IBM has signed a $240M multiyear deal with AI cloud startup Together AI to build an AI inference clu...
The AMW Read
The deal updates Together AI's position in the AI infrastructure player map and signals strong demand for inference compute, but it's an incremental development for a known player rather than a structural shift.
IBM has signed a $240M multiyear deal with AI cloud startup Together AI to build an AI inference cluster on IBM Cloud using Nvidia HGX B300 systems, underscoring demand for dedicated AI infrastructure.
The partnership highlights how enterprise demand for AI inference—particularly for open-source models—is driving neocloud players and hyperscalers to collaborate on specialized infrastructure. Together AI, which raised $800M at an $8.3B valuation in July, continues to scale its position in the AI infrastructure market. This deal also signals that IBM is deepening its AI cloud offerings through partnerships rather than building purely in-house, reflecting a broader pattern where traditional enterprise tech vendors leverage external AI infrastructure expertise to stay competitive.
For builders, this means more options for deploying open-source models on enterprise-grade cloud infrastructure with dedicated GPU clusters, potentially lowering barriers to production-scale inference. For investors, it underscores that demand for AI compute—especially in the inference layer—remains robust enough to support large multiyear commitments from major enterprise players.
