Condor Galaxy

The Cloud AI Supercomputer Built for Generative AI

Built by G42 and Cerebras

Exa-scale performance, single device simplicity

Condor Galaxy, the supercomputer built by G42 and Cerebras, is the simplest and fastest way to build AI models in the cloud. With over 16 ExaFLOPs of AI compute, Condor Galaxy trains the most demanding models in hours rather than days. The terabyte scale MemoryX system natively accommodates 100 billion+ parameter models, making large scale training simple and efficient.

Wafer-Scale Cluster Datasheet

• 4 ExaFLOPs
• 64 x CS-2s
• 82 TB of Memory

• 4 ExaFLOPs
• 64 x CS-2s
• 82 TB of Memory

• 8 ExaFLOPs
• 64 x CS-3s
• 108 TB of Memory

Condor Galaxy 3

8

exaFLOPS of AI compute
at FP16

64

Cerebras CS-3 systems

164

terabytes of memory

720

billion parameter models, extensible to 24 trillion

338

terabits-per-second node-node on-chip bandwidth

72,704

AMD Epyc™ Gen 3 cores

Unprecedented ease-of-use

Near-Perfect Performance Scaling

Unlike any known GPU-based cluster, Condor Galaxy delivers near-perfect scaling across GPT-class large language models, including GPT variants, Falcon and Llama.  Near-perfect scaling means that as additional CS-3s are used, training time is reduced in near perfect proportion.

Setting up a generative AI model takes minutes, not months and can be done by a single person.

What our customers are saying

“AMD is committed to accelerating AI with cutting edge high-performance computing processors and adaptive computing products as well as through collaborations with innovative companies like Cerebras that share our vision of pervasive AI. Driven by more than 70,000 AMD EPYC processor cores, Cerebras’ Condor Galaxy 1 will make accessible vast computational resources for researchers and enterprises as they push AI forward.”

– Forrest Norrod, EVP & GM, Data Center Solutions Business Group, AMD

“With Condor Galaxy 3, we continue to achieve our joint vision of transforming the worldwide inventory of AI compute through the development of the world’s largest and fastest AI supercomputers. The existing Condor Galaxy network has trained some of the leading open-source models in the industry, with tens of thousands of downloads. By doubling the capacity to 16 exaFLOPs, we look forward to seeing the next wave of innovation Condor Galaxy supercomputers can enable.”

– Kiril Evtimov, Group CTO of G42

Get a Test Drive

Train or fine-tune your next AI model on Condor Galaxy. Easily access one or 64 systems via the Cerebras Cloud at competitive rates. Schedule a trial Today.

Try Condor Galaxy