Summary
Lenovo has announced AI Express, a programme offering three validated server configurations with NVIDIA hardware for enterprise AI workloads. Lenovo lists order-to-ship timelines starting at 15 days for eligible small configurations, subject to conditions and regional availability.
Lenovo announced AI Express on 30 September 2026, a programme that combines validated server configurations, software options and services to help organisations deploy production AI infrastructure. The three configurations are built around NVIDIA accelerators and are sized for different model sizes and user demands. Lenovo says order-to-ship can start at 15 business days for eligible configurations.
AI Express is a route into Lenovo’s Hybrid AI Factory with NVIDIA: infrastructure for running AI workloads, including inference, where a trained model processes new inputs to generate outputs. Lenovo positions the configurations as a way to start with capacity for current workloads and scale as demand grows.
Three configurations for different workloads
The small configuration is built on Lenovo ThinkSystem SR650a V4 with two NVIDIA RTX 6000 PRO Blackwell Server Edition GPUs. Lenovo lists it for focused inference serving tens of users, with average model sizes from 7 billion to 70 billion parameters and throughput of 30+ TPS. The configuration is listed as starting from 15 days.
The medium tier uses a Lenovo ThinkSystem SR675 V3 with eight of the same NVIDIA GPUs. It is aimed at higher-throughput inference and agentic AI for hundreds of users, with average model sizes from 70 billion to 400 billion parameters and throughput of 30+ TPS. Lenovo lists a starting timeline of 20 days.
The large configuration pairs a Lenovo ThinkSystem SR680a V4 with NVIDIA HGX B300. It is designed for deployments serving thousands of users and models of up to one trillion parameters, across a range of workloads and interactivity levels. Lenovo lists this tier from 25 days.
Lenovo says customers can also choose validated CPU technologies from AMD and Intel. Software options include Red Hat AI Factory with NVIDIA or NVIDIA AI Enterprise; Veeam Kasten is available as an option for protecting AI applications, data, models and pipelines.
Deployment services and timing conditions
The programme includes services covering use-case identification, proof-of-concept work and optimisation of GPU environments for performance, utilisation and cost efficiency. Lenovo also announced Premier Support Plus for Servers, with proactive and predictive support capabilities intended to identify issues earlier.
The listed timelines are conditional. Lenovo says the 15-business-day order-to-ship period applies to eligible configurations and selected parts in non-restrictive markets. The clock starts only after order validation, payment or credit clearance, End User Certification and any applicable due diligence or export approvals. Availability, eligibility and ordering processes vary by region.
The scale figures are also sizing guidance rather than fixed capacity guarantees: Lenovo says model size, user count and TPS vary with the model, workload, software and service-level agreement. Its published estimates are based on a Lenovo internal sizing tool. That qualification matters when comparing tiers, since actual performance depends on the specific deployment.