the news now

The world's news, cross-checked among reputable sources.

This is a new development in a story we have covered before · earlier coverage

IBM and Together AI Announce $240 Million Deal for Nvidia GPU Cluster

morning · · thread: Nvidia and Wall Street Firms Partner to Raise $500 Billion for AI Infrastructure

2 sources · United Kingdom

Who reported this

  • Reuters United Kingdom · Centre · Thomson Reuters Corporation
  • The Register United Kingdom · Centre · Situation Publishing Ltd

What the colours mean

  • Left
  • Centre-left
  • Centre
  • Centre-right
  • Right
  • A hatched block means the outlet is affiliated with, or controlled by, a state.

Political lean describes where an outlet sits within the politics of its own country. It is never a position on a single global scale.

Political lean is comparable inside one country and not across them, which is why the bar groups by country first. Publicly funded broadcasters are not marked as state-linked.

The owner of each outlet is listed as a matter of record, not as a judgement about the outlet.

Every source for this story reports from United Kingdom.

Every outlet covering this story shares the same political lean; read with that in mind.

IBM and Together AI have entered into a $240 million agreement to run Together AI's open weights inference platform on a large cluster of Nvidia GPUs housed in IBM Cloud. The deal is driven by the need for rapid AI scaling and the current shortage of AI infrastructure, including power and datacenter capacity. Together AI selected IBM to secure the necessary GPU capacity to maintain low token costs for its users.

The deployment will utilize Nvidia B300 GPUs, which are part of the HGX platform announced in early 2025. Unlike the most powerful 72 GPU rack systems, the HGX platform is more conventional and suitable for traditional air cooled datacenters. Each 14 to 15 kW box contains eight B300 GPUs interconnected by NVLink and Nvidia Spectrum X Ethernet fabrics. IBM stated that this represents its first large scale use of B300s for inference applications. The compute capacity is expected to become operational in the first quarter of 2027.

How each side framed it

Centre
Outlets with a center lean focused on the technical specifications of the hardware and the strategic necessity of the deal amid industry wide supply chain constraints.

Sources

100% of the statements in this article were traced back to the source articles listed above.