An Ex-OpenAI Researcher With $20 Billion Just Built a Neocloud Trifecta: Here Are the 3 Stocks He's Buying.

The article argues that Nebius, CoreWeave, and IREN are attractive neocloud plays tied to AI infrastructure demand, with estimated 2026 revenue growth of 550%, 147%, and 46%, respectively. It says the stocks trade at 6x, 3x, and 7x next year's sales, suggesting valuations may be cheap relative to long-term growth. The piece is largely bullish commentary on AI infrastructure bottlenecks and investor positioning rather than a company-specific catalyst.

Analysis

The market is starting to price neoclouds less like venture-style hypergrowth and more like capacity-constrained infrastructure assets, which is exactly where mispricings can persist. The key second-order effect is that GPU supply and power availability, not demand for AI compute, become the gating factor; that favors the firms with the best access to Nvidia ecosystem support and cheapest marginal watts, while penalizing generalist cloud players that cannot re-architect fast enough. In that setup, NBIS/CRWV/IREN are not just competing with hyperscalers — they are competing for scarce financing, long-dated power contracts, and time-to-deploy before the next wave of customers is already spoken for.

The biggest risk is that the current valuation discount is implicitly assuming linear expansion of data center utilization, but these businesses can stay capital-hungry for longer than revenue ramps. If financing conditions tighten or GPU delivery slips by even 1-2 quarters, the multiple compression can overwhelm growth, especially in names where next-year sales are still a small base and forward estimates are heavily back-end loaded. Over a 3-12 month horizon, the market will likely reward whichever name proves it can translate contracted demand into operating leverage rather than just headline revenue growth.

The consensus is underestimating how important asset-quality dispersion will become inside the neocloud basket. Investors are treating NBIS, CRWV, and IREN as a thematic group trade, but their path to durable margins likely diverges based on power cost, customer concentration, and balance-sheet flexibility; that should eventually create a winner/loser hierarchy even if the theme stays hot. The more interesting trade is not whether neoclouds work, but which business can self-fund the next buildout without repeatedly diluting equity holders.

Key Decisions for Investors

• Long NBIS / long CRWV / long IREN basket on 6-9 month horizon, but size each leg equally and rebalance quarterly; the risk/reward is attractive while all three trade at compressed sales multiples relative to growth, but the basket reduces idiosyncratic execution risk.
• Prefer NBIS over CRWV on a relative basis for a 3-6 month trade if financing markets stay constructive; the market is likely to reward the name perceived to have the cleanest path from contracted demand to margin expansion, while CRWV remains the most exposed to 'growth at any cost' skepticism.
• Use IREN as the highest-beta call option on power scarcity over 12 months; pair it against a short in a broader hyperscaler proxy if you want to isolate the theme, since IREN's renewable-power positioning could earn a premium if power becomes the bottleneck investors focus on next.
• Do not chase after a one-day gap up; wait for 10-15% pullbacks or failed post-earnings reactions to establish positions, because these names are likely to remain volatile around deployment and financing updates.
• If you want a cleaner expression, consider a long NBIS / short AMZN pair for 6-12 months: the thesis is that specialized AI infrastructure can re-rate faster than general-purpose cloud if AI workloads stay capacity-constrained, though the short is more about relative multiple expansion than fundamental weakness at AMZN.