Riot Platforms (NASDAQ: RIOT) saw its shares climb approximately 14% in premarket trading on Tuesday after a Bloomberg report identified Anthropic as the "leading frontier AI" company behind the Bitcoin miner's newly announced $9.1 billion infrastructure agreement. The deal, disclosed on Monday, involves a 20-year contract to provide 191 megawatts of computing capacity from Riot's Rockdale, Texas campus—enough to power roughly 143,000 homes at any given time.
The agreement, which runs through June 2048, is expected to generate about $9.1 billion in revenue. It also includes two optional five-year extensions, potentially lifting the total contract value to as much as $16.1 billion. This marks a significant step in Riot's transformation from a cryptocurrency miner into an AI infrastructure provider, a pivot that has been gaining momentum across the sector.
AI Pivot Gathers Pace
Riot, formerly known as Bioptix—a maker of diagnostic equipment for the biotechnology industry—shifted its focus to Bitcoin mining years ago. Now, like several crypto miners facing changing market dynamics, it is expanding into cloud computing and AI infrastructure. The company has also signed a separate agreement with Advanced Micro Devices (AMD) to build AI computing infrastructure. Together, the Anthropic and AMD contracts cover 241 MW of contracted IT capacity at Rockdale.
Riot estimates that the two agreements will generate about $520 million in average annual revenue and between $416 million and $462 million in annual net operating income. The company expects to deliver the first 96 MW of capacity to Anthropic in December 2027, with the remaining capacity scheduled for June 2028. To support the buildout, Riot anticipates development costs of between $2.1 billion and $2.3 billion, or roughly $11 million to $12 million per IT MW.
Bitcoin Mining Still Matters, but AI Is Becoming Central
Bitcoin mining remains Riot's largest business, but AI infrastructure is becoming an increasingly important contributor to its long-term growth strategy. In the second quarter, revenue rose 14% year over year to $174 million, beating analyst expectations of $152.1 million. Bitcoin mining generated $113.7 million in revenue during the quarter, while Riot's data center business contributed $23.2 million, and its engineering segment added another $37.3 million.
The company mined 1,587 Bitcoin during the quarter, up from 1,426 a year earlier, although its average mining cost increased to $49,912 per Bitcoin. Riot ended June with $548.9 million in cash and holdings of 11,380 Bitcoin. The AI pivot is not unique to Riot; other miners have also struck similar deals. For instance, Hut 8's AI infrastructure partnership with Anthropic and Fluidstack involves developing at least 245 MW of AI data centre infrastructure, with potential capacity up to 2,295 MW.
Anthropic Ramps Up AI Infrastructure Partnerships
Anthropic has signed a series of infrastructure agreements in recent months to secure the computing power needed to support surging demand for its AI models. The company recently struck a $10 billion deal with infrastructure startup Volta Infra Holdings and, in May, agreed to purchase nearly $45 billion worth of computing capacity from Elon Musk's xAI. These deals highlight the intense competition for AI compute capacity, which has become a critical bottleneck for AI development.
For Riot, the deal with Anthropic provides a long-term revenue stream that diversifies its income away from the volatile Bitcoin market. The company's stock reaction reflects investor optimism about this strategic shift. However, the significant capital expenditure required—$2.1 billion to $2.3 billion—will test Riot's balance sheet, though its cash reserves and Bitcoin holdings provide some cushion.
As the AI infrastructure boom continues, Riot's pivot could serve as a template for other crypto miners looking to repurpose their energy-intensive facilities. The Rockdale campus, with its access to cheap power, is well-suited for high-performance computing. The deal also underscores Anthropic's aggressive expansion strategy, as it seeks to secure the massive computing resources needed to train and run its advanced AI models.
This article is for informational purposes only and does not constitute financial advice.
