Gaudi Server Rack

Habana’s AI Training processors to power the Voyager supercomputer

April 8, 2021

With native integration of 6x QUAD-100 GbE in every Gaudi node, Habana’s AI Training processors have been selected to power the Voyager supercomputer with one-of-a-kind, high-efficiency AI system scaling.

Read More
AWS logo

AWS announces new Amazon EC2 instances will leverage up to 8 Gaudi accelerators and deliver up to 40% better price performance than current GPU-based EC2 instances.

December 1, 2020

Today at re:Invent, AWS CEO Andy Jassy announced new AWS Amazon EC2 Instances based on Habana's Gaudi processor technology.

Blog

High-Performance Training
And Inference Solutions

HLS 1

Purpose-Built for AI Training

The only AI processor with Integrated RDMA over Converged Ethernet to enhance scalability and lower total cost of ownership.

For more information about Gaudi
8 Card Server
GOYA Habana Labs

Purpose-Built for AI Inference

Goya delivers outstanding throughput and latency for AI inference workloads

For more information about Goya
News & Events

Announcing availability of Habana’s SynapseAI® 1.0.0 Release for our software platform.

Read the blog
Developer Site

Introducing the Habana Developer site with the software, tools and support developers and researchers need.

Read the blog
Intel AWS

Intel CEO interviewed regarding AWS announcement of Habana-based EC2 Instances

Read More
AWS re:Invent

See the latest coverage from re:Invent, Amazon EC2 instances will leverage Gaudi accelerators.

Read More
Bow Swan | Intel CEO

Intel pledges $1 million in support of efforts to address social injustice and anti-racism

Read Intel CEO’s memo
Eitan Medina | Habana AWS

See Habana’s Eitan Medina and AWS’s Chetan Kapoor on TheCUBE interview regarding the AWS announcement

See the interview
Intel Acquires Habana

Intel acquires Habana Labs.

Read the announcement

MLPerf reports Goya performance in industry’s first published inference benchmarks

Read the blog

Industry standards like Glow benefit data center and cloud customers.

See the blog.

Habana takes the stage with industry leaders at AI Hardware Summit

Read More

Goya delivers 1,527 sentences-per-second on BERT NLP model.

See the details.

"Startup Projects to Hit Nvidia V100 Performance at Half the Power." Linley Group

Read More