Google is adopting Nvidia's strategy to challenge its dominance in the AI chip market.

      To create a formidable competitor to Nvidia, Google has turned to Nvidia's own strategies. An investigation by the Wall Street Journal reveals how Google is employing financial guarantees and what the industry refers to as “circular financing” — the very tactics that fueled the insatiable demand for Nvidia’s chips — in order to attract data center clients for its own silicon.

      The most prominent example can be found on the southern shore of Lake Ontario, just a short drive from Niagara Falls.

      The $3.2 billion guarantee at the falls

      At an AI data center cluster in western New York, dubbed Lake Mariner, Google has reportedly provided a $3.2 billion financial guarantee, according to the WSJ. The developers of the site, TeraWulf and the Google-supported cloud provider FluidStack, will lease computing power from thousands of Google’s tensor processing units (TPUs) to Anthropic. This guarantee enables the data center to secure cheaper debt, which is precisely how Nvidia has consistently increased demand for its own chips. “They want to be in the game; they don’t want to be left behind,” stated TeraWulf co-founder Nazar Khan to the publication.

      Circular financing, Nvidia’s method

      Another tactic being used is circular financing, where a portion of the money invested by a chipmaker returns to it through chip purchases. Google is supporting a series of Anthropic initiatives under this premise: a $7 billion data center named River Bend near Baton Rouge and an additional $1.4 billion in guarantees for a computing lease in Colorado City, Texas.

      This is built on an existing extensive compute agreement between Google and Broadcom and feeds into a roughly $35 billion private credit deal arranged by Apollo and Blackstone, which acquires Google TPUs and leases them to Anthropic.

      Going direct, and aiming for dominance

      Google is no longer keeping its chips for itself. In May, the company announced it would begin selling TPUs directly to customers and introduced its first chip designed specifically for inference, a rapidly growing task of handling AI queries rather than training models. Additionally, it secured a $5 billion deal with Blackstone to establish a cloud company directly targeting Nvidia-supported providers like CoreWeave and Nebius, and revealed this month plans to raise $85 billion in equity primarily for AI infrastructure.

      The proposal is making an impact: Citadel Securities, an early adopter, reports running some workloads at 30 percent lower costs and up to four times faster on TPUs.

      Nvidia remains unimpressed

      Jensen Huang has downplayed the threat. He stated in April that Nvidia’s “market reach is far greater than any TPU or ASIC can ever achieve,” contending that Anthropic is Google’s only major outside TPU customer and challenging Google to substantiate that its chips are more economical. Nvidia continues to command over 90 percent of the AI chip market, bolstered by its CUDA software and user-friendly hardware, and some smaller clouds express concerns about losing their Nvidia allocations, referred to as “Jensen jail,” should they diverge. Huang’s public remarks remain more relaxed: “Nothing brings me more joy than when you buy everything from Nvidia. But I am also very pleased if you simply buy something from Nvidia.”

      Why it matters

      Google is not the sole contender. AMD, Broadcom, and Cerebras are also in the mix, and Amazon is employing a similar strategy with its $50 billion Trainium business and substantial Anthropic clusters. However, none of them can rival Google’s financial strength, which is central to this competition.

      Under Amin Vahdat, who was elevated in December to oversee AI infrastructure, Google has ceased being cautious in its competitive approach, even as he asserts that “it’s not zero-sum” because “there’s so much demand out there.” When the wealthiest challenger starts emulating the precise financing flywheel that built Nvidia’s empire, Nvidia’s 90 percent advantage will face a significant test, and the expansion of AI will entail more debt-driven, circular risks.