VC
Value Add VC
⚡HomePulse⚡Helpful Apps📝Blog🤝Partner
Illustration for: Riot Platforms Signs $9.1B, 20-Year AI Compute Deal With Anthropic
Value Add VC/Pulse/AIFOLLOW-UP

Riot Platforms Signs $9.1B, 20-Year AI Compute Deal With Anthropic

Bitcoin miner Riot Platforms struck a 20-year, $9.1 billion compute deal with Anthropic that leases 191 megawatts of its Texas power campus, the AI lab's third major infrastructure commitment in three months totaling more than $60 billion.

By the Numbers

$9.1B over 20 years
Deal value
~$16.1B (with options)
Extended value
191 MW
Power leased
241 MW
Riot total contracted
+20-25%
Stock move
TC
By the AI Desk
Edited by Trace Cohen · Early-stage VC & angel · Founder, New York Venture Partners
August 11, 2026
2 min read
ShareXLinkedInEmail
TC

The VC Read · Trace's Take

Trace Cohen

$60B in Anthropic compute contracts across three separate deals in three months is the number that matters more than any single agreement -- it's a bet Anthropic's revenue and fundraising can support two-decade obligations in a field where the model leaderboard has flipped multiple times in the last two years. For Riot, this trades bitcoin-price risk for Anthropic-solvency risk; that's a real diversification, not a risk-free pivot. Watch whether other large miners sign comparable deals in the next two quarters -- that confirms the playbook, not just this one contract.

Analysis

The Deal

Bitcoin miner Riot Platforms struck a $9.1 billion, 20-year compute deal with Anthropic, leasing 191 megawatts at its Rockdale, Texas campus, according to Bloomberg and CNBC. The agreement is worth an estimated $9.1 billion in revenue over its base term, rising to roughly $16.1 billion if extended for two additional five-year periods. Riot shares surged more than 20% following the announcement.

What Changed

Combined with a prior AMD agreement, Riot now has 241 megawatts contracted and roughly $9.8 billion in long-term revenue on the books -- transforming a bitcoin miner into what amounts to a contracted AI infrastructure landlord generating an estimated $520 million in average annual revenue at full delivery. That's the same pivot Riot Platforms signs deal with Anthropic as bitcoin miners shift to AI infrastructure describes across the sector broadly: crypto miners sitting on grid-connected power capacity are increasingly finding that leasing it to AI labs is more lucrative than mining bitcoin with it.

Anthropic's Compute-Buying Spree

This is Anthropic's third major infrastructure commitment in roughly three months:

  • Volta -- emerged from stealth in early August with a $10 billion, six-year Anthropic compute deal
  • Macquarie and GIC -- tapped by Anthropic to help build out its Theseus data center infrastructure
  • SpaceX -- agreed in May to nearly $45 billion to secure compute resources

Across those deals, Anthropic has now locked in more than $60 billion in total compute contracts in three months -- a pace of infrastructure commitment that outstrips even Anthropic's own aggressive fundraising cadence.

Why Power, Not Chips, Is the Scarce Resource

What all of Anthropic's recent deals share is a focus on securing grid-connected power capacity specifically, not just GPU supply. Riot's value to Anthropic here isn't proprietary technology -- it's 191 megawatts of power already connected to the Texas grid, the kind of scarce, permitted infrastructure that takes years to build from scratch. Bitcoin miners like Riot spent the last several years securing exactly that kind of power access for a different purpose, and AI labs are now paying a premium to redirect it.

The Counterweight

Locking in $60 billion-plus in compute contracts over three months is a bet that Anthropic's own revenue growth and continued fundraising can support obligations at this scale for the next two decades -- a length of commitment that assumes both AI compute demand and Anthropic's own competitive position hold for twenty years, an unusually long horizon in a field where the leading model provider has changed multiple times in just the last two years. Riot's own transformation into an AI infrastructure landlord also means its future is now tied to Anthropic's continued solvency and demand for compute, a concentration risk that trades bitcoin-price volatility for AI-demand volatility rather than eliminating volatility altogether.

Ahead

Watch how many other bitcoin miners follow Riot's exact playbook over the next two quarters -- if multiple large miners sign comparable AI compute deals, it will confirm this is now the default strategy for miners sitting on excess power capacity, not a one-off deal specific to Riot's Texas footprint.

ShareXLinkedInEmail

More on

Anthropic →

Reported by Bloomberg · First reported by CNBC · Analysis by Value Add Pulse.

← Back to Pulse

THE WIRE in your inbox— Tech, startup & VC news with Trace's take. Free, no spam.

Read Next

AI· Aug 12, 2026

CoreWeave Revenue More Than Doubles, Stock Jumps 18%

Illustration for: CoreWeave Revenue More Than Doubles, Stock Jumps 18%
AI

CoreWeave Revenue More Than Doubles, Stock Jumps 18%

CoreWeave reported second-quarter revenue of $2.6 billion, up 112% year-over-year, and raised its full-year guidance on a backlog that has swelled past $104 billion, sending shares up 18% in premarket trading.

AI· Aug 11, 2026

Gemini Hits 1 Billion Monthly Users, Catching Up to ChatGPT

Illustration for: Gemini Hits 1 Billion Monthly Users, Catching Up to ChatGPT
AI

Gemini Hits 1 Billion Monthly Users, Catching Up to ChatGPT

Google's Gemini app crossed 1 billion monthly users, its fastest-growing product ever, closing ground on ChatGPT even as OpenAI reports its own product separately hit 1 billion weekly users on a different growth curve.

AI· Aug 11, 2026

Decagon Crosses $100M ARR Betting Against Forward-Deployed Engineers

Illustration for: Decagon Crosses $100M ARR Betting Against Forward-Deployed Engineers
AI

Decagon Crosses $100M ARR Betting Against Forward-Deployed Engineers

Decagon, a three-year-old AI customer-service agent startup valued at $4.5 billion, crossed $100 million in annualized revenue while deliberately avoiding the forward-deployed engineer model rivals like Sierra use to customize deployments.

@Trace_Cohen·t@nyvp.com