IBM partners with Together AI to scale AI inference with NVIDIA, but near-term impact is uncertain.

What happened

IBM announced a multi-year collaboration with Together AI to deploy large-scale open-source AI inference clusters using NVIDIA HGX B300 systems on IBM Cloud. The move aims to capture enterprise AI workloads that demand fast, efficient production environments, expanding IBM’s AI infrastructure footprint beyond software and consulting. It directly addresses the hardware-centric spending trend that has pressured recent results, but the multi-year nature and capital requirements mean near-term financial impact remains unclear. While the partnership signals intent, IBM’s ability to convert AI hype into sustained revenue growth still hinges on consulting reacceleration and deal execution. Investors should view this as a necessary but unproven step in IBM’s strategic pivot.

Implication

The Together AI deal shows IBM is proactively tackling the AI spending shift toward hardware, but the financial benefits will take quarters to materialize and do little to immediately lift muted consulting revenue. The partnership may help IBM capture a slice of the 55.8% growth forecast for data-center systems, yet competitive intensity with hyperscalers remains high, and unless it pulls through higher-margin software and services, it risks being a capital-intensive distraction. For now, the thesis stays anchored to the need for consulting reacceleration and sustained Red Hat growth. Investors should monitor upcoming earnings for signs that this deal, combined with backlog conversion, is improving revenue quality, but base-case upside remains limited near $230.

Thesis delta

The agreement adds an AI infrastructure dimension to IBM’s strategy, which could gradually address the spending mix headwind, but does not change the core call that consulting growth and deal conversion are the critical proof points. We maintain a WAIT rating, with the partnership as an incremental positive that needs time to prove its impact.