WhiteFiber (WYFI) Q2 2026 Earnings Call Transcript

WhiteFiber reported Q2 revenue of $28.8M (+54% Y/Y), with cloud services revenue rising to $23.8M (+$7.2M Y/Y) and colocation revenue reaching $4.7M, as NC1 entered active customer deployment (20MW available, full 40MW targeted by end of August). Despite a $15M net loss (+$? driven by higher depreciation and interest expense), adjusted EBITDA improved to $5.5M (from $3.3M), aided by GM of ~59% and $143M of deferred revenue from customer prepayments. Management highlighted $865M of 10-year contracted revenue for 40MW at NC1 plus $540M of cloud contract backlog, and disclosed new GPU-as-a-service agreements totaling $165M (Base 10), $108M (Prime Intellect), $87.5M (Iceland), and $160M (Paris), while noting NC1 ramp delays tied to switchgear delivery and commissioning and ~$2.2M bad-debt expense from a prior customer termination.

Analysis

The near-term trade is less about quarterly earnings power and more about whether WhiteFiber converts backlog into financeable, repeatable cash flows without another delay. The real upside comes if NC1 financing closes: that would recycle equity, de-risk the next site, and signal the balance sheet can support a roll-up of retrofit assets rather than a one-off project story. If financing slips again, the market should compress the multiple quickly because the equity remains dependent on project-level execution, not operating cash generation.

Second-order winners are the GPU and power ecosystem, but not evenly. NVIDIA should benefit from continued long-duration allocation discipline and a customer base willing to pre-commit to next-gen parts, while Duke’s upside is mostly option value from incremental load, not an earnings step-up. The bigger loser is the speculative neo-cloud / capacity-leasing cohort: WhiteFiber is signaling that scale, credit quality, and deployment certainty matter more than cheap capacity, which can pressure weaker operators that rely on short-dated demand and ad hoc financing.

The contrarian issue is that management is talking like a platform company, but the market will still underwrite it like a capital-intensive developer until proof arrives that managed services and cross-site networking produce durable, high-ROIC revenue. The cross-DC technology could be strategically important, but near term it is more likely an internal utilization tool than a standalone monetization engine. Over 6-18 months, the thesis is falsified if NC1 ramps but financing stalls, if customer prepayments do not translate into lower equity needs, or if 2027 power promises prove too slow to monetize.

Key Decisions for Investors

• Small tactical long WYFI into confirmation of NC1 project financing; use a tight catalyst stop if documentation slips past the next 30-45 days or if the company signals unfavorable terms.
• Pair: long WYFI / short BTBT as a quality-vs-story trade on AI infrastructure execution, with the thesis that financed, contracted capacity deserves a premium to higher-beta peers with less visible project economics.
• Add on weakness to NVDA on any post-call softness; the best second-order read-through is sustained allocation discipline and multi-year GPU demand, not a one-quarter revenue pop.
• Avoid chasing DUK on this print; the incremental load is supportive but too small to move near-term earnings, so treat it as a watch item for utility-capex optionality rather than a direct trade.
• Set an alert on WYFI around financing close and the September commercial launch of cross-DC networking; if neither proves monetizable by the next update, de-rate the story and reconsider the long.