Riot Platforms: Valued on a Letter, Not a Lease

The Setup

Riot Platforms is a Bitcoin miner that the market has decided to value as an artificial intelligence landlord. The shares have climbed roughly 93% this year and delivered a total return north of 200% over the trailing twelve months, pushing the market cap to around $10 billion. Wall Street has chased it the whole way up: in the first week of June alone, Clear Street lifted its target to $38, Keefe Bruyette went to $37, and Bernstein moved to $30, each citing the same thing, the company's pivot from mining into AI and high-performance computing data centers.

This is not a piece about whether AI infrastructure matters. It does, and grid-connected power is genuinely the binding constraint on the whole buildout. Nor is it an argument that Riot's land and power assets are worthless. They are real and they are valuable. This is a narrower argument: the price has been set by a story about what those assets could become, and the gap between that story and what the company has actually built, leased, and earned is wide enough to matter.

The cleanest way to see the gap is to put two numbers next to each other. One is the valuation an activist investor has publicly placed on the AI pivot. The other is the recurring revenue that pivot generated in the most recent quarter. They are not close. We covered the sibling version of this setup last week in the IREN avoid thesis: a miner that has actually started pivoting, priced for perfection. Riot is the variation where the price runs on the promise rather than the progress.

The Letter That Moved the Stock

The bull case deserves a fair and full hearing before the numbers complicate it, because it is built on a genuinely good insight. In February 2026, Starboard Value, then Riot's fourth-largest shareholder with roughly 12.7 million shares, published a letter to management arguing that Riot was sitting on some of the most valuable undeveloped data center assets in the United States. The reasoning was sound. Riot owns land with the one input that cannot be conjured quickly: power connectivity at scale. Corsicana, Texas has about 400 megawatts built toward an ultimate 1 gigawatt target. Rockdale carries roughly 700 megawatts of developed capacity. Kentucky targets another 232 megawatts.

Starboard's conclusion was the headline that moved the stock: the AI pivot could be worth up to $21 billion. At the time the letter circulated, Riot's entire market cap was around $6 billion. The argument was that the market was valuing Riot as a crypto miner while its Texas power sites were worth a multiple of the whole company if redeployed for AI. Data center developers spend years waiting in interconnection queues for power Riot already controls, and that scarcity is real. It is the most credible bull thesis in the mining sector, and it came from a credible activist with a real position and a seat at the table.

That number, $21 billion against what is now a roughly $10 billion company, is doing almost all of the work in the stock. It is a projection of what the assets could be worth if the data center buildout fills with paying tenants. It is not a description of the business that exists today. The rest of this article is what happens when you set the letter down and open the filings.

A Letter, Not a Lease

Riot reported first-quarter 2026 results on April 30, and the data center segment is where the story meets the income statement. The segment generated $33.2 million in revenue, a number bullish writeups cite as proof the pivot is working. Read the composition and it tells a different story. Of that $33.2 million, about $32.2 million came from tenant fit-out services, the one-time construction work of building out space. The actual recurring operating lease revenue, the contracted income that an AI landlord earns quarter after quarter, was roughly $0.9 million.

That is the whole thesis in one line. The recurring data center business that is supposedly worth $21 billion is, as of the last filing, generating under $1 million a quarter. The contracted footprint behind it is a single anchor tenant. During the quarter, AMD (AMD) exercised an option for an additional 25 megawatts, bringing its total contracted capacity to 50 megawatts, and Riot delivered its first 5 megawatts of that. Fifty megawatts is a real start. It is also about 5% of the 1 gigawatt Corsicana target the valuation leans on.

The distinction matters because the market is treating a buildout as though it were a book of leases. Fit-out revenue is lumpy project work that ends when construction ends. A signed, multi-year operating lease is the recurring, high-margin income that justifies an infrastructure multiple. Riot has a great deal of the former booked and almost none of the latter flowing. The announced $400 million Corsicana expansion is more of the same: it is capital going out the door to build capacity, not revenue coming in from capacity that is full. The story is a gigawatt of leased AI compute. The lease is fifty megawatts contracted, five delivered, and under a million dollars of recurring revenue to show for it.

The P&L Is Moving the Wrong Way

Total first-quarter revenue was $167.2 million, which did beat the analyst estimate near $131 million. Below the top line, the picture darkens. Riot reported a GAAP net loss of $500.5 million, or $1.44 per diluted share, and an adjusted EBITDA loss of $311 million. For the full year 2025, the company lost $663 million, and the accumulated deficit has climbed to roughly $1.85 billion.

A fair reading has to acknowledge what drove the quarterly loss. A large part of it was non-cash: a mark-to-market adjustment on Bitcoin holdings of $326.7 million and depreciation and amortization of $97.7 million. Strip those out and the loss is smaller. But stripping them out does not rescue the operating picture, because the adjusted EBITDA loss of $311 million already excludes them, and it is still a loss of more than a quarter of a billion dollars. The business is not marginally unprofitable on an accounting technicality. It is deeply unprofitable on its own preferred measure.

The mining economics underneath are the uncomfortable part. Riot disclosed that its fully-loaded cost to mine one Bitcoin, including depreciation, was $96,283 in the quarter, against a production value of $75,964 per coin. That is 127%: it cost Riot more to produce a Bitcoin than the Bitcoin was worth when it was mined. The core engine of the company is running at a loss on a fully-loaded basis, and that is the engine being asked to fund the data center dream.

It is funding it, but look at how. Riot sold 3,778 Bitcoin for $289.5 million during the quarter to pay for the buildout, drawing down the crypto stack to finance the pivot. Total liabilities jumped from roughly $78 million to about $1.04 billion in a single quarter, and the share count continues to drift upward. None of this is fatal on its own. But it is the opposite of the clean, self-funding infrastructure operator the valuation imagines. It is a money-losing miner liquidating reserves and taking on obligations to chase a business it has barely begun to earn.

The Bitcoin Dependence

The activist sum-of-the-parts case assigns the overwhelming share of Riot's value to AI and data centers. The income statement tells you where the money still comes from. Of the $167.2 million in first-quarter revenue, $111.9 million, roughly two-thirds, was Bitcoin mining, and that figure fell from $142.9 million a year earlier as Bitcoin prices softened and network difficulty rose. The company also held 15,679 Bitcoin at quarter-end, so the value of the balance sheet and the swing factor in reported earnings both move with the price of a single volatile asset.

That dependence is not abstract this month. In early June, Bitcoin fell to a four-month low on a wave of liquidations, and even the largest corporate holder sold coins for the first time in years. When the token drops, three things happen to Riot at once: mining revenue compresses, the Bitcoin on the balance sheet marks down, and the cash generated by selling coins to fund the data center buildout buys less. The AI story is supposed to make Riot independent of crypto cycles. As of the last filing, the crypto cycle still drives the revenue, the earnings, and the financing.

This is the heart of the disconnect. A valuation that treats Riot as an AI infrastructure pure-play is pricing away the very exposure that still determines its results. The pivot may eventually earn that reframing. It has not earned it yet, and a soft Bitcoin tape is a live reminder of which business the company actually is today.

What the Wealth Engine Scores Say

Before the valuation verdict, here is what the Wealth Engine Pro platform's systematic scoring shows for Riot right now.

The platform rates Riot as Weak with a Bearish outlook and a blended fair value of $4.02, a level that would require the stock to fall roughly 84% from its current price near $24. Put another way, the market is paying about six times the platform's estimate of fair value. The low Financial Health score (34 of 100) reflects the losses and the rising liabilities, and the modest Moat (5 of 15) and Growth (7.5 of 15) scores describe a capital-intensive business with thin durable advantages and revenue that is contracting in its largest segment.

These scores are systematic. They evaluate companies based on reported financials, balance sheet quality, moat characteristics, and valuation models (discounted cash flow, peer comparison, earnings power). They measure what a company is today, not what it might become. That is by design: the scoring system is built to keep emotion and forward speculation out of the numbers.

In this case, the editorial thesis and the platform scores point in the same direction. The article's argument is that the price has run far ahead of the realized business, and the systematic data, which cannot price an unbuilt data center campus until the lease revenue and earnings appear in filings, says the same thing from a different angle. When both the quantitative model and the qualitative analysis flag the same gap between price and substance, that convergence is worth paying attention to. Research Riot yourself on the platform and decide which signal matters more for your situation.

The Valuation Verdict

Strip away the labels and the valuation question is simple. Riot carries a roughly $10 billion market cap on full-year 2025 revenue of about $647 million, revenue that is still roughly two-thirds declining Bitcoin mining, against a full-year net loss of $663 million. The entire bridge from that reality to the valuation is the activist projection of what the Texas power sites could earn if the data center buildout fills with tenants.

There is a respectable bull-case math here, and it deserves to be stated fairly: if Riot leases out a gigawatt of capacity at infrastructure margins, the stock is not expensive at all, and today's price will look cheap. But that is exactly the problem. The price does not reflect a probability-weighted path to that outcome. It reflects the outcome itself, the campuses built, the leases signed, the rent flowing, arriving more or less on schedule. The most optimistic published models already assume something like47% annual revenue growth and a swing to positive earnings by 2029, an outcome the current $0.9 million of recurring lease revenue and 50 megawatts of contracted capacity do not yet support.

That is the disconnect the data keeps surfacing. The buyer at $24 is not paying for a Bitcoin miner with an interesting option on data centers. The buyer is paying for the data center business as though it were already built and leased, and accepting the loss-making mining operation and the Bitcoin exposure as part of the bargain. For that price to be reasonable, almost everything has to go right, on time, at the assumed scale. The history of first-of-a-kind infrastructure buildouts is not kind to that assumption.

What Could Go Wrong

Every avoid thesis owes the other side an honest hearing, and this one is stronger than most avoid candidates we cover. There is a real version of the bull argument, and a buyer at today's price is not obviously foolish.

The power is a genuine moat. The binding constraint on AI is grid-connected power at scale, and Riot controls hundreds of megawatts that a new entrant would wait years to secure. Starboard is right that this asset is scarce, hard to replicate, and worth more in an AI-hungry world than in a crypto-only one. If the compute shortage persists, that capacity appreciates.

The AMD relationship is real and high-margin. The 50-megawatt commitment from an investment-grade tenant is a real anchor, and operating lease revenue carries very high gross margins once it scales. A handful of additional anchor leases would change the recurring-revenue picture quickly, and the fit-out work shows Riot can execute the construction.

A credible activist is engaged. Starboard is not a message board. It is a serious investor with a real stake and a track record of forcing value-accretive change. Its presence raises the odds of disciplined capital allocation, asset monetization, or a transaction that crystallizes the sum-of-the-parts value it describes.

Bitcoin optionality is free. The mining business, unprofitable as it is on a fully-loaded basis today, would swing back toward profitability on a sharp Bitcoin rally, and the 15,679 coins on the balance sheet would mark up. That is upside the AI thesis does not even require.

The thesis breaks if Riot signs a string of large leases and energizes Corsicana close to plan. In that world the recurring revenue ramps, the multiple compresses, and the stock grows into its price. What the avoid thesis argues is not that this is impossible. It is that the current price already assumes it, leaving little reward if everything goes right and meaningful downside if the leasing is slow, the buildout slips, or Bitcoin stays soft. It is the same pattern examined in The Quantum Computing Bubble and the IREN avoid thesis: a real technology, a real opportunity, and a price that has detached from the part of the story that has actually been proven.

The Bottom Line

Riot Platforms is not a bad company. It owns scarce power, it signed a real anchor tenant, and it has a credible activist pushing it to unlock value. If the buildout lands on plan, the bulls will be right and this article will age poorly. That outcome is possible. It is simply not what the current price represents, because the current price represents that outcome as already achieved.

The most recent quarter showed a $500 million net loss, mining that costs more to run than it produces, two-thirds of revenue still coming from a declining crypto operation, and a data center business generating under $1 million in recurring lease revenue. The number carrying the valuation, an activist's $21 billion, is a projection of what the power sites could be worth, not a measure of what they earn. The platform's systematic scoring reads Weak, Bearish, and Expensive, with a fair value roughly 84% below the market price. Every independent angle lands in the same place.

That is the entire philosophy here. Wealth Engine Pro evaluates companies on what they are, not on what the most exciting version of the story hopes they will become. The narrative says Riot is an AI infrastructure operator. The data says it is a money-losing Bitcoin miner in the early innings of an expensive transition, priced as though the transition were already complete. The market is paying for a gigawatt of leases. The filings show a letter, fifty megawatts, and under a million dollars of recurring rent. Until that changes in the numbers, the data is the part worth trusting.