture developed by Graphcore that is tailored to graph-structured workloads.

Graphcore launches in Japan with new office, partnering with SCSK and HPC systems to bring ground-breaking IPU technology to customers in Japan.

Working together, Gridspace and Graphcore produced a 100M parameter speech application in a matter of days rather than weeks, accelerated by the IPU.

Pienso CEO Karthik Dinkar explains how Graphcore's BERT packing technique is helping his company deliver real-time data insights to customers.

At NeurIPS we're showing the Graphcore Rackscale IPU-Pod™ reference design which takes full advantage of the IPU’s scale-up and scale-out features, and can...

Graphcore expands into Germany, adding MEGWARE, one of Germany's most respected technology providers, to its Global Partner Program.

Graphcore joins Hugging Face hardware partner program to enable state-of-the-art Transformer deployment at production scale on IPU systems.

Graphcore is a key member of a new European project to create a supercomputing framework leveraging the efficiency of sparse compute for AI applications.

Nigel Toon, Graphcore CEO, gave a sneak peak of the joint Dell Graphcore IPU-Appliance for enterprise datacenters this week at a Dell event in Chicago. The...

Sensoro’s ESG solution will use IPU compute for applications such as ecological ecosystem protection, citizen safety and environmental health, animal welfare...

Work by PNNL and Graphcore could be applied to a wide range of applications, from drug discovery to developing more efficient solar panels.

UbiOps users can easily create AI microservices and pipelines for both deployment and training workloads on IPU enabled nodes.

Making Graphcore’s containerised software easily accessible for developers, Graphcore joins Docker Hub’s Verified Publisher Programme.

Updated Graphcore IPU benchmarks for training and inference machine learning models

Graphcore's MLPerf submissions show big performance gains on ResNet and BERT while Baidu's PaddlePaddle submission reflects the growing IPU ecosystem.

Graphcore shares outstanding performance results in first ever training submission to MLPerf, AI's most widely recognised comparative benchmarking process.

The leading framework for building and running Graph Neural Networks (GNNs) now works with the leading GNN compute platform - Graphcore IPUs.

Graphcore and Korea’s Electronics and Telecommunications Research Institute (ETRI) have entered a multi-year partnership to develop new software approaches...

Graphcore and its partners have created an 8-bit floating point format designed for AI, to be submitted to the IEEE as a proposed industry standard.

We hire exceptional people to help make our next generation AI stack - including silicon, hardware and software - the worldwide standard for AI compute

For Seoul-based AI translation startup Twigfarm, switching from GPUs to Graphcore IPUs yielded a 10x performance gain while simultaneously lowering costs.

Argonne National Laboratory has installed Graphcore’s Intelligence Processing Units (IPUs) within its AI Testbed.

Graphcore's IPU-POD system powers dramatic accelerations for Tractable’s computer vision damage assessment workloads.

Graphcore joins the influential list of PyTorch Foundation members, reflecting our commitment to PyTorch and contribution to "hardware and open...

Graphcore now supports Alibaba Cloud’s Open Deep Learning API (ODLA), which is designed to rapidly enable heterogeneous computing and meet the growing demand...

Graphcore is creating a range of high-skilled new jobs across the company.

Sep 8, 2025 - GNNs - powered by Graphcore IPUs - are enabling breakthroughs in drug discovery, consumer internet, e-commerce, quantum physics and many more areas.

Graphcore delivers 26x performance gains for Finance customers

The next phase in our growth story starts today. Graphcore secures new $200m funding from BMW, Microsoft & leading financial investors to drive growth

Sep 8, 2025 - Graphcore and Paperspace's new pay-as-you-go notebooks allow users to quickly and affordably scale-up their IPU capability as their AI compute needs grow.

Korean technology giant NHN will use Graphcore IPU-POD systems for its upcoming AI compute projects.

Project Radium will enable pooling and sharing of IPU resources over the primary datacenter network in virtualized, multi-tenant environments without pushing...

Graphcore today announced that the company has been acquired by SoftBank Group.

Japan’s Kindai University has chosen Graphcore’s IPU for its machine learning research programme drawing from the latest developments in neuroscience.

Graphcore comes out of stealth mode and announces $30m in funding to accelerate AI and machine learning.

Innovators in Korea will be able to access IPU technology through KT’s Hyperscale AI Service.

Graphcore makes the Intelligence Processing Unit. Graphcore has 80 repositories available. Follow their code on GitHub.

Graphcore and Pienso's partnership, accelerating AI insights with language models for low-code/no-code users, has been recognised with a CogX Award.

Sep 8, 2025 - A free new service will provide IPU access natively within Paperspace’s Gradient platform for building, training, and deploying machine learning models.

DP Technology has successfully launched its DeePMD-kit system for molecular dynamic simulation powered by Graphcore IPU systems and Poplar software.

Research teams at Pacific Northwest National Laboratory have been using Graphcore IPUs to help accelerate research in Quantum Chemistry and Cybersecurity.

An IPU consists of many cores called tiles: its 1472 tiles can run completely independent programs. Here, we explain how to optimise a simple Poplar codelet to...

Graphcore and Pure Storage's first reference architecture details how the IPU-POD and Pure Storage FlashBlade systems work together to deliver outstanding AI...

Graphcore - a SoftBank Group Company - announces the opening of its AI Engineering Campus in Bengaluru, India.

Graphcore and Aleph Alpha will work together on research and deployment of Aleph Alpha’s advanced multi-modal models on current IPU systems and the...

Graphcore and Tencent Cloud's IPU Preview service gives innovators access to poweful, flexible AI compute in the cloud.

LabGenius researchers are using BERT, running on Graphcore IPUs, to identify new treatments for conditions like cancer and inflammatory disease.