Solution
Cost-Efficient Tenstorrent AI Clusters
Moreh designs and builds scalable AI clusters around Tenstorrent's network-integrated accelerators — optimized for inference and ready for training — from cluster architecture and networking to full-stack software and ongoing support.
What We Deliver
Chip
Wormhole Processor
Tenstorrent's Wormhole processor is a lightweight and efficient building block for AI clusters. It delivers 164 TFLOPS of BLOCKFP8 (BF16) performance and 12 GB of memory. The true power of Wormhole emerges when many chips come together to form a single cluster.
Learn more (tenstorrent.com) ›Server
Galaxy Server
Tenstorrent's Galaxy server is equipped with 32 Wormhole processors, delivering performance equivalent to a typical 8-GPU server.
Learn more (tenstorrent.com) ›Networking
Switchless Chip-to-Chip Torus Network
Every Wormhole processor has 3.2 Tbps Ethernet interfaces, forming a direct chip-to-chip torus network without switches — efficiently handling AI communication patterns at scale.
Software
Moreh AI Software
Moreh's full-stack software covers both inference and training on Tenstorrent hardware — production-ready from day one.
Moreh vLLM
Single-Node Inference Engine
- •Built for Tenstorrent Wormhole chips
- •SOTA MoE model support including DeepSeek, GPT-OSS, and Qwen
- •450+ optimized operations
MoAI Inference Framework
Cluster-Scale Inference
- •Scale across multiple Galaxy servers
- •Torus-aware scheduling & disaggregation
- •OpenAI-compatible API for the cluster
MoAI Training Framework
Fine-Tuning & Training
- •Fine-tuning and training on TT clusters
- •PyTorch-compatible, 450+ operations
- •Same cluster for inference and training
Coming soon
End-to-End Support
Hardware Supply
We source and supply Tenstorrent Wormhole chips and Galaxy servers, handling procurement so you don't have to.
Cluster Construction
We design and build the cluster — from rack layout and power planning to torus network topology.
Software Deployment
Moreh vLLM and MoAI Inference Framework are deployed and optimized for your specific workloads.
Technical Support
Ongoing support for Tenstorrent-specific issues, performance tuning, and software updates.
Through our partnership with Tenstorrent and 3 years of joint development, we deliver a fully integrated solution from hardware to software — ready to efficiently run a wide range of AI workloads.
Press release on a strategic partnership ›Ready to deploy Tenstorrent AI at scale?
From a single Galaxy server to a full cluster — for inference and training, we handle the hardware, software, and everything in between.
Contact Sales