Key Highlights
- International Business Machines and Together AI announced a multi-year $240 million partnership to deploy a massive AI inference platform on IBM Cloud infrastructure.
- The platform leverages Nvidia HGX B300 hardware paired with Spectrum-X Ethernet networking technology.
- Together AI plans to utilize this infrastructure for operating open-source AI models, processing approximately 400 trillion tokens per month.
- Commercial availability is scheduled for the first quarter of 2027.
- IBM shares climbed 1.51% following the announcement; Nvidia stock advanced 0.46%.
International Business Machines has entered into a $240 million multi-year partnership with Together AI to construct a high-capacity AI inference platform on IBM Cloud, utilizing Nvidia’s advanced hardware. Following the announcement, IBM shares increased 1.51%, while Nvidia stock rose 0.46%.
International Business Machines Corporation, IBM
The partnership focuses on delivering inference capabilities for open-source artificial intelligence models. Inference represents the computational process where trained AI systems generate outputs, and it has emerged as a critical factor driving enterprise computing requirements.
The infrastructure will be constructed using Nvidia HGX B300 platforms, which incorporate Nvidia’s latest Blackwell architecture. The deployment will also integrate Nvidia Spectrum-X Ethernet networking infrastructure.
According to IBM, this represents the inaugural purpose-built, enterprise-scale inference platform on IBM Cloud utilizing these advanced systems. The platform is projected to become operational during the first quarter of 2027.
Together AI provides a comprehensive platform enabling developers and organizations to construct and implement AI applications using open-source models including DeepSeek, MiniMax, and Kimi. The organization currently processes approximately 400 trillion tokens on a monthly basis.
Vipul Ved Prakash, CEO of Together AI, stated the infrastructure will enable the organization to deliver “production-grade inference to more companies, faster.” He emphasized that organizations are seeking advanced performance capabilities without incurring the expenses associated with proprietary models.
IBM Cloud’s general manager, Alan Peacock, indicated that IBM and Nvidia are providing “scalable, economical, enterprise-grade AI infrastructure” through this collaboration.
Open-source artificial intelligence has experienced growing adoption as organizations seek to reduce AI-related expenditures. Recent cybersecurity concerns affecting models from Anthropic, OpenAI, and Meta have additionally motivated enterprises to explore open-source alternatives.
Nvidia’s Strategic Role
The Blackwell processors central to this partnership represent Nvidia’s most recent generation, which Nvidia has optimized particularly for AI inference applications. The Spectrum-X networking infrastructure is engineered to manage the substantial data throughput required by enterprise-scale AI platforms.
Together AI indicated it chose IBM and Nvidia based on their technology development plans and capability to provide GPU resources efficiently at competitive token costs.
IBM’s Open-Source Strategy
This partnership aligns with IBM’s comprehensive strategic direction. Previously this year, the corporation unveiled Project Lightwell, a $5 billion initiative with Red Hat designed to assist enterprises in securing open-source software through AI-powered tools and a workforce exceeding 20,000 engineers.
Initial participants in that program include Bank of America, Goldman Sachs, and JPMorganChase.
The IBM-Together AI agreement represents one component of an extensive IBM-Nvidia collaboration encompassing GPU-optimized data analytics, unstructured data processing, hybrid infrastructure solutions, and professional services.
Together AI recently completed an $800 million Series C funding round, achieving an $8.3 billion company valuation. The organization maintained this valuation level since July.
IBM verified that the platform will be the first deployment of its type on IBM Cloud incorporating Nvidia HGX B300 systems, with widespread availability planned for the first quarter of 2027.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
