Morgan Stanley initiated coverage of three publicly traded bitcoin BTC$70,409.49 mining companies on Monday, backing two names tied to data center leasing while taking a more cautious stance on a miner focused on bitcoin exposure.

Analyst Stephen Byrd and his team started coverage of Cipher Mining (CIFR) and TeraWulf (WULF) with Overweight ratings and set price targets of $38 and $37, respectively. Shares of CIFR are higher by 12.4% Monday to $16.51, while WULF is ahead 12.8% to $16.12.

He also initiated coverage of Marathon Digital (MARA) with an Underweight rating and an $8 target. Shares of MARA are marginally higher on Monday at $8.28.

Byrd’s core argument rests on viewing certain bitcoin mining sites less as crypto bets and more as infrastructure assets. Once a mining company has built a data center and signed a long-term lease with a strong counterparty, he wrote, the asset is better suited to investors who value steady cash flow than to traders focused on bitcoin price swings.

“At a macro level, once a bitcoin company has a built-in data center and entered into a long-term lease with a creditworthy counterparty, that DC’s natural investor habitat is not among bitcoin investors but among infrastructure investors,” Byrd wrote, adding that such assets should be valued for “long-term, stable cash flow.”

To make the point concrete, Byrd compared these facilities to data center real estate investment trusts such as Equinix (EQIX) and Digital Realty (DLR), which he described as “the closest comparable companies to consider when valuing DC assets developed by bitcoin companies.” Their shares trade at more than 20 times forward EBITDA, meaning investors are willing to pay over $20 for every $1 of expected annual operating cash flow because those firms offer scale, diversification and steady growth.

Byrd does not expect data centers developed by bitcoin companies to trade at similar levels, “primarily because these data center REITs have growth potential that a single DC asset does not provide.” Still, he sees room for higher valuations than the market currently assigns.

Cipher sits at the center of that view. Byrd described the company’s data centers as suitable for what he called a “REIT endgame.” “We use the phrase ‘REIT endgame’ to describe our valuation approach because, ultimately, these contracted DCs should be owned by REIT-like investors that appropriately value long-term, low-risk contracted cash flows,” he wrote.

In a simple scenario, a Cipher site that shifts from self-mining bitcoin to leasing space to a large cloud or computing customer could resemble a toll road. Cash flows become predictable. The role of bitcoin fades.

TeraWulf earned a similar framework. Byrd pointed to the company’s history of signing data center agreements and to management’s background in power infrastructure. “TeraWulf has a strong track record of signing agreements with data center customers, and the management team has extensive experience in building a wide range of power infrastructure assets,” he wrote.

He expects the firm to convert sites without bitcoin-to-data-center contracts at a present value of about $8 per watt. His base case assumes the company succeeds in roughly half of its planned annual data center growth of 250 megawatts per year over 2028-2032. In a more optimistic scenario, he assumes that the success rate rises to 75%.

The tone shifted with Marathon Digital. Byrd argued that the company offers “lower potential upside driven by bitcoin-to-DC conversions.” He cited Marathon’s hybrid strategy, which combines mining with data center ambitions rather than fully repurposing sites, along with its focus on maximizing exposure to bitcoin’s price, including issuing convertible notes and using the proceeds to buy bitcoin.

Marathon’s limited history of hosting data centers also weighed on the view. “For MARA, bitcoin mining economics are the dominant driver of the stock’s value,” Byrd wrote.

That focus carries risk. “Fundamentally, we see significant risks to profitability of bitcoin mining, both in the near and long terms,” Byrd added, noting that “the historical ROIC of the bitcoin mining business has been unattractive.”

The coverage lands as investors debate whether bitcoin miners should evolve into power and computing landlords. Morgan Stanley’s answer is selective. Where long-term leases and infrastructure discipline take hold, Byrd sees value. Where mining remains the core business, he sees fewer reasons to expect outsized gains.