Nvidia acts as a "matchmaker": Anthropic and "AI Cloud" Lamda reach a $35 billion "computing power agreement"

The competition for computing power continues to intensify, and Nvidia’s role has far surpassed that of a mere chip supplier.

According to The Wall Street Journal on August 31, sources revealed that AI startup Anthropic has signed a $35 billion cloud computing deal with Lambda, a cloud computing service provider backed by Nvidia. Nvidia itself holds the lease for the relevant data center. This is Anthropic’s second multi-billion dollar computing power deal this year, and represents the latest example of Nvidia’s deep involvement in the AI computing power supply chain, acting as a “matchmaker.”

The transaction structure is quite complex: the data center is being developed by data center developer Hut 8 in Nueces County, Texas. Weeks ago, Nvidia signed an agreement with Hut 8 to secure the capacity, and Lambda will then use this site to deploy chips purchased from Nvidia to provide cloud computing power services for Anthropic. This arrangement allows Lambda to secure a high-value contract with Anthropic without having to obtain its own data center space.

Analysts believe that the significance of this deal lies in the market signal: Nvidia is extending its influence from hardware sales into the capital and commercial arrangements of the entire AI infrastructure ecosystem, while Anthropic’s urgent demand for computing power—driven by rapid product expansion—is reshaping the competitive landscape of the cloud computing market.

Anthropic’s Computing Power Crisis: Two Multi-Billion Dollar Deals in a Year

Earlier this year, Anthropic encountered bottlenecks in computing power supply just as its products were rapidly gaining market penetration. This triggered a round of intensive cloud capacity procurement.

Reportedly, sources revealed that just earlier this month, Anthropic signed a $45 billion agreement with another emerging Nvidia-backed cloud service provider Nscale to lease Nvidia computing resources in West Virginia.

The $35 billion agreement signed with Lambda is the second large-scale computing power lock-in Anthropic has completed in a short period. Combined, the two deals are worth $80 billion, highlighting the urgent demand for computing power among leading AI developers and the willingness to make long-term commitments to ensure supply.

Nvidia’s “Matchmaking” Logic: Credit Endorsement for Ecosystem Control

The core structure of this deal reflects Nvidia’s increasingly strengthened strategic layout within the AI computing power ecosystem.

Anthropic is not an investment-grade credit-rated company and faces obstacles in securing large-scale data center resources through traditional financing channels. By holding the data center lease, Nvidia effectively provides a credit endorsement for this deal, allowing Anthropic to bypass this limitation and procure the required computing power.

According to The Wall Street Journal, in July this year Nvidia announced a plan to provide credit support to emerging cloud service providers like Lambda in exchange for a share of their cloud revenue.

However, the report also noted that Nvidia has recently suspended certain deals under this program. It is unclear whether Lambda must share revenue generated at its Texas site with Nvidia, nor have details been disclosed about how much Lambda pays Nvidia for data center usage.

Hut 8: The “Dual-Line Player” Serving Both Nvidia and Google

In this transaction, the role of data center developer Hut 8 is particularly noteworthy—this bitcoin mining company is providing infrastructure for two competing camps in the same region.

In July this year, Hut 8 disclosed that a “high-investment-grade company” had signed a 15-year lease, renting the entire capacity of its 700-megawatt park in Texas, without revealing the tenant’s name. Hut 8 stated at the time that this $2 billion facility would be designed specifically to support Nvidia chips.

Meanwhile, Hut 8 is also separately developing another batch of data centers for Anthropic, which will be equipped with Google’s Tensor Processing Units (TPUs)—direct competitors to Nvidia chips. Google has provided a financial guarantee for the project, assisting Hut 8 in debt financing.

This means that in the same region, Hut 8 is providing computing power infrastructure for both the Nvidia camp’s Lambda-Anthropic deal and supporting Anthropic projects based on Google’s TPU ecosystem, making it a rare “dual beneficiary” in the current AI computing power arms race.

Analysts suggest that this series of deals indicates that emerging cloud service providers represented by Lambda and Nscale (neocloud) are leveraging Nvidia’s capital and credit support to quickly enter a market previously dominated by hyperscale cloud providers such as Amazon, Microsoft, and Google, and are securing long-term, stable, high-value contracts through direct partnerships with AI developers.

Nvidia’s role in this model has evolved from a pure hardware supplier to a driving force behind the “financialization of infrastructure” in the computing power ecosystem—by holding leases, providing credit support, and taking equity stakes in emerging cloud providers, Nvidia embeds its interests at multiple points in the AI computing power supply chain, creating more sticky sources of revenue beyond chip sales.

However, the report also points out that Nvidia’s recent suspension of some credit support plans for emerging cloud providers indicates that this model is still exploratory, and its sustainability and risk exposure remain to be seen.