Digi Power X (DGXX) Q2 2026 earnings review
The AI Pivot Begins, But Profitability Lags
Digi Power X officially crossed the starting line in its transition to an AI infrastructure company, generating its first $1.1M in GPU bare-metal compute revenue from just five weeks of operations. However, the legacy colocation and energy business continues to shrink, dragging total Q2 revenue down to $6.6M. The bottom line took a severe hit, with net loss ballooning to $14.4M. While management points to a positive Adjusted EBITDA of $3.3M, that figure requires ignoring massive non-cash charges. With $1.1 billion in contracted future revenue, the investment narrative is entirely focused on a colossal 2027 scale-up, demanding flawless execution and aggressive capital deployment.
🐂 Bull Case
The NeoCloudz platform successfully generated $1.1M in high-margin AI compute revenue in just five weeks with 100% uptime on B300 GPUs, proving the viability of their infrastructure transition.
The company has secured $1.1 billion in contracted AI data center revenue, anchored by the Columbiana, Alabama facility, providing deep visibility into future cash flows if execution stays on track.
🐻 Bear Case
Net loss cratered to $14.4M. Management is propping up the 'profitability' narrative using Adjusted EBITDA ($3.3M), which conveniently ignores $5.8M in share-based compensation and a $5.0M warrant fair value loss.
Scaling from an annualized revenue base of ~$26M today to a targeted $250-$300M run-rate by Q3 2027 requires flawless construction of Tier III data centers and navigating high-interest debt markets for necessary CapEx.
⚖️ Verdict: ⚪
Neutral. The strategic pivot from crypto mining to B300 GPU AI infrastructure is undeniably the right move, and securing $1.1B in backlog is a massive win. However, the current financial reality—shrinking total revenues, massive GAAP losses, and heavy reliance on future debt syndication—makes this a 'show-me' story strictly tied to 2027 execution.
Key Themes
GPU-as-a-Service Monetization Initiated
The most important development this quarter is the $1.1M generated from the NeoCloudz platform. Utilizing B300 GPU bare-metal infrastructure, the 0.6 MW deployed capacity ran at 100% uptime. This establishes the baseline for the company's planned 10 MW expansion in 2027, transitioning the company away from volatile digital currency mining into high-density AI inference clouds.
Alabama Tier III Campus on Schedule
Construction at the flagship Columbiana site is progressing on schedule. Phase 1 (15 MW) is expected to be delivered in December 2026, and Phase 2 (25 MW) in March 2027. This facility is the backbone of the $1.1 billion contracted AI data center agreement, with expansion options potentially pushing the contract value to $2.5 billion.
Poor Earnings Quality
Management loudly touted a positive Adjusted EBITDA of $3.3M, but this metric heavily masks underlying financial leakage. The GAAP net loss of $14.4M was driven by $5.8M in share-based compensation (a real dilutive cost to shareholders) and $5.0M in warrant fair value losses. Relying on Adjusted EBITDA presents a distortedly optimistic view of operational health.
CapEx Demands and Debt Syndication Risks
The company invested $110M year-to-date in equipment and infrastructure. While they hold $142.4M in cash, building out Phase 2 in Alabama and expanding GPU compute requires massive capital. The company has engaged Goldman Sachs to syndicate debt financing. In the current macro environment, securing large-scale project debt may come with punitive interest rates or heavy covenants.
US Data Centers Inc. Valuation Event
The company's subsidiary, US Data Centers Inc. (USDC), raised outside capital at a $125 million pre-money valuation. Digi Power X retained an approximate 48% ownership stake. This off-balance-sheet capital raise provides non-dilutive validation of the modular AI data center platform's intrinsic value.
Other KPIs
Legacy colocation and mining generated $3.6M, energy sales contributed $1.9M, and the newly launched GPU rental segment added $1.1M. The legacy business continues to serve as a low-margin bridge while the AI infrastructure is constructed.
The balance sheet is currently well-capitalized with zero debt and $142.4M in cash as of June 30 (increasing to ~$150M by August). This war chest is critical given the $110M already deployed year-to-date and the imminent capital requirements for the Alabama buildout.
Guidance
Accelerating wildly. Management targets this staggering run-rate by Q3 2027, driven by the energization of the Alabama facility and GPU expansion. This requires executing a roughly 10x multiplier on current trailing-twelve-month revenues.
Accelerating. This forms the hard floor of the 2027 outlook, derived entirely from existing contracts and scheduled deployments, shielding it partially from broader market volatility if construction deadlines are met.
Accelerating. The initial 0.6 MW pilot proved successful. Management intends to scale this segment significantly in 2027 to capture the high-margin gap between contracted colocation and full AI-compute service provision.
Key Questions
Debt Facility Covenants
With Goldman Sachs engaged to syndicate debt for the Alabama Tier III data center, what is the expected cost of capital, and will the facility require any equity dilution or warrants attached?
Bridging the Run-Rate Gap
You guided for $140M in contracted revenue run-rate for 2027, but a total target of $250-$300M. Exactly how much of that $110M-$160M gap requires securing net-new customers versus simply energizing uncontracted capacity?
Margin Profile of AI Colocation
As the $1.1 billion, 10-year AI colocation contract activates in December 2026, what steady-state gross margin should investors model for this segment compared to the highly profitable bare-metal GPU rentals?