Digi Power X secures $1.1B data center contract with AI chip maker Cerebras Systems

Digi Power X signed a Master Services Agreement with Cerebras Systems to colocate a purpose-built 40-megawatt AI data center campus in Columbiana, Alabama. The initial 10-year term is valued at about $1.1 billion, with total potential contract value reaching up to $2.5 billion including renewals. The deal meaningfully expands the company’s revenue visibility and underscores demand for AI infrastructure.

Analysis

This is less about one contract and more about a validation event for the AI-infrastructure monetization stack. The market typically underprices how quickly a credible anchor tenant can compress financing risk, improve project bankability, and pull forward additional hyperscale or inference workloads to the same campus or adjacent sites. If DGXX can convert this into debt capacity, utility leverage, and a second customer, the equity could re-rate on a multiple expansion rather than just a revenue step-up.

The second-order winner is the local ecosystem: power, cooling, electrical equipment, and construction vendors with constrained near-term capacity should see improved pricing power as AI colocation demand keeps shifting from speculative announcements to contracted load. The main loser is the “power-only option” market for smaller data-center developers that lack an enterprise-grade tenant or a differentiated site; in a tighter capital environment, this deal raises the bar for who can finance 40MW+ builds at acceptable spreads.

The key risk is execution, not headline value. Multi-year AI infrastructure agreements often disappoint on timing: interconnect delays, transformer lead times, permitting, and customer ramp slippage can push cash flow realization 6-18 months behind press-release expectations. If broader AI capex sentiment cools or if financing costs stay elevated, the market may eventually discount this as a long-dated option rather than near-term earnings power.

Contrarian view: the move may be underestimating dilution risk. For smaller infrastructure operators, the fastest path to delivering large campuses is often equity issuance, JV structures, or project-level leverage that transfers upside to lenders/partners if execution is slow. So while the contract is bullish on enterprise value, near-term stock performance could still be capped if investors fear funding the buildout before the cash starts coming in.

Key Decisions for Investors

• Long DGXX on pullbacks over the next 2-6 weeks; use the headline as a catalyst but size for execution risk. Target a 15-25% upside move if the company can provide financing or tenant-detail follow-through; cut if the story shifts to heavy dilution.
• Prefer a pair trade: long DGXX vs short a basket of non-contracted AI data-center developers with limited tenant visibility over 1-3 months. The spread should favor names with hard backlog and power access if the market starts rewarding bankability over narrative.
• Buy short-dated DGXX call spreads only after confirmation of project financing or additional customer disclosures. This limits downside versus common equity because the real upside likely comes from a second leg of announcements, not this press release alone.
• Watch for confirmation in utility/interconnect and equipment names over the next quarter; if DGXX is real, the second-order winners should be electrical infrastructure suppliers and power-delivery vendors before the broader AI complex re-prices.
• If DGXX rallies sharply on the headline without follow-through, fade the move into strength via a tactical short or hedged call overwrite. The market may be overpaying for a long-dated revenue stream before capex, timing, and dilution are clarified.