AI Data Center Stocks Face a Cost-of-Capital Reality Check

The selloff in AI infrastructure stocks is exposing a basic mismatch: many companies promise years of data center demand, while investors must fund the construction now. Cantor Fitzgerald still rates seven miners-turned-data-center operators Overweight, but it cut price targets as higher rates and execution risks changed the value of distant cash flows.

The most bullish calls focused on Hut 8, TeraWulf and Core Scientific. According to Stocktwits, Cantor argued that the market had punished the group too broadly despite differences in contracts, management quality and project execution.

Funding is the common constraint

These businesses need large amounts of capital before a data hall begins producing revenue. Higher Treasury yields raise the discount rate applied to future cash flow, while project delays or cost overruns increase the amount of financing needed.

CoreWeave offers a useful credit-market warning even though it follows a different business model. The company issued 9.75% senior notes due 2031 earlier in 2026 and later raised additional convertible debt. CNBC reported that the yield on the 2031 notes approached 16% during the latest selloff. A secondary-market yield far above the original coupon signals that bond investors demand much more compensation for risk.

That does not prove that every AI infrastructure developer has the same credit profile. Hut 8, TeraWulf, Core Scientific, Cipher Mining and CleanSpark have different contracts, sites, power arrangements and balance sheets. It does show why the sector can fall together when financing conditions tighten.

What separates the winners

The strongest projects should have contracted customers, committed power, credible construction budgets and financing that does not depend on continuously rising share prices. Announced megawatts without those supports are less valuable.

Investors should compare contracted revenue with remaining capital spending, debt maturities and the amount of equity that may still be issued. The analyst targets matter less than whether each company can reach operation without diluting shareholders or refinancing at punitive rates.

The sector may still produce large winners. The current selloff is a reminder that AI demand alone is not enough. The financing structure determines how much of that demand reaches common shareholders.