Hut 8's $9.8B Beacon Point Lease Turns a Bitcoin Miner Into an AI Landlord
A single 352 MW lease in South Texas just rewrote the valuation story for an entire sector of the energy industry, proving that the most valuable asset in the AI boom isn't chips or code. It's power.
On May 6, 2026, Hut 8 Corp. announced a 15-year, triple-net lease at its Beacon Point campus in Nueces County, Texas, worth $9.8 billion at base value. The counterparty is unnamed but described as a high-investment-grade tenant planning to use the site for AI training and inference. By the end of that trading session, Hut 8 shares had surged more than 30%, touching all-time highs. At its intraday peak, the stock was up 37%.
What's actually happening here is bigger than one company's earnings call. Bitcoin miners built the country's most distributed network of cheap, large-scale power infrastructure, and now they're converting it, site by site, into the backbone of the AI economy. Hut 8 just signed the largest publicly disclosed example of that transformation to date. The deal doesn't just change the company's trajectory. It sets a pricing benchmark the entire industry will reference for years.
The Beacon Point announcement follows Hut 8's River Bend deal in late 2025, a 245 MW contract reportedly backed by Google valued at $7 billion. Combined, the company now carries $16.8 billion in contracted AI backlog across 597 MW. Those figures aren't projections. They're take-or-pay obligations from investment-grade counterparties.
The $9.8B Deal, Decoded
Strip away the headline number and the structure of the Beacon Point lease tells the real story. This isn't a letter of intent or a memorandum of understanding. It's a firm, 15-year commitment on a take-or-pay, triple-net basis, with no termination-for-convenience clause. The tenant pays regardless of whether it uses the capacity, and covers taxes, insurance, and maintenance on top of rent.
"We have a 15-year obligation from a high-investment-grade counterparty and the contract is structured on a take-or-pay, triple-net basis with no termination for convenience."
Asher Genoot, CEO, Hut 8 Corp. — Reuters, May 6, 2026
The base term runs 15 years at $9.8 billion. Three optional five-year renewals could push the total to $25.1 billion over 30 years. The contract includes a 3% annual escalator, which means cash flows compound in the company's favor throughout the life of the lease. Analysts at CryptoRobotics estimate stabilized annual revenue at roughly $655 million once the campus reaches full operation.
Deal snapshot: Beacon Point is the first phase of a 1 GW campus in Nueces County, Texas. The 352 MW IT load leased under this agreement is designed to NVIDIA's DSX reference architecture for gigawatt-scale AI training and inference. Water-smart cooling systems are specified. Power connection is targeted for Q1 2027; the first data hall comes online in Q3 2027.
The tenant identity remains undisclosed, which is standard for hyperscale infrastructure deals. The language used, "high-investment-grade," points toward a major cloud or AI firm with a balance sheet large enough to absorb a multi-decade, multi-billion-dollar commitment. The unnamed nature of the counterparty is less alarming than it sounds: the triple-net structure means Hut 8's revenue isn't contingent on the tenant's operational performance, only its creditworthiness.
| Deal | Capacity | Base Value | Term | Max Value |
|---|---|---|---|---|
| River Bend (Dec 2025) | 245 MW | $7.0B | 15 years | Not disclosed |
| Beacon Point (May 2026) | 352 MW | $9.8B | 15 years | $25.1B |
| Total Backlog | 597 MW | $16.8B | — | — |
From Mining to AI Landlord
Hut 8 didn't start 2025 as an AI infrastructure company. It started as a Nasdaq and TSX-listed Bitcoin miner, operating facilities at sites like Drumheller in Alberta. The pivot began in earnest as Bitcoin mining economics deteriorated to the point where the business model itself became untenable.
The core problem: miners currently lose an estimated $19,000 for every coin they produce, according to figures cited by CoinDesk's analysis from May 6, 2026. That figure reflects a combination of elevated energy costs, post-halving block rewards, and compressed hashprice. When your primary product is structurally unprofitable, the rational move is to repurpose your assets for something else.
The "something else," for miners with access to large blocks of cheap power, is AI compute. Hut 8 realized that what it had built, a network of permitted, power-connected sites with existing grid relationships, was precisely what AI hyperscalers were scrambling to acquire. The company didn't need to reinvent itself. It needed to redirect its existing infrastructure toward a higher-margin application.
"Beacon Point underscores why we start with power and maintain flexibility across end markets. Operating across multiple applications lets us underwrite assets that single-use-case developers cannot."
Asher Genoot, CEO, Hut 8 Corp. — CoinDesk, May 6, 2026
The Beacon Point campus itself illustrates this flexibility. Originally planned as a Bitcoin mining operation through Hut 8's affiliate American Bitcoin Corp., the site was redesigned for AI amid surging demand from model developers and cloud providers. The 1 GW of total utility capacity secured at the campus gives Hut 8 room to expand well beyond the initial 352 MW now under lease.
The Math Behind the Pivot
To understand why miners are rushing toward AI, you need to understand what the economics look like on both sides. Bitcoin mining is a commodity business where margins compress whenever the coin price drops or energy costs rise. AI data center leasing, structured the way Hut 8 has done it, is more akin to owning a toll road.
Bitcoin Mining
Estimated $19,000 loss per coin produced under current economics. Revenue tied directly to volatile hashprice and block rewards halved every four years.
AI Leasing
Take-or-pay, triple-net leases with investment-grade counterparties. $655M projected annual revenue at stabilization. 3% annual escalator built in.
Power as Moat
Permitted, grid-connected sites in Texas. Years of permitting and utility relationships translate directly into AI capacity that hyperscalers can't build fast enough themselves.
Sector Scale
Miners have signed more than $70 billion in AI contracts industry-wide. Some firms project up to 70% of revenue coming from AI by the end of 2026.
The triple-net lease structure does something particularly important for Hut 8's balance sheet: it eliminates most of the operating cost risk. Hut 8 isn't responsible for maintaining or operating the AI workloads running inside Beacon Point. The tenant handles that. Hut 8 functions as the landlord, collecting rent while partners like American Electric Power, Vertiv, and Jacobs handle development and engineering.
The $655 million annual revenue figure at stabilization, if it holds, would represent a dramatic step-change from anything the company could realistically achieve through mining. And unlike mining revenue, it doesn't disappear when Bitcoin's price drops 30% in a week.
Power-First Infrastructure
Genoot has described Hut 8's strategy as "power-first," and that framing isn't marketing language. It's a genuine operating thesis. Most data center developers start with a customer commitment and then go looking for land and power. Hut 8 does the opposite: it secures power first, then figures out the highest-value application for it.
"Hut 8 is a power-first infrastructure builder targeting long-duration, investment-grade leases."
Asher Genoot, CEO, Hut 8 Corp. — Q1 2026 Earnings Call, May 6, 2026
This approach works in Texas specifically because the ERCOT grid, despite its volatility, has more large-scale power available to industrial customers than most other states. Nueces County, on the Gulf Coast, also benefits from proximity to port infrastructure and a relatively mild coastal climate that reduces cooling loads. The facility's water-smart cooling design addresses one of the key sustainability concerns regulators and corporate tenants increasingly raise about large-scale AI infrastructure.
The NVIDIA DSX architecture specification matters too. DSX is NVIDIA's reference design for facilities meant to run dense GPU clusters at scale, the kind of hardware that trains large language models and runs inference at the volumes hyperscalers require. Designing to that standard from the outset means the facility doesn't need expensive retrofits when the tenant moves in equipment.
What is DSX? NVIDIA's Data Center Scale (DSX) reference architecture provides specifications for building AI data centers capable of hosting dense GPU clusters. It covers power delivery, cooling, networking topology, and structural requirements. Facilities built to DSX can accept the latest NVIDIA compute systems without modification, reducing deployment timelines for hyperscale tenants.
Execution Risks the Market May Be Underpricing
The 30-plus percent stock surge on announcement day reflects genuine investor enthusiasm. It also reflects the kind of forward-looking optimism that doesn't fully account for the gap between signing a lease and collecting rent. There are real execution risks here, and some analysts think the market is pricing in the upside before the downside is fully understood.
The most immediate risk is the construction timeline. Power energization is scheduled for Q1 2027 and the first data hall for Q3 2027. That's a two-year window where Hut 8 is building a facility it doesn't yet have revenue from, in an environment where construction costs and supply chain disruptions remain elevated. Any slippage in the schedule delays the cash flow that investors are already pricing into the stock.
Risk to watch: The ERCOT grid in Texas is subject to power curtailment events, particularly during extreme weather. A triple-net lease shifts operational costs to the tenant, but grid disruptions can affect facility uptime and strain relationships with counterparties who need reliable compute for AI training runs. Curtailment risk isn't unique to Hut 8, but it's a factor that traditional data center operators in other markets don't face at the same frequency.
There's also the question of what happens when the 15-year base term ends. The "up to $25.1 billion" figure assumes all three five-year renewal options get exercised. That's a 30-year assumption about the AI compute market, a market that didn't meaningfully exist a decade ago. The tenant has no obligation to renew, and by 2041, the technology landscape could look very different.
Critics of the valuation reset that accompanied the announcement note that Hut 8 still derives a portion of its revenue from crypto operations, that the company's stock had been depressed prior to the deal, and that a 37% intraday spike based on a forward-looking contract creates a high bar for the underlying business to actually clear. The deal is real. Whether the market's reaction is calibrated correctly is a separate question.
Industry Ripple Effects
Hut 8 isn't the only miner making this transition, and the Beacon Point deal will accelerate moves already underway across the sector. Core Scientific, one of the largest US miners, acquired the Polaris site for $421 million as part of its own AI expansion push. Riot Platforms has also been exploring data center monetization. The Beacon Point lease gives all of them a live pricing benchmark to reference in negotiations with potential AI tenants.
The broader implication for energy markets is significant. Stranded power, meaning large blocks of generation capacity that were built for mining but became uneconomical as mining profits collapsed, is now finding a new buyer class. AI firms desperate for gigawatt-scale compute can't wait the five to seven years it takes to permit and connect a greenfield data center campus. Miner sites, already connected and permitted, compress that timeline dramatically.
- Data center REITs are watching $/MW economics shift upward as AI-specific facilities command premium lease rates compared to traditional colocation.
- Energy sector investors are reassessing the value of industrial power contracts in Texas and other deregulated markets.
- Traditional capital, including institutional investors who previously avoided crypto-adjacent companies, is now looking more closely at miners-turned-AI-infrastructure players.
- AI model developers and cloud providers gain access to additional compute capacity that would take years to build from scratch, easing what has become a genuine supply constraint.
The $70 billion in AI contracts signed across the mining sector cited by CoinDesk suggests this isn't a niche phenomenon. It's an industry-wide reallocation of physical infrastructure assets toward a more profitable use case. Hut 8 is the most visible example right now, but the pattern is repeating at scale across the sector.
For the AI industry specifically, deals like Beacon Point matter because compute capacity has been the binding constraint on model development for the past two years. Every megawatt that comes online sooner, through repurposed mining infrastructure, is training time and inference capacity that wouldn't otherwise exist. The supply crunch isn't over, but it's easing, and former Bitcoin miners are a meaningful part of why.
Frequently Asked Questions
What is the Beacon Point AI data center lease?
Hut 8 Corp. signed a 15-year, triple-net lease at its Beacon Point campus in Nueces County, Texas on May 6, 2026. The lease covers 352 MW of IT capacity for AI training and inference, with a base contract value of $9.8 billion and potential up to $25.1 billion if renewal options are exercised.
Who is the tenant in the Hut 8 Beacon Point deal?
The tenant has not been publicly identified. Hut 8 describes the counterparty as a "high-investment-grade" entity, a designation that typically refers to corporations with credit ratings of BBB or above, suggesting a major cloud provider, technology company, or AI firm.
When will Beacon Point generate revenue for Hut 8?
Grid power energization is scheduled for Q1 2027, with the first data hall coming online in Q3 2027. Analysts project approximately $655 million in annual revenue at full stabilization. There is a roughly two-year construction gap before meaningful cash flows begin.
What does triple-net lease mean for an AI data center?
A triple-net lease means the tenant, not the landlord, pays property taxes, building insurance, and maintenance costs in addition to base rent. For Hut 8, this structure minimizes operating expense exposure while providing predictable rent income, making the arrangement similar to owning a commercial property with a creditworthy long-term tenant.
Why are Bitcoin miners pivoting to AI infrastructure?
Bitcoin mining has become structurally unprofitable for many operators, with estimated losses of around $19,000 per coin produced under current economics. Miners already have permitted, power-connected sites that AI hyperscalers need urgently. Repurposing those sites for AI leasing offers far higher and more stable margins than continuing to mine.
How does Beacon Point compare to Hut 8's River Bend deal?
River Bend, signed in late 2025 and reportedly backed by Google, covers 245 MW at a base value of $7 billion over 15 years. Beacon Point is larger at 352 MW and $9.8 billion. Together they give Hut 8 a $16.8 billion contracted AI backlog across 597 MW of total capacity.
What is NVIDIA's DSX architecture?
DSX is NVIDIA's reference architecture for gigawatt-scale AI data centers. It specifies power delivery, cooling, and networking configurations that allow facilities to host dense GPU clusters without custom modifications. Beacon Point is designed to DSX standards, making it immediately compatible with the compute hardware AI tenants deploy.
What are the main risks to the Hut 8 AI pivot?
Key risks include construction delays pushing back the Q3 2027 cash flow timeline, potential ERCOT grid curtailments affecting facility uptime, tenant non-renewal after the 15-year base term, and the possibility that AI compute demand softens if hyperscalers find chip efficiency improvements that reduce their need for raw infrastructure capacity.
What Comes Next
The Beacon Point lease isn't the end of Hut 8's transformation. It's the confirmation that the model works. Two investment-grade deals, two take-or-pay structures, $16.8 billion in contracted backlog, and 597 MW of capacity committed. The company now has a repeatable playbook: acquire or develop power-rich sites, design to AI hyperscaler specifications, and sign long-duration leases with creditworthy counterparties. The remaining 648 MW of Beacon Point's 1 GW total capacity represents the immediate next chapter.
For the broader infrastructure market, the deal validates something that wasn't obvious 18 months ago: that Bitcoin miners, widely dismissed as stranded-asset owners after the 2022 crypto collapse, built something genuinely valuable. Their willingness to absorb the permitting, utility negotiation, and grid interconnection work that conventional data center developers avoided has positioned them as critical suppliers to the AI economy. The $70-billion-plus in sector-wide AI contracts is the market's verdict on that assessment.
The harder question is whether the valuation catch-up has run its course or whether there's more repricing to come. Hut 8's stock is pricing in a lot of future cash flows that don't arrive until 2027 at the earliest. If construction stays on schedule and the tenant relationship holds, the fundamentals will eventually catch the stock price. If either slips, the gap between expectation and reality closes painfully. Investors who bought the 30% surge are betting that Asher Genoot and his team can deliver a data center on time, on budget, in Texas, for an unnamed AI giant, two years from now. That's a specific bet. Make it with open eyes.
