Asset Analysis · June 2026
I've opened five new positions, and although at first glance they may seem like a motley collection of small-cap stocks, the truth is that All five hang from exactly the same string: Artificial intelligence isn't built with software; it's built with energy and computing power, and almost no one is looking at the part of the chain where supply is truly scarce.
The common thesis
While the market is scrambling to pay forty or fifty times earnings for the visible layers of the AI business, the physical infrastructure needed to make it all work—the megawatts, land with access to the power grid, liquid-cooled data centers, solar cell factories—is in many cases priced at multiples of a forgotten company. And that's where I wanted to position myself.
The five purchases are Applied Digital (APLD), CoreWeave (CRWV), CleanSpark (CLSK), Keel Infrastructure (KEEL), and T1 Energy (TE). Three of them—APLD, CRWV, and KEEL—are, to varying degrees, direct plays on AI computing power. CLSK adds the layer of Bitcoin mining as an energy-intensive asset that is repurposed for that same computing power. And TE brings the energy angle: domestic manufacturing that fuels all that electricity demand. It’s not a basket of individual bets; it’s the same thesis expressed in five different ways, each with distinct risk profiles: If the overall direction is right, I don't need all five elements to be perfect for the whole thing to work.
A word of caution regarding everything that follows: these are small or medium-sized companies, most of which are reporting losses, have high operating leverage, and exhibit a level of volatility that isn’t for everyone. My position in one of them is already trading below my cost basis, and I’m not particularly concerned about it, because I'm not chasing a price; I'm buying a thesis with a time horizon that isn't measured in trading sessions. The price targets I set are my own, based on execution scenarios I consider likely—not on market consensus, which in many cases lags behind what I believe will happen.
| Ticker | Company | Home | Objective | Upside |
|---|---|---|---|---|
| CRWV | CoreWeave | $105,15 | $160 | +52% |
| CLSK | CleanSpark | $17,40 | $26 | +49% |
| TE | T1 Energy | $10,80 | $16 | +48% |
| KEEL | Keel Infrastructure | $5,45 | $8 | +47% |
| APLD | Applied Digital | $45,60 | $65 | +43% |
Actual entry prices in the portfolio. Our own 12- to 18-month price targets. The risk rating is qualitative and subjective.
Position 01
Applied Digital is the company that best exemplifies this thesis. It started out in cryptocurrency mining and has completely pivoted toward building high-performance data centers, known as «AI Factories,» which it leases to hyperscalers under contracts lasting about fifteen years. The model is, at its core, a highly specialized form of industrial real estate: it secures land with access to high-capacity electrical power, builds a liquid-cooled facility suitable for AI chips, and leases it to a tenant with high creditworthiness who pays rent for a decade and a half.
The key takeaway from the last quarter (Q3 of fiscal year 2026) is that this is no longer just a promise. Revenue grew by 1,391 TP3T year-over-year to $126.6 million, with adjusted EBITDA of $44.1 million and adjusted net income of $33.2 million ($0.09 per share), although under GAAP it continued to report a loss of $100.9 million due to depreciation and non-recurring items. The first 100-MW building at Polaris Forge 1 is now fully operational; management notes that about 900 MW (nearly 1 GW) under construction and to exceed 1 billion in NOI within five years. The balance sheet showed approximately 2.1 billion in cash compared to 2.7 billion in debt.
The real gem is the contract portfolio. The Ellendale campus (Polaris Forge 1) has 400 MW fully leased to CoreWeave for approximately 11 billion over a 15-year period. Added to this is a 200 MW contract with an investment-grade hyperscaler at Polaris Forge 2, and a deal worth approximately 7.5 billion for 300 MW at Delta Forge 1 with another highly rated hyperscaler, which became a repeat customer just one month later. In total, the company has accumulated approximately $16 billion to $23 billion in contracted rental revenue, with the goal of ensuring that the 70% is sourced from investment-grade counterparties. When a client returns so quickly to seek additional capacity, it is usually the best possible indication of the asset’s quality.
Revenue estimates for fiscal year 2027 point to $540–550 million, nearly double the current figure as the campuses mature. The downside is well known: increasing leverage to finance construction and a significant concentration of clients. The consensus has been raising its price targets to $45–$70 (Needham $66, Citizens $60, Compass Point $70).
My price target is $65, based on the gradual conversion of the backlog into recurring NOI and a reasonable NOI multiple for assets under 15-year leases. I bought in at $45.60; the stock has corrected somewhat and is now trading below my cost basis, which, in my view, doesn’t change my investment thesis one bit.
Position 02
If Applied Digital is the landlord, CoreWeave is the toll booth. It rents out GPU-accelerated computing power to those who build large AI models, and its client list speaks for itself: Microsoft, Meta, OpenAI, Anthropic, Mistral, Cohere and quantitative firms such as Jane Street and Hudson River Trading. It is one of the most in-demand computing infrastructures on the planet, and its relationship with NVIDIA is so close that it is often the first to receive and test the latest hardware.
The figures for the first quarter of 2026 are hard to overstate: revenue of 2.078 billion, an increase of 11.21% compared to the same period a year earlier, with adjusted EBITDA of 1.157 billion and a margin of 56%. But the figure that underpins the thesis is the contracted backlog of 99.4 billion, nearly 50% more than at the end of the previous year, with an average duration of nearly five years. This includes a commitment of approximately 21 billion with Meta through 2032, an agreement worth about 6 billion with Jane Street, and the partnership with OpenAI. Management states that it has already committed more than 75% of its 2027 capacity.
There is one particularly telling sign: who is putting up the money and at what price. NVIDIA invested $2 billion in CoreWeave shares at $87.20, and Jane Street invested $1 billion at $109. When the chip manufacturer itself and one of the world’s most sophisticated quantitative firms are buying equity at those levels, they’re telling you something about the asset’s value. Furthermore, CoreWeave has pioneered a new class of financing—investment-grade GPU-backed loans (rated A3 by Moody’s on an $8.5 billion facility)—which lowers its cost of capital and validates the model in the credit market.
It's not all sunshine and rainbows. It drags on. more than 50 billion in total liabilities and a massive pace of investment (31–35 billion in projected capex for 2026) that is squeezing margins and forcing the company to raise capital on a recurring basis, with bond issuances at 9.751% 3-year notes. Second-quarter guidance was weaker than expected and shifts much of the year’s weight to the second half.
Market estimates range widely (Jefferies 160, Wells Fargo 155, BofA 140). I've set my target at $160, in line with the bullish outlook. I entered at $105.15, below the price at which Jane Street entered, which I think is a reasonable starting point.
Position 03
At its core, CleanSpark is one of the largest Bitcoin miners in the U.S., with a operational hashrate of approximately 50 EH/s —it was the first publicly traded mining company to reach that level with fully self-operated data centers— and a substantial BTC position on its balance sheet. But the reason it’s in this basket is what it’s building on top of that: it controls roughly 1.8 GW of contracted capacity (with approximately 808 MW actually in use), including a recent approval of 300 MW in ERCOT (Texas). That access to large-scale power, at a time of structural shortages, is precisely what the AI industry needs but cannot find.
CleanSpark maintains more than 13,400 Bitcoin on the balance sheet, most of which were mined in-house, and actively manages them. In May, it produced 671 Bitcoin (about 21.66 per day), selling 654 at an average price of around $80,000. What is noteworthy is its production cost, around $43,000 per Bitcoin in marginal cost, giving it a substantial cushion relative to the market price and placing it among the most efficient miners (with a fleet efficiency of nearly 16 J/TH). Even a severe correction in Bitcoin would leave it operating at a profit.
On paper, the results for the second fiscal quarter were a failure: a loss per share of $1.52 compared to $0.49 the previous year, on revenue of $136.4 million. But here’s the catch: Most of that loss stems from a non-cash accounting write-down of approximately 224 million based on the value of Bitcoin and the collateral, not on the deterioration of the operating business. The gross margin remained above 40% in the third quarter, and the company maintains liquidity of around 1.2 billion and a comfortable current ratio. The impact is largely accounting-related, not cash-related.
Management is shifting part of its capacity toward high-performance data centers: it is redeveloping its site in Sandersville (Georgia), 250 MW, for AI and HPC workloads alongside mining, and is in talks with an investment-grade hyperscaler with a pipeline exceeding 5 GW. The acquisition of GRIID provided it with high-quality sites powered by the TVA in Tennessee. Here’s how I see it: The market allows me to buy an efficient, well-capitalized Bitcoin mining company at a reasonable price and, on top of that, get an option on its conversion to AI infrastructure practically for free.
Analysts' estimates range from $16 to $30, depending on the weight they assign to the pivot (Maxim and Macquarie at $22, B. Riley/KBW at $26, average ~$20–$23). I'm setting my target at $26, based on the 2027 EBITDA valuation and hashrate growth, with AI and cash flow serving as upside catalysts. I entered the position at $17.40.
Position 04
Keel is probably the least well-known of the five positions, and also the most speculative. Keel Infrastructure is the former Bitfarms, the Canadian Bitcoin mining company, which completed a major transformation in April: it relocated to the U.S., established its headquarters in New York, adopted a new name and ticker symbol, and shifted its focus from pure mining to digital and energy infrastructure for HPC and AI in North America.
What makes it interesting isn't some generic «shift to AI» narrative—half the mining industry is already talking about that—but the specific asset it holds: a a 2.2 GW pipeline and, crucially, existing power grid interconnections, focusing on three of North America’s most strategic markets with the tightest supply: Pennsylvania, Washington State, and Quebec. In AI data centers, the real bottleneck isn’t land or concrete; it’s access to megawatts with approved interconnection, and that can take years to build from scratch. Keel already has it.
The location of Panther Creek (Pennsylvania) has master plans for an HPC campus with a capacity of up to 350 MW, with Sharon (Pennsylvania) and Moses Lake (Washington) moving forward, with zoning and permitting currently underway. Much of its generation in Pennsylvania comes from plants that burn waste classified as an alternative fuel source, with access to tax credits and the sale of electricity on the PJM market. Keel has completed technical evaluations with partners such as World Wide Technology and Appleby Strategy Group for all of its sites.
He has completed his Complete exit from Latin America following the sale of Paso Pesado, bringing forward cash flows to reinvest in North American HPC/AI infrastructure. As of May 8, it had total liquidity of approximately $533 million (with $336 million in unrestricted cash and ~$197 million in unencumbered Bitcoin), and is executing an orderly wind-down of its Bitcoin position. In early June, it bolstered its cash position with $400 million in convertible notes at 1.2501% with a maturity date of 2032, financing at a very low cost.
The consensus ranges from $3 (Cantor) to $5.50 (H.C. Wainwright), with $4.50 (Chardan) in the middle. My target price of $8 is deliberately set above the consensus: I think the market still values Keel for what it used to be (a Bitcoin mining company) rather than what it is becoming (an energy-rich infrastructure operator with guaranteed interconnection). The day it signs its first major HPC contract, the market will be forced to apply a data center multiple—not a mining multiple—to its 2.2 GW: that’s where the re-rating comes in. I entered at $5.45, and because of that dichotomy, this is the one I’m approaching with the most caution.
Position 05
T1 Energy completes the supply-side aspect of its business model. It is the former FREYR Battery, which has been transformed into a a domestic manufacturer of solar modules and cells in the U.S., based in Austin. While the four companies mentioned above are voracious consumers of electricity, T1 is on the generation side, building solar capacity on U.S. soil at a time of strong political support for domestic energy production. If AI is an energy problem, someone has to produce that energy on U.S. soil.
The first quarter of 2026 was a true turning point. Revenue skyrocketed to about 177 million from 53 million the previous year, reported its first positive net income from continuing operations (approximately 3.9 million) and a record adjusted EBITDA of 9.1 million, driven by its G1 plant in Dallas, which is now fully operational. The stock rose more than 50% in the third quarter following these results and, even so, continues to trade at compressed sales and EBITDA multiples relative to its future potential.
The major catalyst still to come is the G2 plant in Austin, a 2.1-GW cell factory scheduled to begin production in Q4 2026, and whose ramp-up would lead management to target adjusted EBITDA of $375 million to $450 million in 2027. Added to this is a key cash flow driver: the monetization of the 45X tax credits that reward residential solar power generation. The company has already sold $160 million in these credits to an investment-grade financial institution at $0.91 per credit dollar. In addition, it announced the acquisition of KORE Power for approximately $32 million to align with energy storage and the needs of AI data centers.
Northland initiated coverage with a target price of $16, while BTIG set its target at $8. I've set my target at $16, based on the jump in EBITDA resulting from G2's entry, plus the leverage from the 45x loans. I bought at $10.80.
Overview
I could have focused everything on the highest-quality option—probably CoreWeave or Applied Digital—and that would have been a valid choice. The reason I chose to spread it across five is because Each one expresses the same idea at a different point on the risk-return curve. CRWV and APLD are the most established plays, with a contracted backlog and real revenue visibility. CLSK adds a Bitcoin component and a virtually free option on AI. KEEL is the early re-rating play, offering the greatest potential upside and the highest risk. And TE represents the other end of the spectrum—power generation—which benefits from the same structural demand but through a different mechanism and with its own fiscal leverage.
Adjust the weight of each of the five positions and observe how the basket’s weighted potential return changes relative to its own 12- to 18-month target prices. The weights are automatically normalized to 100%.
But there is something more important than the breakdown by risk profile: the internal correlation logic of the basket. If demand for AI accelerates, everyone benefits, but the transmission mechanisms are so different that one position does not replace another. CoreWeave rises when hyperscalers sign more computing contracts; Applied Digital when that growth drives demand for physical data center space; CleanSpark due to Bitcoin and its reclassification as an HPC company; Keel when the market applies an infrastructure multiple rather than a mining multiple; and T1 Energy when energy demand and domestic industrial policy converge at its Austin facility. Five different catalysts, five paths leading to the same conclusion.
It is also worth reflecting on the current stage of the cycle. For months, the prevailing narrative has focused on the software and application layers of AI—companies like Palantir and Salesforce, as well as model providers—paying multiples that discount decades of growth. Physical infrastructure, on the other hand, continues to trade at a discount relative to the demand curve already locked in as backlog. There is a reason for this: The market tends to undervalue assets that don't stand out in a demo, and a data center in North Dakota or a solar cell factory in Texas don’t shine in a demo. They shine in the revenue they generate over fifteen years and in the tax credits. That is precisely the inefficiency I’m trying to highlight.
There is one more thing: the the implicit dominance that TE exerts over the rest. The other four are net energy consumers; T1 Energy is on the supply side. If energy prices rise structurally, T1 and its 45X leveraged ETFs appreciate in value, while the operating costs of the other four increase. It is not a perfect hedge, but it introduces a real element of diversification into what would otherwise be a purely one-way bet on electricity consumption.
Finally, and this is what matters most to me as an investor, this portfolio has a feature that I particularly value: The catalysts are binary and occur in close succession. I’m not expecting a gradual margin expansion over the next five years. I’m expecting Austin G2 to come online in Q4, for Keel to sign its first lease agreements, for CleanSpark to announce its first HPC deal, for Applied Digital to bring its pending 150 MW facilities online, and for CoreWeave to convert its backlog into recognized revenue in the second half of the year. These are events that will or will not play out over the next twelve to eighteen months. If these milestones are met, the price targets I’ve set are conservative. If not, we’ll know soon and can act accordingly. That's asymmetry handled well: not only in terms of potential returns, but also in the feedback the market will provide on whether we are on the right track.
Frequently Asked Questions
Artificial intelligence is built on energy and computing power, not software, and the physical infrastructure that supports it (megawatts, land with grid access, liquid-cooled data centers, solar cell factories) trades at neglected multiples while the market pays 40–50x for the visible layers of the AI business. The five positions express that same idea through different channels.
CoreWeave (CRWV) is the GPU computing provider; Applied Digital (APLD) is the hyperscaler provider; CleanSpark (CLSK) is a Bitcoin miner with an AI focus; Keel Infrastructure (KEEL) is the energy re-rating play (formerly Bitfarms); and T1 Energy (TE) is the generation-focused, domestic solar cell manufacturer (formerly FREYR Battery).
12- to 18-month price targets: CoreWeave $160 (+521% 3-month target price from $105.15), CleanSpark $26 (+491% 3-month target price from $17.40), T1 Energy $16 (+481% 3-month target from $10.80), Keel $8 (+471% 3-month target from $5.45), and Applied Digital $65 (+431% 3-month target from $45.60). These are proprietary targets based on execution scenarios, not the market consensus.
Because each one expresses the same idea at a different point on the risk-maturity curve, with different catalysts. In addition, T1 Energy acts as an implicit hedge: if energy prices rise structurally, it gains value while driving up the costs of the other four. You don’t have to get all five right for the portfolio to work.
Markets by Diego is the financial analysis platform of Diego García del Río, a Spanish economist and independent private investor, and founder of Hill Valley Consulting. He publishes asset analyses, macroeconomic reports, and strategies involving options and leveraged ETFs, along with tracking of actual trades in international markets.
More from Markets by Diego
Spanish 
English 
German 
French