Lancium and Nvidia announced a partnership to develop gigawatt-scale AI data centers.

Lancium is partnering with Nvidia to deploy the chipmaker’s AI factory technology across a portfolio that the Texas-based infrastructure developer says includes 4 GW of leased capacity and more than 15 GW of powered land in development. The partnership establishes Nvidia’s role in designing and deploying AI infrastructure across Lancium’s footprint. At its campuses, Lancium will implement Nvidia’s DSX reference designs and power-management technologies, including systems designed to increase compute density and modulate AI factory power consumption in response to grid conditions. Nvidia is also making a strategic investment in Lancium, a company backed by Blackstone. The parties declined to disclose the investment amount.

Lancium, which refers to its facilities as “clean campuses,” stated they will serve as deployment sites for Nvidia’s full-stack AI factory platform, encompassing accelerated computing, networking, and software. The companies indicated the arrangement will provide cloud providers, infrastructure developers, and AI firms within Nvidia’s ecosystem access to power-ready capacity at the gigawatt scale. While the scale of Lancium’s portfolio is central to the announcement, its two headline figures represent distinct stages of development. The developer reports 4 GW of capacity under lease and more than 15 GW of powered land in development. Lancium did not identify the specific projects underlying these figures, nor did it provide timelines for bringing the full portfolio online.

Lancium’s publicly announced projects include its 1.2 GW Clean Campus in Abilene, where Crusoe is developing AI data center capacity as part of the Stargate initiative; a 1 GW campus in Childress County, also developed with Crusoe; and a new campus near Turkey in Hall County, where QTS will design, construct, and operate the data center buildings. Lancium retains ownership of the campuses and oversees their electrical and civil infrastructure. For the Hall County campus, Lancium and QTS committed to funding all energy infrastructure improvements, with Lancium planning to supply power via battery storage and solar generation. These projects are already drawing billions in planned investment. QTS and Lancium estimate the Hall County campus will inject over $10 billion in capital into the region. Additionally, Lancium secured a $600 million debt financing package in 2025 to advance its Clean Campus strategy, starting with the Abilene site. In Abilene, a 2024 joint venture among Crusoe, Blue Owl Capital, and Primary Digital Infrastructure was capitalized at $3.4 billion to fund more than 200 MW of build-to-suit data center capacity at the Lancium campus. These figures are not additive measures of Lancium’s total investment; rather, they reflect distinct projects, financing structures, and development stages.

However, “powered land” does not necessarily equate to grid-deliverable load on par with leased capacity, noted Neil Osnato, founder of Persistence Analytics Group. “Four gigawatts described as ‘under lease’ suggests a materially stronger commercial commitment than 15+ GW of ‘powered land in development,’” Osnato said. Regarding the larger figure, critical questions remain regarding the capacity of executed interconnection paths, the scope of studied and required infrastructure, the timeline for energizing each tranche, and the volume of committed customer demand. “A large development pipeline should not automatically be read as 15 GW of executable load,” Osnato said. Developers routinely secure power for AI campuses years before construction is complete. Securing land, generation resources, and interconnection paths establishes a development position without guaranteeing an equivalent volume of immediately energizable load.

Lancium is also marketing power flexibility as a core campus value. The company will deploy Nvidia’s DSX MaxLPS to optimize GPU power allocation, potentially enabling higher GPU densities within existing facility power budgets, according to Matt Kimball, vice president and principal analyst for data center technologies at Moor Insights and Strategy. The cited “up to 40%” efficiency gain represents an ideal-case scenario rather than a guaranteed improvement, Kimball noted. The primary benefit lies in eliminating the need to reserve peak-rated power for each GPU when typical workloads demand significantly less. For example, if a GPU is rated at 1 kW but operates at 600 watts, conventional power allocation leaves 400 watts of capacity idle. DSX allocates power based on actual workload demands, freeing that capacity for other facility operations. “It’s the ability to better utilize the incoming power” that holds greater significance than the 40% figure, Kimball said. At a 1 GW facility, a 20% improvement in power utilization would yield 200 MW of potential capacity for additional GPUs, while a 10% improvement would generate 100 MW, Kimball calculated. Actual outcomes will depend on specific workloads and operating conditions. Kimball emphasized that the approach is directionally significant because it aligns compute workloads more closely with available facility power, moving away from treating each GPU’s maximum rated consumption as a fixed requirement.

For a gigawatt-scale campus, the grid value of this flexibility depends entirely on practical execution, Osnato observed. To qualify as a useful grid resource, the system must reduce or reshape consumption during grid stress, delivering responses that are fast, measurable, dependable, and available during critical system conditions. “The question is not whether the software can technically move load; it is how much load can move, for how long, how often, under whose control, and what operating constraints remain,” he said. This capability could fundamentally alter how utilities plan for large AI loads. A genuinely flexible gigawatt-scale customer presents a distinctly different planning challenge than an inflexible one, Osnato noted. However, utilities should not assume that technically available flexibility can replace investments in generation or transmission infrastructure. If a utility relies on a data center’s flexibility to defer infrastructure spending, the capability must be measurable, reliably available when needed, and subject to ongoing revalidation as campus configurations, workload mixes, and operating economics evolve.

Lancium stated its campuses will integrate grid interconnections with behind-the-meter generation and energy storage. Its power-management systems are engineered to allow data centers to respond dynamically to grid conditions while preserving the compute density demanded by AI workloads. The Nvidia partnership provides Lancium with a scalable technology platform for its expanding portfolio, while offering Nvidia’s customers and infrastructure partners an additional pathway to large-scale AI capacity. “AI factories are the essential infrastructure of this new industrial era,” said Nico Caprez, Nvidia’s vice president of global AI infrastructure growth. Michael McNamara, Lancium’s CEO and co-founder, stated the company has spent years cultivating the power, land, and infrastructure expertise required to develop gigawatt-scale AI data centers. The companies declined to specify which Lancium campuses will initially deploy Nvidia technology. For Lancium, the more consequential test will arrive as these projects transition from development into interconnection and active operation. “Announced capacity is not executable capacity, and technically flexible load is not the same as dependable grid capacity,” Osnato said. If Lancium can demonstrate both durable load and verifiable flexibility, its model could deliver meaningful grid benefits, he added. However, validation must track each project from announcement through interconnection, energization, and operation, rather than relying on early assumptions.

Shane Snider is Senior News Writer at Data Center Knowledge, covering AI infrastructure, hyperscale data centers, cloud platforms, and the power and energy systems driving modern compute expansion. His reporting focuses on the operational, economic, and environmental forces reshaping digital infrastructure, including AI factories, utility constraints, liquid cooling, renewable energy procurement, and next-generation data center architectures. He has won recent Azbee awards for news series and government reporting. Based in Raleigh, North Carolina, Snider covers how hyperscalers, utilities, chipmakers, and infrastructure providers are responding to the rapid rise of AI workloads and global compute demand.