NVIDIA Corp. is investing up to $3 billion in Lancium, a power infrastructure developer driving the massive Stargate data center project in Texas.
The move underscores the growing imperative for semiconductor manufacturers to secure dedicated energy and land resources required to run next-generation artificial intelligence (AI) workloads.
Under terms of the deal, initially reported by The Information, NVIDIA will inject an upfront $2 billion in exchange for roughly a 20% equity stake in the Blackstone-backed infrastructure firm. The semiconductor giant could commit an additional $1 billion if Lancium meets crucial operational milestones, including securing critical electrical grid connections. The transaction values Lancium and its portfolio of energy assets at an estimated enterprise value of $10 billion.
The capital infusion will support Lancium as it expands operations ahead of a potential initial public offering in 2027. The company’s primary asset — a 1,000-acre Clean Campus located in Abilene, Texas — serves as the initial operational foundation for the broader Stargate project.
“NVIDIA’s reported stake in Lancium, a power and land developer, tells buyers where AI capacity is actually gated,” said Mitch Ashley, vice president and practice lead for Software Lifecycle Engineering and AI-Native Software Engineering at The Futurum Group. “Chips are no longer the scarce part. Grid interconnects and land clear on utility timelines that no purchase order accelerates. Constraint relocation now reaches past software into megawatts. Buyers signing multi-year AI capacity deals inherit that physical risk without seeing it. Ask for the interconnect date and the remedy when it slips.”
Unveiled by President Donald Trump as part of a sweeping national strategy to build computing infrastructure, Stargate is an ambitious joint venture involving SoftBank, OpenAI, and Oracle Corp. The participating partners have outlined plans to direct up to $500 billion toward constructing massive AI data center hubs across the United States.
The transaction highlights a strategic shift for NVIDIA as it works to address key logistical bottlenecks in AI expansion, including power generation and grid connectivity. Speaking at a private industry gathering of financial institutions in Taipei, NVIDIA CEO Jensen Huang stressed that financial returns on AI infrastructure investments have proven profitable at scale, silencing earlier skeptics.
“You need land, power, you need energy, but you also need financing,” Huang said, framing wealthy family offices as an emerging source of capital alongside institutional investors.
The investment fits into NVIDIA broader infrastructure push, which includes a recent $2 billion commitment to chip designer Marvell Technology Inc.
Boasting a market capitalization of approximately $5.42 trillion, NVIDIA continues to leverage its dominance in graphics processing units (GPUs), networking solutions, and its proprietary CUDA software platform. By directly funding power developers, NVIDIA aims to guarantee that its hardware is supported by the massive energy capacity needed to train and run complex AI models.
Neither NVIDIA nor Lancium offered immediate comment when contacted regarding the deal.
