IBM has signed a $240 million agreement with Together AI to build a large-scale inference cluster, CryptoBriefing reported. The deal adds IBM to a growing list of established technology firms partnering with specialized AI infrastructure providers. Financial terms beyond the headline figure have not been made public.
Together AI operates in the market for AI compute infrastructure. The company focuses on providing hardware and software designed to run large language models efficiently. Inference clusters like the one described in this deal are used to process live AI workloads after models have been trained. This differs from training clusters, which are used to build models from scratch.
Demand for inference capacity has risen sharply as companies deploy AI models into production. Training a model happens once or periodically. Running that model to answer queries, generate content, or power applications happens continuously and at scale. This has pushed many firms to invest heavily in dedicated inference infrastructure rather than relying solely on general-purpose cloud computing.
IBM has positioned itself as an enterprise AI provider in recent years. The company has emphasized hybrid cloud and AI tools aimed at large corporate customers. A deal of this size suggests IBM is prioritizing direct investment in specialized compute infrastructure rather than depending entirely on third-party cloud providers for AI workloads.
The scale of the reported deal, at $240 million, places it among the larger publicly disclosed infrastructure agreements in the AI sector this year. Such deals typically cover hardware procurement, data center buildout, and long-term service arrangements. It remains unclear from available reporting whether the cluster will be built for IBM’s exclusive use or shared with other Together AI clients.
Neither IBM nor Together AI has released a joint statement confirming additional details about the project’s timeline or physical location. CryptoBriefing’s report did not specify when construction or deployment of the cluster would begin. As with many infrastructure announcements in the AI sector, further details often emerge in subsequent regulatory filings or company disclosures.
Sources disagree on this story
This article was published before the reports below were compared. The reporting above stands; what follows is where the published accounts do not agree.
CryptoBriefing and Cryptopolitan both cover the $240M IBM-Together AI inference cluster deal but give different founding years for Together AI.
What all sources agree on
- IBM signed a $240 million deal with Together AI to build a large-scale inference cluster on IBM Cloud.
- The deployment is targeted for 2027.
- Together AI provides infrastructure/cloud capacity for open-source AI models rather than closed models.
- The deal is described as multi-year and tied to enterprise AI inference demand.
Where the reports disagree
1Year Together AI was founded
Together AI, founded in 2020, has carved out a niche as an infrastructure layer for open-source AI models.
The AI company, founded in 2022, rents cloud infrastructure to companies that want to build on open, modular AI stacks instead of closed models.
What would settle it: Together AI's own corporate registration records or its official company history statement.
What to make of it
Treat the $240 million IBM-Together AI inference cluster deal and its 2027 target as established; the two reports cannot both be correct on when Together AI was founded, so that detail should not be repeated as fact until checked against Together AI's own corporate records.
Market Impact
If confirmed through additional disclosures, the deal would reinforce a broader trend of established technology companies committing significant capital to AI compute infrastructure. Large inference deals of this size can affect demand for chips, data center capacity, and specialized cloud services, which in turn influences suppliers across the AI hardware supply chain.
For Together AI, a $240 million agreement with a major enterprise technology firm like IBM would represent a substantial commercial win. It could also signal growing competition among infrastructure providers seeking large corporate clients as AI deployment scales beyond pilot projects into production systems.
The reported agreement points to continued large-scale investment in AI infrastructure by established technology firms. Further details on the cluster's scope and timeline are expected as the companies release additional information.
Frequently Asked Questions
What is an inference cluster?
An inference cluster is a set of computing hardware designed to run trained AI models in production, handling tasks like answering queries or generating content in real time.
How is inference different from AI training?
Training builds a model using large datasets, typically a one-time or periodic process, while inference runs the finished model continuously to serve live applications.
What does Together AI do?
Together AI provides infrastructure, including hardware and software, that helps companies run large AI models efficiently at scale.
Has IBM confirmed the deal publicly?
The agreement was reported by CryptoBriefing, and further public confirmation or additional details from IBM and Together AI have not yet been disclosed.