Nvidia (NVDA.US) invests $3 billion in Blackstone's Lancium: From chip supremacy to electricity “contractor”, AI energy landscape is next

The Zhitong Finance App learned that as the focus of the global AI competition shifts from computing power to electricity, NVDA.US (NVDA.US) is using practical actions to redefine the boundaries of “AI infrastructure.” According to reports, Nvidia has agreed to invest up to $3 billion in Lancium, a power infrastructure developer supported by Blackstone Group, to acquire about 20% of the shares in the company behind the “Stargate” (Stargate) data center park. The deal extends Nvidia's AI footprint directly from the chip layer to the energy infrastructure layer — in an age where AI chip power consumption is rising exponentially, locking in electricity is becoming a more pressing strategic proposition than locking in customers.

The deal marks Nvidia's historic expansion from its core chip business to the bottom of the AI value chain — energy infrastructure. The bottleneck of AI has shifted from computing power to the electricity required to run computing power.

Who is Lancium? The electricity operator behind “Stargate”

Lancium was founded in 2017 and is headquartered in Houston, Texas. It was co-founded and CEO by Michael McNamara. The company initially focused on energy transition technology, and its core product, Smart Response software, enables clean parks to function as controlled load resources (CLR) for power grids. The early business covered energy-intensive applications such as Bitcoin mining, then gradually transitioned to focus on AI data center power infrastructure.

Lancium's most critical asset is located in Abilene (Abilene), Texas — a 1,000-acre “Lancium Clean Campus” that was the first operational base for the “Stargate” (Stargate) program. The park is planned to house approximately 400,000 Nvidia AI chips distributed across eight buildings. Lancium has now locked in and developed 4 gigawatts of electricity on the Texas grid, and another 15 gigawatts to be connected project is underway.

Anatomy of a transaction: strategic upgrade from “selling chips” to “locking in electricity”

The structure of the deal clearly reflects Nvidia's strategic intentions.

Phase 1: Nvidia invested 2 billion US dollars and acquired about 20% of Lancium's shares. Based on this calculation, Lancium's corporate value is valued at around $10 billion, including debt.

Phase 2: Once Lancium's multiple parks meet established conditions such as grid access, Nvidia will invest an additional 1 billion US dollars, and the shareholding ratio can rise to a maximum of about 30%.

This investment is a pure equity investment and does not involve credit guarantees for data center construction and leasing. However, people familiar with the matter revealed that Nvidia is considering providing credit guarantees for other projects, including a large 10-gigawatt Ohio data center project involving OpenAI and SoftBank.

Lancium's asset portfolio is already quite impressive. The company has secured a contract for 4 gigawatts of electricity in Texas, which includes:

1.2 gigawatts to supply OpenAI and Oracle at the “Stargate” park in Abilene;

900 megawatts for Microsoft's new data center built by partner Crusoe in Abilene;

1 GW was contracted by QTS Data Centers in Turkey, Texas;

1 GW is being built by Crusoe at Childress.

In addition to the locked assets, Lancium has also laid out land in Texas that could potentially carry 15 gigawatts of electricity and is awaiting approval to connect to the grid.

Lancium's Abilene “Clean Park” covers an area of 1,000 acres, has 1.2 gigawatts of grid-connected facilities, and has received full approval from the Texas Electric Reliability Commission (ERCOT). The park was the first operating base for the “Stargate” program.

AI's “First Principle”: From Shortage of Computing Power to Shortage of Electricity

The AI “five-layer cake” theory proposed by Nvidia CEO Hwang In-hoon divides the AI industry into five levels from the bottom up: energy, chips, infrastructure and system software, AI models, and applications. Among them, energy is defined as the “primary principle” of AI infrastructure.

This theory is becoming a reality. Hwang In-hoon recently estimated that the cost of building a 1 gigawatt data center could soon reach 80 billion to 100 billion US dollars. As AI clusters move from 10,000 kilos to 100,000 or million calories, the bottleneck in power supply is more difficult than chip supply — the electricity demand for a hyperscale data center is equivalent to the electricity consumption of a medium-sized city, and grid upgrades and connections often take years. Lancium is the key player in this pain point.

From “Stargate” to investment map: Nvidia is becoming an infrastructure company

The Lancium deal is the latest footnote to Nvidia's accelerated transformation into a core investor in the AI infrastructure ecosystem. According to reports, in the quarter ending April of this year, Nvidia invested a total of 18.6 billion US dollars in private enterprises and infrastructure funds, exceeding the total amount invested in the previous full year.

Since this year alone, Nvidia has issued checks of $2 billion each to cloud computing service providers CoreWeave and Nebius, and completed investments in network equipment companies Lumentum, Coherent, and semiconductor company Marvell. This series of actions shows that Nvidia is accelerating its transformation from a chip seller to a core investor in the AI infrastructure ecosystem.

In the clean energy sector, Nvidia has previously invested in the nuclear fusion company Commonwealth Fusion Systems and the nuclear energy company TerraPower founded by Bill Gates.

Nvidia has a clear commercial logic for such investments: pre-empting scarce power resources for its large chip customer base. When customers need to build AI data centers, Nvidia can not only provide chips, but also guarantee electricity supply through the power infrastructure company it invests in — a strategic upgrade from “selling shovels” to “controlling water sources.”

Nvidia's approach to the energy sector is far from an exception. In 2025, Nvidia took steps more than 10 times in the fields of clean energy development, grid management, and data center construction, including investing in the nuclear fusion company Commonwealth Fusion Systems and TerraPower, a nuclear energy company founded by Bill Gates. From nuclear fusion to grid access, Nvidia is building an investment portfolio covering the entire AI energy chain — Lancium is the latest implementation of this strategy, but it is far from the end.

Risks and variables: Texas governor's “grid moratorium”

Just before and after news of the deal broke, a potential variable surfaced. Texas Governor Greg Abbott ordered the suspension of grid access applications for the new data center this week, requiring the relevant projects to undergo further review.

This policy trend has created uncertainty about Lancium's expansion timeline. However, people familiar with the matter pointed out that Lancium's 15-gigawatt project to be connected is in a relatively favorable position — the review process for its partner power company American Electric Power has entered an advanced stage.

The broader risk is that the power demand for AI infrastructure is facing social resistance and regulatory scrutiny. From an environmental protest at Google's $15 billion data center in Andhra Pradesh, India, to the suspension of power grid access in Texas, AI's “power hunger” is evolving from a technical issue to a social governance issue.