Applied Digital (APLD) and IREN are pitched as AI “neocloud” data center plays, but both show heavy losses amid rapid buildout. Applied Digital grew FY2026 revenue to $611.3M (+183.7% YoY) yet posted a $250.3M net loss, with negative FCF of $2.8B and very high leverage (debt-to-equity 2.9x); IREN grew FY2026 revenue to $707.0M (+41.1% YoY) but reported a $702.6M net loss largely from $638.8M non-cash impairments, with negative FCF of $2.2B and leverage of 1.9x. The article concludes IREN is the preferred pick due to vertical integration and a lower forward P/E (137.6x vs 526.3x) while highlighting substantial capital intensity and transition/regulatory/power risks for both.
Both names are really financed claims on AI buildout, not clean operating businesses. That matters because the market usually overpays for “capacity growth” and underprices the cost of capital: if growth is funded with repeated equity/debt raises, the economic winner is often the chip vendor or hyperscaler, not the neocloud equity holder. APLD’s concentration risk makes it the more brittle structure; one customer/accounting change can move the equity far more than the headline revenue trajectory suggests. IREN has more balance-sheet flexibility and the stronger scarcity asset — power access — but vertical integration does not eliminate dilution risk if capex stays ahead of operating cash flow.
The near-term tape should be driven by financing conditions rather than AI demand headlines. Over days to weeks, investors will likely favor the name that appears least dependent on a fresh raise; that creates a modest relative tailwind for IREN versus APLD. Over 1-3 months, the real catalyst is whether either company can show non-dilutive project funding, on-time MW delivery, and improved conversion of signed capacity into billable revenue; any slip there can re-rate these stocks sharply lower because the market is paying for execution certainty that is not yet earned. Over 6-18 months, the structural question is whether the AI supply chain keeps migrating toward better-capitalized balance sheets (MSFT, NVDA, DELL) and away from levered infrastructure equities.
Consensus is missing that “AI infrastructure” is not a single trade. The scarcity premium sits in power, permits, and cheap financing; the equity value accrues where those are cheapest, not necessarily where revenue growth is fastest. My base case is that the article’s bullish framing is directionally right on AI demand but overstates the investability of these two stocks at current valuations, especially if rates stay sticky or credit spreads widen. A sustained drop in funding costs or a credible path to positive FCF would falsify the bearish view; absent that, these are trading vehicles, not long-duration compounders.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request TrialOverall Sentiment
mildly negative
Sentiment Score
-0.12
Ticker Sentiment