Cantor Cuts AI Data-Center Targets but Keeps WULF, HUT and CORZ at Overweight

Cantor Fitzgerald cut price targets across seven AI data-center developers but kept Overweight ratings on all of them, a combination that captures the sector's current tension. Analysts still see valuable contracted power and computing capacity, while public markets are applying a larger discount for financing and execution risk.

The sharpest concerns are familiar to capital-intensive infrastructure. Higher interest rates raise the cost of debt, higher capitalization rates reduce the present value of future lease cash flows, and construction overruns can consume equity before a project begins generating revenue. Local opposition and political uncertainty add another layer of delay risk.

Cantor called TeraWulf, Hut 8 and Core Scientific unusually attractive after the selloff, according to Stocktwits' account of the research note. That judgment is an analyst opinion, not a company forecast. The same note lowered targets, including a one-third reduction for Cipher Mining, which shows that bullish ratings can coexist with weaker valuation assumptions.

The companies are not interchangeable. Contract quality, tenant credit, power availability, financing structure and delivery timing determine how much of a site's headline megawatt capacity becomes cash flow. Investors should compare signed leases and funded construction rather than treat every former bitcoin miner as the same AI exposure.

TeraWulf's recent expansion of contracted power at its Muskie campus from 500 megawatts to 1 gigawatt adds optional capacity, but it does not by itself create tenant revenue. The sector's rerating will depend on signed leases, financing costs and on-time energization. Until then, lower targets are a reminder that asset value can rise while equity value remains highly sensitive to capital markets.