Riot Platforms, Inc.

Overview

This report applies a Joel Greenblatt Magic Formula-style framework to Riot Platforms (RIOT) as of 2026-08-16: assessing the company first as a quality business via return on capital, then as a cheap stock via EBIT/enterprise value, with a focus on normalized earnings rather than temporarily distorted GAAP results.

Business Quality Assessment

RIOT does not currently screen as a high-quality Magic Formula business. The trailing operating base is deeply negative. In H1 2026, segment gross profit was only $78.4M, while SG&A was $158.2M, depreciation and amortization was $195.5M, and the company recorded a $28.0M impairment. That produced a GAAP operating loss of $739.3M. Even excluding the most clearly non-cash/non-operating items—$401.3M bitcoin fair-value loss, $60.2M derivative fair-value loss, and $28.0M impairment—H1 2026 EBIT was still roughly negative $250M. The Data Center operating lease gross margin is promising at 84%, but recurring operating lease revenue was only $4.9M in Q2 2026. Bitcoin mining is cash-positive before miner depreciation in H1 2026, with cost to mine excluding depreciation of $47,369 versus an average bitcoin price of $73,736, but after miner depreciation the segment is deep negative. Return on invested capital is currently negative whether bitcoin is treated as operating working capital or not. The business has a real strategic option in AI/data center leasing, but it is not yet a high-return business in the Greenblatt sense.

Valuation Analysis

Enterprise value is approximately $7.43B using market cap of $7.14B plus $843M total debt less $549M cash and restricted cash. If bitcoin holdings of $666M are also treated as liquid/investment assets, adjusted EV is about $6.76B. Against that, normalized TTM EBIT appears to be approximately negative $600M to negative $650M after removing bitcoin fair-value swings, derivative mark-to-market losses, and impairments. That implies a negative earnings yield of roughly -8% to -9%, far worse than prevailing bond yields near 4-5%. The market is not paying a low multiple on current earnings because there are no current earnings. The forward P/E is -20.33, and the price-to-book ratio is 3.01. RIOT is not cheap on a Magic Formula earnings-yield basis.

Magic Formula Ranking

Normalized Earnings Analysis

Reported earnings are heavily distorted by non-cash bitcoin mark-to-market losses of $401.3M in H1 2026, derivative losses of $60.2M, and a $28.0M impairment from reorienting Rockdale infrastructure from mining to data center use. Stock-based compensation of $74.7M in H1 is also large relative to revenue. However, even after these adjustments, the core business is not yet meaningfully profitable. Sustainable owner earnings are currently negligible or negative. The most attractive economic asset is the future data center NOI: management estimates $365M to $411M average annual NOI from the 191 MW frontier AI lab lease and $51M from AMD's 50 MW lease, for combined contracted NOI of $416M to $462M. But the 191 MW deployment is not scheduled for full delivery until June 2028 and requires $2.1B to $2.3B of capex. Magic Formula normalization should not capitalize future leased NOI before it is built, financed, and generating cash.

Why The Market Is Wrong

The market is excited about RIOT's AI data center pivot and headline $9.1B lease, but it is pricing future contracted revenue and power scarcity rather than current earnings power. After the initial post-earnings spike, the stock gave back much of the gain, suggesting the market may be tempering early enthusiasm. The contrarian Magic Formula view is that RIOT is a capital-intensive transitional business with negative current EBIT and negative return on tangible capital; the market may be overpaying for optionality. Conversely, the market may be underappreciating the value of fully approved Texas power if the Corsicana LOI converts to a lease. But that is a development-special-situation case, not a Greenblatt 'good company at a cheap price' case.

Key Risks

Conclusion

Using Joel Greenblatt's Magic Formula, RIOT fails both required tests: earnings yield is negative and return on capital is negative. The stock is being valued on future AI data center potential, not current earnings power. Even after normalizing for bitcoin mark-to-market losses and one-time impairments, the underlying business is losing money after depreciation and SG&A. While the data center leases could eventually create high-margin contracted cash flows, that is a capital-intensive development story with significant execution and financing risk. A Magic Formula investor would not buy RIOT today; the prudent rating is SELL/avoid until the company demonstrates positive normalized EBIT and acceptable return on tangible capital.

Overview

This is a Druckenmiller-style top-down macro and reflexivity analysis of Riot Platforms (RIOT). The core question is not whether RIOT is a cheap bitcoin miner, but whether the market is mispricing the shift from a bitcoin treasury/cash-flow model to a scarce, contracted AI-power infrastructure business. The report focuses on liquidity flows, power scarcity, credit feedback loops, and the asymmetry between current valuation and the optionality of fully approved Texas power capacity.

Macro Context

The dominant macro backdrop is an AI-driven infrastructure buildout in which power access, not server demand, is the binding constraint. The US grid, especially ERCOT, is effectively rationed for new large loads; existing fully approved, energized power is a scarcity asset. Monetary policy remains relevant but the capital markets are bifurcated: project finance and private credit are deep for investment-grade tenant-backed data center leases, while unproven or merchant power projects pay punitive costs. Bitcoin is in a weaker macro regime: H1 2026 average BTC price was roughly $73,700 versus $95,991 in H1 2025, and rising global network hash rate has compressed mining economics. Regulatory moves such as Texas SB6 and ERCOT batch-study rules reinforce the advantage of already-approved sites. In this environment, the marginal dollar flows toward companies that can provide contracted power to hyperscalers and AI labs rather than purely commodity bitcoin production.

Company Position in Macro Landscape

RIOT is increasingly a beneficiary of the AI-power bottleneck, not a pure crypto entity. The company controls approximately 1.7 GW of fully approved and energized capacity at Rockdale (700 MW) and Corsicana (1 GW), plus Kentucky assets. It has contracted 241 MW of critical IT capacity: 50 MW with AMD and 191 MW with a leading frontier AI lab, reported as Anthropic. The 191 MW lease is expected to generate about $9.1 billion over 20 years, with annual NOI of $365-411 million once fully deployed by mid-2028. AMD adds another roughly $51 million annual NOI at 50 MW. At current data, RIOT is still hostage to old bitcoin economics: Q2 2026 revenue was $174.2 million but GAAP net loss was $237.2 million, driven by a $74.6 million bitcoin fair-value loss, depreciation, and impairment. The company is not yet a clean infrastructure compounder; it is a transitional asset owner with negative current earnings whose value increasingly derives from long-dated power contracts.

Reflexivity Analysis

The reflexivity is now self-reinforcing in both directions. Positive loop: delivering AMD on time creates credibility -> higher-quality tenants sign leases -> lenders treat the cash flows as investment-grade-like -> RIOT secures low-cost project financing -> capital recycling funds more buildouts -> the market rerates the power portfolio from $3-4 million per MW toward AI data-center comps of $7-9 million per MW. The Morgan Stanley $573 million interim facility is an early signal of this loop. Negative loop: RIOT still funds equity requirements by selling bitcoin into a weak tape; bitcoin markdowns depress equity and reported results; weak GAAP earnings keep fundamental investors cautious; delays in permanent financing could force dilutive equity or slower development. Short interest near 15% of float adds a squeeze tailwind on positive news, but the post-earnings fade from premarket highs shows the market is not yet convinced. The key reflexive catalyst will be the closing of permanent investment-grade project financing and conversion of the Corsicana LOI into a legally binding lease; that is the point where the story moves from 'show me' to 'scarcity rerating.'

Competitive Position & Disruptive Threats

RIOT's moat is physical and increasingly scarce: 1.7 GW of approved, energized power in Texas, on owned land, with existing substations and direct grid interconnection. That is difficult to replicate quickly under current ERCOT rules. Vertical integration through ESS Metron and E4A provides switchgear and electrical engineering capacity, reducing supply-chain risk. The company has credible anchor tenants in AMD and Anthropic. Against peers such as Core Scientific, Hut 8, TeraWulf, and Applied Digital, RIOT still trades at a discount on EV per available MW despite having contracted revenue. The main competitive risk is that larger hyperscalers or specialized developers build their own campuses, but that is constrained by grid access. Disruptive threats include AI workload efficiency gains reducing power demand per token, potential future generation technologies, and tenant concentration. RIOT's engineering segment is also exposed to data-center equipment cycles, but current backlog of $177.1 million, about 90% data-center related, indicates industrial demand.

Asymmetric Risk/Reward

At a market price near $19.02 and market cap near $7.1 billion, RIOT is close to its 200-day average of $18.50 after giving back much of the August earnings spike. The upside is tied to contracted cash flows and optionality: 241 MW contracted could generate $416-462 million annual NOI at maturity. Even at 15x stabilized NOI, that segment alone would be worth $6.2-9.2 billion enterprise value, before assigning anything to the remaining Rockdale capacity or the 1 GW Corsicana campus. The Corsicana LOI for the entire site, if signed, was described by management as potentially supporting more than $1 billion in annual rent at full build-out. Downside is real: book value per share is only about $6.32, so the equity is not asset-cheap at 3x book. If permanent financing fails or leases slip, the stock can retrace toward the $11-13 range. The payoff is therefore convex: multiyear upside to $30 or higher if the infrastructure re-rating continues, with downside risk of roughly 35-40% if execution fails. The current pullback improves entry asymmetry versus chasing the post-earnings spike.

Key Risks

Investment Details

Overview

A Peter Lynch-style fundamental analysis of Riot Platforms (RIOT) as of August 2026, using Q2 2026 financial results, the 10-Q filing, and data center lease announcements to assess whether the company is a buyable story for an individual investor.

The Two-Minute Story

Riot owns large, fully approved, energized power campuses in Texas, mainly Rockdale with 700 MW and Corsicana with 1,000 MW. It is converting that power from low-return bitcoin mining into AI data center leases. It has already signed AMD and a leading frontier AI lab, reported by CNBC as Anthropic, for 241 MW of critical IT capacity, representing about $9.8 billion of contracted data center revenue. If Riot builds those data centers on time and finances them, it becomes a toll booth on scarce AI power. Today the company is still losing money because bitcoin prices fell, bitcoin mark-to-market losses are large, and it is spending heavily on the transition. The simple thesis is that the power assets and signed leases are worth more than the current market value, but execution is the whole game.

Stock Category

Do You Understand This Business?

At a high level, yes: Riot owns power infrastructure and either mines bitcoin or leases powered data center space to AI companies. An average person can understand power scarcity and data center demand. However, the detailed financials are complex because of bitcoin mark-to-market accounting, derivatives, tenant fit-out revenue, project financing, and convertible notes. The edge would come from understanding Texas power markets, data center construction costs, lease economics, and financing terms better than the market.

PEG Ratio Analysis

Lynch's Checklist

Tenbagger Potential

A 10x from the current $7.1 billion market cap would require roughly a $71 billion valuation. If Riot leased and stabilized all of its current Texas power at the economics of the recent AI deals, the business might support a valuation of several times its current level, but likely not 10x by itself. A tenbagger would require flawless delivery, conversion of the Corsicana LOI into a lease, low-cost project financing, minimal dilution, rising bitcoin prices, and additional power capacity acquisition. That is possible but too speculative to be a base case.

Key Risks

Conclusion

Peter Lynch would probably say this is not an earnings story yet and it is definitely not a boring, overlooked company. The data center pivot is not diworsification because it uses the same power and engineering core, but it adds real execution risk. Riot has a genuine asset base in scarce approved power and has secured impressive tenants, which supports the bull case. But you are paying about $7.1 billion for a company with trailing negative earnings, negative operating cash flow, and a multi-year, multi-billion-dollar buildout ahead. A speculative investor may buy for the AI infrastructure optionality. A conservative Lynch-style investor should wait for evidence that the asset play is converting into sustainable cash earnings before paying up. Therefore, HOLD.

Overview

Deep value / contrarian analysis of Riot Platforms (RIOT) at $19.02: a hated Bitcoin miner that is quietly transitioning into a contracted AI data center landlord. The market is pricing the GAAP losses from the legacy mining business and ignoring the $9.8B in signed data center revenue and 1.7GW of approved Texas power.

The Bear Case

Consensus sees a cash-burning crypto miner riding AI hype. H1 2026 GAAP net loss was $737.6M, driven by $401M bitcoin mark-to-market losses, $195M D&A, and a $28M impairment. Bitcoin mining revenue fell because average BTC price dropped and network hash rate rose; cost to mine including depreciation was about $90,631 versus production value of $71,667 in Q2. The company sold 9,665 bitcoin, shrinking the stack from 18,005 to 11,380. Data center revenue is still small and mostly low-margin tenant fit-out work; operating cash flow was -$272.8M in H1. Bears argue the 191MW Anthropic buildout requires $2.1-2.3B capex, delivery is years away, financing is not finalized, and the stock at roughly 13x P/S is already pricing success before it happens. Elevated short interest near 15% of float reflects this skepticism.

The Bull Case

Contrarian thesis: RIOT owns something scarcer than bitcoin exposure: 1.7 gigawatts of fully approved, energized Texas power at Rockdale and Corsicana. It has signed $9.8B of contracted data center revenue with AMD and Anthropic for 241MW. The Anthropic lease alone is expected to generate $9.1B base rent over 20 years and $365-411M annual NOI; AMD adds about $51M annual NOI. Combined 241MW should produce $416-462M annual NOI by mid-2028. A Corsicana LOI for the full roughly 1GW campus could add more than $1B annual rent if signed. Management's full buildout NOI target is $1.6-2.1B. The current $7.1B market cap values RIOT largely as a miner, not a digital infrastructure company. The AMD on-time delivery validates execution. $1.2B liquidity and project financing provide runway. If the market re-rates contracted NOI at data center multiples, upside to $29-36+ is credible. Short interest adds squeeze potential.

Fundamental Deep Dive

Sentiment & Technical Setup

Catalyst Analysis

1) Corsicana LOI conversion: a full-campus lease could add more than $1B annual rent and force a data center valuation. 2) Project financing: closing the AMD term loan and Anthropic project debt would remove dilution fears. 3) Delivery milestones: AMD 25MW expansion in November 2026 and May 2027, and Anthropic 96MW in December 2027, prove execution. 4) Data center operating lease revenue ramped from $0.9M in Q1 to $4.9M in Q2, with full 25MW now online. 5) Bitcoin rally would improve mining cash flow and treasury. 6) Short squeeze: roughly 15% short interest against catalyst-rich news flow.

Key Risks

Conclusion

This is a battleground where the market is anchored to trailing GAAP losses from bitcoin mining while the company is monetizing scarce approved power into long-dated AI leases. At $19.02 and a $7.1B market cap, the market is not giving full credit for the $416-462M contracted NOI potential from 241MW, nor the Corsicana option. The trade is not without risk: it requires patient capital and tolerance for volatility. But the setup, negative sentiment, elevated short interest, catalysts in 2026, credible tenants, and management execution, fits a deep value/contrarian framework. I would rather own the power and contracted cash flows with a free option on Bitcoin than short it.

Overview

This report applies William J. O'Neil's CAN SLIM methodology to Riot Platforms, Inc. (RIOT) as of August 16, 2026, using the provided financial data and recent SEC filings, earnings disclosures, and market commentary. It evaluates whether RIOT qualifies as a disciplined growth-stock purchase under the seven CAN SLIM criteria.

Financial and Business Overview

Riot Platforms is transitioning from a pure bitcoin miner into a digital infrastructure company focused on large-scale data center leasing for AI/high-performance computing. Q2 2026 revenue was $174.2 million, up 14% year over year, with Bitcoin Mining revenue of $113.7 million, Data Center revenue of $23.2 million, and Engineering revenue of $37.3 million. However, profitability is weak: Q2 2026 net loss was $237.2 million versus net income of $219.5 million in Q2 2025. First-half 2026 net loss was $737.6 million. Results were heavily affected by a $401.3 million negative fair-value change on bitcoin, higher depreciation, and a $28.0 million impairment. The company ended Q2 with $548.9 million in cash and restricted cash and 11,380 bitcoin valued at about $666 million, roughly $1.2 billion in liquid assets. Strategic announcements include a 20-year, 191 MW AI data center lease expected to generate about $9.1 billion in base rent and an expanded AMD lease bringing contracted capacity to 241 MW. Despite the strong new business narrative, current GAAP earnings are deeply negative.

Market Position & Competitive Advantages

Riot's key strengths are its fully approved and energized power portfolio in Texas, including roughly 700 MW at Rockdale and 1 GW at Corsicana, plus additional Kentucky capacity. Power access is a scarce asset in the AI data center buildout. Vertical integration through ESS Metron and E4A Solutions gives Riot internal control of electrical equipment and engineering, which can reduce supply-chain risk and support faster delivery. Signed leases with AMD and a reported leading frontier AI lab provide about $9.8 billion of contracted data center revenue and validate the pivot away from pure mining. That said, execution risk is high: the AI buildout is expected to cost $2.1-2.3 billion for the 191 MW project, current operations are burning cash, and the company is relying on bitcoin sales and project debt until leasing cash flows ramp. Bitcoin price exposure, tenant concentration, and execution/capital risk are material.

Stock Performance

RIOT closed near $19.02 on August 16, 2026, down about 1% on the day. The stock is below its 50-day moving average of $23.14, a roughly -17.8% deviation, but above its 200-day moving average of $18.50 by about +2.8%. The 52-week range is $11.16 to $30.32, and the stock sits about -37.3% below the 52-week high. Despite a +54.3% year-over-year gain, the most recent tape shows weakness. The 10-day average volume of 26.5 million shares is well above the 3-month average of 18.3 million, indicating active trading. After the Q2 earnings and AI lease announcement, RIOT initially spiked more than 20% but gave back much of the gain, a sign of selling into strength. This is not the classic CAN SLIM setup of a stock near new highs on strong accumulation.

CAN SLIM Analysis

Key Risks

Conclusion

Under CAN SLIM, a stock must first demonstrate strong current and annual earnings growth. RIOT fails both foundational criteria: it swung to a large quarterly net loss, TTM EPS is -$3.84, and annual earnings are inconsistent. The new AI data center leases are powerful N catalysts, but they have not yet translated into positive EPS or a rising stock near highs. Instead, the stock reversed lower after the announcement, trades below its 50-day average, and shows distribution-like volume. Supply/demand is not confirming, institutional sponsorship is mixed, and market direction is not supportive. Therefore, despite the strategic transformation, RIOT does not meet the CAN SLIM profile and should be avoided by disciplined growth investors until fundamentals and price action improve.

Overview

A Warren Buffett-style intrinsic-value analysis of Riot Platforms, Inc. (RIOT), focusing on business quality, moat durability, management behavior, financial strength, and whether the current market price offers a margin of safety.

Business Understanding

Riot Platforms is a Bitcoin miner transitioning into an AI data-center landlord. The Bitcoin mining side is easy to understand: Riot runs computing hardware, consumes electricity, and earns bitcoin. That is a commodity operation with no pricing power. The newer data-center business is also understandable in principle: Riot leases power capacity and builds facilities for tenants like AMD and a leading AI lab. However, it is capital-intensive, execution-heavy, and the future cash flows are years away. The basic economics are understandable, but the business is outside the circle of competence reserved for predictable, low-change industries.

Economic Moat Analysis

In Bitcoin mining, Riot has no durable moat. It sells a commodity product priced by Bitcoin network economics, competes against many well-funded miners, and its returns depend on bitcoin price, network hash rate, and power costs. Its low-cost power strategy and vertical integration provide some operational advantages, but not sustainable pricing power. In data centers, Riot may be building a narrow moat from owned, fully approved, energized power assets; long-term leases with creditworthy tenants; and electrical engineering capabilities. Once a tenant is built in, switching costs are high. But this moat is unproven, capital-intensive, and currently concentrated in a few customers. Overall, the moat is narrow at best and not yet demonstrated by durable returns on capital.

Management Quality

Management has shown operational competence by delivering the initial AMD capacity on time and signing a large AI lease. However, capital-allocation behavior is concerning. The company has relied heavily on stock-based compensation, historically used at-the-market equity issuance, sold Bitcoin into weakness to fund operations, and recorded realized Bitcoin losses of $113.2 million in the first half of 2026. The 2025 Say-on-Pay vote received only about 31.5% support, highlighting shareholder dissatisfaction with executive compensation. There was a CFO transition and related-party employment arrangements. Management is ambitious, but the track record does not yet show disciplined, shareholder-first capital stewardship.

Financial Strength

Financial strength is weak. The company posted a first-half 2026 net loss of $737.6 million, operating cash flow of negative $272.8 million, and declining stockholders' equity from $2.86 billion to $2.19 billion. Return on equity is negative. Debt is meaningful: $594.4 million principal on 2030 convertible notes, a $200 million bitcoin-backed credit facility, and other revolving borrowings. Interest coverage is negative. Bitcoin holdings provide some liquidity, but they are volatile and partly restricted as collateral. Heavy depreciation, stock compensation, and fair-value swings make reported losses large even when revenue grows. There is no dividend and no meaningful free cash flow.

Intrinsic Value Assessment

Owner earnings are currently negative. Even after adding back depreciation, the business consumes cash because capex and development spending exceed operating cash flow. The announced AI lease could generate estimated average annual NOI of $365-$411 million, and AMD another $51 million, but those cash flows require $2.3-$2.5 billion of additional development spending and are not expected to fully stabilize until 2028 or later. Discounting future contracted NOI and offsetting debt, dilution, execution risk, and current cash burn, I estimate intrinsic value is below the current price. A rough fair value is nearer $12-$15 per share. At $19.02, there is no margin of safety; the market is paying for successful execution that has not yet occurred.

Key Risks

Investment Details