Moreh and Tenstorrent Team Up for AI Innovation
Moreh, recognized for its optimized AI infrastructure software, has joined forces with Tenstorrent, an AI semiconductor company, to unveil a pioneering AI data center solution designed to be scalable and cost-efficient. This innovation is showcased at SuperComputing 2025, marking a significant advancement in AI infrastructure.
Revolutionizing AI Data Centers
The collaborative solution merges Moreh's MoAI Framework with Tenstorrent's Galaxy™ Wormhole server, providing an alternative to traditional GPU systems for large-scale AI inference and training tasks. This innovative combination addresses escalating operational costs and complexities associated with AI infrastructure, a common barrier faced by enterprises today.
Advantages of the Joint Solution
One of the standout features of this solution is its ability to offer both inference and training capabilities within a single platform. This dual functionality enhances scalability, allowing data centers to efficiently deploy extensive foundation models with reduced networking challenges compared to conventional GPU systems. As a result, clients can achieve greater efficiency while significantly cutting the total cost of ownership (TCO).
The Real-World Impact of the Collaboration
Developed through years of partnership, the joint offering harnesses Moreh's deep software expertise alongside Tenstorrent's innovative chip design. This synergy delivers tangible benefits that are applicable to real-world AI deployments, providing enterprises with a practical and effective solution for their growing AI needs.
Moreh is also expanding its support to include Tenstorrent processors, enhancing its vision of transforming into a universal AI software platform. This expansion signifies a notable step towards creating a flexible and economic AI infrastructure capable of addressing the diverse requirements of large-scale AI services.
Voices Behind the Vision
Gangwon Jo, CEO of Moreh, expressed enthusiasm about the partnership, stating, "Through this collaboration, we are introducing the flexibility and efficiency that AI data centers desperately need. Our objective is to make scalable AI infrastructure accessible for any enterprise aiming for performance without being confined to a single vendor."
Jasmina Vasiljevic, Senior Fellow at Tenstorrent, echoed Jo's sentiments, emphasizing, "Our collaboration showcases that open and co-designed systems can satisfy the increasing demands of AI at scale. Together, we're working towards making AI infrastructure more accessible and infinitely scalable for all."
About Tenstorrent
Tenstorrent builds computing solutions tailored for AI and the visionaries shaping its future. Their high-performance RiscV based CPUs, modular chiplets, and scalable computing systems empower developers, accommodating everything from single-node experiments to expansive data center deployments.
The ethos of Tenstorrent promotes an open future. Their architecture and software are crafted for adaptability, allowing editing, forking, and ownership by developers. This commitment is underscored by their team of engineers and innovative thinkers redefining the fusion of hardware and software to accelerate progress in technology.
Frequently Asked Questions
What is the main feature of the new AI data center solution?
The solution uniquely supports both AI inference and training on a single platform, enhancing scalability and reducing operational costs.
How does this collaboration benefit enterprises?
This partnership allows companies to efficiently scale AI workloads while significantly lowering total cost of ownership through advanced infrastructure solutions.
What are the key components of the solution?
The solution combines Moreh's MoAI Framework with Tenstorrent's Galaxy™ Wormhole server.
What future goals do Moreh and Tenstorrent have?
They aim to make scalable AI infrastructure more accessible and vendor-independent for diverse enterprises.
What does Tenstorrent believe about the future of AI?
Tenstorrent envisions a future where open, adaptable systems empower developers and facilitate innovation across AI technologies.
