TLDR: IBM Cloud is significantly enhancing its AI infrastructure by integrating AMD Instinct MI300X GPUs to power Zyphra’s advanced generative AI training. This multi-year collaboration provides Zyphra, a $1 billion-valued open-source AI company, with a large-scale, dedicated cluster on IBM Cloud, marking the first full-stack AMD training platform deployment on IBM’s cloud environment. The initiative aims to accelerate the development of frontier multimodal foundation models for Zyphra’s superagent, Maia.
ARMONK, N.Y. and AUSTIN, Texas – October 1, 2025 – IBM (NYSE: IBM) and AMD (NASDAQ: AMD) have announced a strategic multi-year collaboration to deliver cutting-edge AI infrastructure to Zyphra, a San Francisco-based open-source AI research and product company. This partnership positions IBM Cloud to provide a substantial cluster of AMD Instinct™ MI300X GPUs, complemented by AMD Pensando™ Pollara 400 AI NICs and AMD Pensando Ortano DPUs, specifically for Zyphra’s ambitious generative AI training initiatives.
This collaboration is poised to establish one of the largest advanced generative AI training capabilities to date, powered by a comprehensive AMD stack running on IBM Cloud. Zyphra, which recently secured Series A financing valuing the company at $1 billion, is dedicated to building a leading open-source/open-science superintelligence lab. Their focus includes advancing fundamental innovations in novel neural network architectures, long-term memory, and continual learning.
Krithik Puthalath, CEO and Chairman of Zyphra, emphasized the significance of this integration, stating, “This collaboration marks the first time AMD’s full-stack training platform—spanning compute through networking—has been successfully integrated and scaled on IBM Cloud, and Zyphra is honored to lead the way in developing frontier models with AMD silicon on IBM Cloud.” He added, “We’re excited to partner with IBM and AMD to power the next era of open-source, enterprise superintelligence.”
The advanced training cluster will be utilized by Zyphra to develop multimodal foundation models across language, vision, and audio modalities. These models are intended to power ‘Maia,’ a general-purpose superagent designed to deliver significant productivity benefits for knowledge workers across various enterprises. The initial deployment of this infrastructure commenced in early September 2025, with plans for further expansion throughout 2026.
Alan Peacock, GM of IBM Cloud, highlighted the strategic value of the partnership: “Scaling AI workloads faster and more efficiently is a key differentiator in achieving ROI for established enterprises and emerging companies alike. We are delighted to support Zyphra’s strategic roadmap as we combine IBM enterprise cloud expertise with AMD leadership in high-performance computing and AI acceleration.” Philip Guido, AMD EVP, further noted that the partnership sets “a new standard in AI infrastructure,” supporting advanced AI workloads for both enterprises and startups.
Also Read:
- IBM and HashiCorp Detail AI-Driven Infrastructure Automation with Project Infragraph at HashiConf 2025
- IBM Intensifies Focus on Watsonx and Hybrid Cloud to Capture Trillion-Dollar AI Market Share
Both IBM and AMD are committed to pushing the boundaries of performance and efficiency in AI infrastructure. Their ongoing efforts include plans to develop quantum-centric supercomputing architectures, blending IBM’s quantum expertise with AMD’s high-performance computing capabilities, which could redefine the landscape for next-generation AI workloads.
