⚠ Debt, Depreciation & the Path to ProfitHigh threat

CoreWeave (CRWV) — threat to the moat

Never profitable, with interest and chip-aging consuming the top line — the path to profit runs uphill against its own balance sheet.

The second great risk is financial: CoreWeave has never earned a profit, funds its growth with a mountain of debt secured against rapidly-depreciating chips, and its path to sustainable profitability is genuinely uncertain. Despite explosive revenue growth — $5.1 billion in 2025, up 168% — the company lost roughly $1.2 billion1 that year, and its losses have widened in absolute terms even as revenue soared, because the two costs that define its model overwhelm everything else: interest on its debt (over $500 million in a single recent quarter) and depreciation of its GPU fleet (over $1 billion in a quarter). A business whose financing and asset-consumption costs swamp its revenue growth is not obviously converging on profit; it is running hard to stand still, and whether scale eventually tips it into sustainable earnings is an open, unproven question.

Interest plus depreciation as a share of revenue (%)63.9%202471.8%202579.9%H1 2026CoreWeave 10-K FY2025 and 10-Q Q2 2026: (interest expense, net + depreciation and amortization) / revenue
The two costs that come with debt-financed GPUs are taking a rising share of revenue, the opposite of the path to profit.

The structure is inherently fragile. CoreWeave borrows against GPUs that lose value quickly, so it carries heavy fixed debt secured by wasting collateral, and it must continually raise more capital — roughly $28 billion in a year2 — to keep building, because it generates no cash flow to fund itself. This works while GPUs hold value, utilization stays high, contracted revenue flows, and capital markets stay open and cheap; it is dangerously exposed if any of those reverses. And all of it rests on the AI capital-spending cycle continuing — the same untested boom that underpins the whole AI-infrastructure complex, whose eventual normalization is a question of when, not whether. In a downturn, the leverage and depreciation that are manageable in the boom would compound viciously: falling utilization and GPU values against fixed debt, tightening capital access, and pressured, concentrated, entangled counterparties, all at once.

The bull case is that CoreWeave's enormous backlog and scale will drive revenue high enough, fast enough, that it grows into profitability and its financing proves sound — that the losses are the investment phase of a business that will earn handsomely once its capacity is built and utilized. That is possible, and the backlog lends it real support. But an investor must weigh the honest reality: CoreWeave is a deeply-unprofitable, heavily-indebted company whose model depends on depreciating assets, perpetual refinancing, concentrated counterparties, and an untested demand cycle all holding together — a structure that has produced spectacular growth and could produce spectacular losses, and whose path to durable profit is a hope supported by a backlog, not a demonstrated fact. The leverage and depreciation that fund the boom are precisely what would make a bust severe, and the company has never been tested by anything but boom.

The number that tests this threat
Reported
The debt-and-depreciation load that swamps the growth
Interest expense $640M and depreciation $1.4B in Q2 2026; net loss $626M, from $290M

Interest more than doubled in a year; the path to profit needs revenue to outgrow interest and depreciation together, and in Q2 2026 it did not.

Source: CoreWeave Form 10-Q, quarter ended 30 June 2026 ↗
References
  1. Reported2025: revenue $5.1B (+168%), net loss ~$1.2B and widening.
    CoreWeave Form 10-K, fiscal 2025 — revenue $5.13B (+168%), net loss ~−$1.2B; customer concentration disclosed (largest customer ~2/3 of revenue) — FY2025 · publ. early 2026 · source ↗
  2. Reported~$28B of capital raised in a year to keep building.
    CoreWeave Q1/Q2 2026 earnings releases — Q2 revenue ~$2.5B (+111%); revenue backlog $99.4B (Mar 2026, from $66.8B at end-2025); 2026 capex guided $31–35B; ~$28B of financing raised in 12 months; quarterly interest expense >$500M; ten clients >$1B each — Q1-Q2 2026 · publ. 2026 · source ↗
Sources
Generated September 23, 2026