Digi Power X: Power Scarcity Makes Cerebras Contract More Valuable Than Market Thinks

Digi Power X (DGXX) is shifting from crypto mining to AI infrastructure, backed by a $1.1B take-or-pay contract with Cerebras for 40 MW in Alabama. Management targets a $140M+ annualized contracted run rate by 2027, supported by a $150M cash balance and no debt. While the colocation/GPU rental model should provide predictable revenue with upside from compute scarcity, continued equity raises are a dilution risk and management prefers debt financing for future growth.

Analysis

This should be viewed less as a pure AI growth story and more as a monetization of stranded power optionality. The real asset is not the press release language, it is the interconnect, real estate, and power economics; that makes the closest winners other miners with grid access and convertible infrastructure such as IREN, CORZ, and WULF, while ASIC-only miners with no AI path risk multiple compression as investors re-rank scarce-power assets versus obsolete hardware.

The financing structure matters more than the contract headline. Take-or-pay improves bankability only if the counterparty stays committed through buildout, so the next 1-3 months are about project finance, permits, and whether management can avoid another equity raise; if they need to fund growth with stock, the market will treat this as a dilution story, not an infrastructure compounding story. Over 6-18 months, the main bear case is supply response: if too many former miners pivot into AI hosting, colocation rents and GPU lease economics can normalize quickly, shrinking the scarcity premium.

Consensus may be overpaying for the visibility of contracted revenue without discounting execution risk and capital intensity. The equity could re-rate on the announcement, but the durable upside only exists if DGXX proves it can fund expansion with debt or project-level capital rather than repeated equity issuance. Any thesis is falsified by delayed commissioning, materially dilutive financing, or a 2027 run-rate that does not convert into real free cash flow after power, maintenance, and interest.

Key Decisions for Investors

• Avoid chasing DGXX on the initial headline; wait for the next financing disclosure. Only consider a small long if the company secures non-dilutive project debt or JV capital and share count growth is capped.
• Relative-value pair for 1-3 months: long IREN or CORZ versus short DGXX, expressing a preference for scaled AI/power platforms with cleaner funding visibility over microcap execution risk.
• Keep legacy crypto-mining names with weak AI conversion optionality under review for shorts or underweights; the market may start pricing a bifurcation between power-asset owners and pure ASIC miners.
• If the next filing shows equity-funded expansion or delays in the Alabama buildout, treat that as a sell signal on DGXX; the stock is vulnerable to a sharp repricing once dilution becomes explicit.
• Watch CBRS only as a demand-validation proxy: if this becomes the first of multiple similar take-or-pay wins, the sector could deserve a higher scarcity multiple; if not, this remains a one-off customer concentration event.