Better Cloud Investment: Amazon or Nebius Group?

Nebius Q2 revenue grew 454% year over year, versus AWS growth of 37%, but Nebius had a 30% negative operating margin compared with AWS's 39% positive margin. Nebius is borrowing to expand, citing a payback period of just under two years for deals closed in Q2; Amazon plans to spend about $220 billion on data centers this year and can also fund expansion from its commerce profits. The article favors both stocks, framing Nebius as higher-upside and higher-risk and Amazon as the lower-risk choice, and suggests allocating 25% of a cloud portfolio to Nebius and 75% to Amazon.

Analysis

The key underwriting distinction is not growth versus growth; it is who can turn scarce compute into durable, cash-generative customer relationships. Amazon can bundle compute with storage, databases, and distribution, raising switching costs and potentially defending utilization when GPU supply loosens. Nebius is more exposed to a single economic bottleneck: keeping expensive accelerators powered and highly utilized while refinancing the build-out. Its reported deal payback is not equivalent to cash payback after capex, financing costs, idle capacity, and hardware obsolescence. Verify contract duration, cancellation terms, customer concentration, and whether reported payback includes the full infrastructure cost.

Over the next 1–3 months, watch Nebius utilization, backlog conversion, net debt, and financing terms; a missed capacity ramp could turn growth into dilution or refinancing risk. For Amazon, the counter-risk is that data-center investment outruns monetization, pressuring free cash flow before AWS returns validate the spend. Over 6–18 months, a supply response from Amazon, Microsoft, Google, and Oracle could compress standalone neocloud pricing, even if AI demand continues growing. Conversely, persistent GPU scarcity and long-duration contracted demand would extend Nebius’s window.

Contrarian read: headline growth rates exaggerate the relative opportunity because they omit scale, margins, and capital intensity. But treating Nebius’s negative operating economics as proof of failure may also be premature if contracted capacity converts to cash on schedule. Without valuation and financing details, the article does not establish which stock offers better risk-adjusted return.

Key Decisions for Investors

• Prefer AMZN as the core exposure; avoid treating its data-center budget as automatically accretive. Reassess if AWS growth or operating economics weaken while infrastructure spending remains elevated, or if free-cash-flow deterioration persists beyond the build-out period.
• Keep NBIS as a smaller, explicitly high-risk satellite rather than sizing it from revenue growth alone. Add only after verifying utilization, customer concentration, contract duration, net debt, and cash payback inclusive of capex and interest.
• For a relative-value expression, consider a modest long-AMZN/short-NBIS pair only if valuation and borrow costs support it; the thesis is lower funding and utilization risk at Amazon, not a claim that Nebius demand will fail. Trim or exit if Nebius demonstrates sustained cash conversion and secures financing without material dilution.
• Near-term catalyst watch: quarterly backlog conversion, capacity utilization, financing access, and capex-to-cash-flow trends. A deterioration in NBIS financing terms or missed deployment milestones would invalidate the satellite thesis; sustained contracted utilization and improving cash economics would challenge the relative short.