Lenovo AI Express cuts GPU server order-to-ship to 15 days
Lenovo has launched AI Express, a streamlined procurement and deployment programme designed to accelerate enterprise adoption of on-premises AI infrastructure. Built in partnership with NVIDIA, the programme offers three pre-validated server configurations spanning small, medium and large deployments, with order-to-ship timelines starting at 15 business days for the entry-level option.
The announcement was made at an event in Dubai and positions AI Express as an on-ramp to Lenovo's broader Hybrid AI Factory architecture. The company cited its own 2026 CIO Playbook, which found that 94 per cent of enterprise respondents expected a positive return on AI investment, with an average anticipated return of $2.78 per dollar spent. Ashley Gorakhpurwalla, President of Lenovo's Infrastructure Solutions Group, said the programme is designed to reduce the risk of overbuilding while giving customers a defined path to scale.
The configurations
The three tiers cover materially different workload profiles. The Small configuration runs on a ThinkSystem SR650a V4 server fitted with two NVIDIA RTX 6000 PRO Blackwell Server Edition GPUs, supports models in the 7B to 70B parameter range, and targets tens of concurrent users at throughput of 30 or more tokens per second. The Medium tier uses a ThinkSystem SR675 V3 with eight of the same RTX 6000 PRO GPUs, handles models up to 400B parameters, and is positioned for agentic AI workloads serving hundreds of users, with an order-to-ship time from 20 business days. The Large configuration is built around the ThinkSystem SR680a V4 and NVIDIA HGX B300 accelerators, targeting trillion-parameter models and thousands of users, with delivery from 25 business days.
All three tiers support AMD and Intel CPUs. Customers can layer in enterprise AI software, including Red Hat AI Factory with NVIDIA or NVIDIA AI Enterprise, and optionally add Veeam Kasten for AI data and pipeline protection. Chris Marriott, Vice President of Enterprise Platforms at NVIDIA, said the programme combines Lenovo's validated hardware with NVIDIA's accelerated computing and software stack into ready-to-ship solutions.
Market context
The announcement lands in a crowded market for enterprise AI infrastructure. Dell Technologies, HPE and Supermicro all offer validated AI-ready server platforms, and hyperscalers continue to compete for the same workloads through managed cloud services. The specific differentiator Lenovo is stressing is supply-chain predictability: the 15-to-25-day order-to-ship window is positioned as a counter to the lengthy lead times that have plagued GPU server procurement since demand accelerated in 2023.
Channel partners are also a central part of the go-to-market. Lenovo is extending AI Express through its Lenovo 360 partner framework, building on an earlier programme called Top Choice Express that applied similar logic to mainstream infrastructure. Ilenia Lupi, Lenovo's Europe Channel Leader, said the programme gives partners a repeatable route to market without the complexity of configuring bespoke solutions for each customer. Rohil Dattani, Managing Director of UK reseller Ballicom, is quoted in the release citing greater consistency and predictability in bundling Lenovo devices into wider customer solutions.
Regulatory and procurement considerations
Enterprises deploying on-premises AI infrastructure of this scale will need to address a growing body of compliance obligations. The EU AI Act's requirements for high-risk AI systems apply to the applications running on this hardware rather than the hardware itself, but organisations in regulated sectors such as financial services or critical infrastructure may face additional scrutiny over data-residency, model governance and audit trails. Lenovo's emphasis on private, on-premises deployment does give customers stronger data-sovereignty controls compared with public-cloud alternatives, a consideration that is becoming increasingly significant under NIS2 and sector-specific frameworks.
The company also introduced Premier Support Plus for Servers alongside the programme, adding proactive and predictive support capabilities for mission-critical environments. Named customer deployments and publicly disclosed inference benchmark results would meaningfully strengthen the commercial case as AI Express moves beyond launch.