Better Artificial Intelligence Stock: Applied Digital vs. Super Micro Computer

Super Micro Computer is presented as the preferable AI-infrastructure investment, combining FY2026 revenue growth of 77.8% to $39.1B with $2.2B in net income and a comparatively low 0.6x price-to-sales multiple. Management forecast FY2027 sales of $65B-$72B, implying growth of roughly 66%-84%, although free cash flow was negative $7.0B amid inventory and capacity investment. Applied Digital grew revenue 167% to $611.3M and has secured $36B of leases over 15 years, but posted a $249.2M loss, negative $2.8B free cash flow, and more than $5B of debt; both companies face customer-concentration and execution risks, while SMCI also faces legal, reporting, and export-control concerns.

Analysis

The economically cleaner AI-infrastructure exposure is SMCI: it converts accelerator availability into near-term system revenue without bearing the multiyear construction, power-procurement, and refinancing risk embedded in APLD’s model. APLD’s contracted backlog has option-like value, but equity holders are subordinated to a capital stack that will likely require recurring debt and/or equity issuance before projects generate stabilized rent; higher rates, construction slippage, or a single tenant’s financing stress can therefore impair equity value disproportionately. CRWV credit spreads, data-center power interconnection milestones, and APLD’s incremental funding terms matter more than reported revenue growth over the next 1-3 months.

SMCI’s low sales multiple is not automatically a bargain: its earnings power depends on retaining scarce GPU allocations and avoiding a mix shift toward lower-margin commodity servers as ODMs, Dell (DELL), HPE, and hyperscalers internalize rack integration. The key 6-18 month debate is whether liquid-cooling and rack-scale integration preserve gross-margin differentiation as NVIDIA platform cycles normalize. Consensus may underappreciate that both names have weak cash conversion, but APLD has duration risk while SMCI has working-capital risk; the latter can unwind rapidly if shipment cadence improves, whereas the former remains dependent on external capital.

Near term, SMCI is vulnerable to litigation/export-control headlines and any GPU supply disruption, creating a better entry on event-driven weakness than chasing AI beta. For APLD, the upside case requires independently verified project completion, tenant funding durability, and a path to financing capex without material dilution; absent these, backlog should be valued at a meaningful execution discount rather than as annuity-like revenue.

Key Decisions for Investors

• Initiate a 3-6 month pair: long SMCI / short APLD, dollar neutral. The trade expresses superior balance-sheet resilience and faster revenue realization; target a 15-25% relative return, with review if APLD secures non-recourse project financing at materially better-than-expected terms or SMCI cuts FY2027 revenue/gross-margin guidance.
• Buy SMCI only on post-results or regulatory-driven weakness, using a 6-12 month horizon rather than adding ahead of an unverified demand print. Falsify on evidence that GPU allocation is lost, inventory days continue expanding while revenue guidance stalls, or gross margin falls below management’s sustainable range for two consecutive quarters.
• Avoid directional APLD longs until three watch items clear: construction/power milestones, tenant credit support—especially CRWV financing conditions—and the financing source for remaining capex. APLD is a potential tactical long only after those disclosures if the equity raise/refinancing discount is absorbed; otherwise dilution and duration risk dominate.
• Monitor CRWV debt and equity performance as a leading indicator for APLD lease collectability and expansion economics. A sustained widening in CRWV credit risk or a reduction in its capacity commitments should trigger a tighter APLD short risk limit or renewed downside positioning.