Leopold Aschenbrenner's Situational Awareness Fund Just Bought This Nvidia-Backed Neocloud Stock

Nebius reported 1Q revenue up 684% year over year and also delivered EBITDA and adjusted EPS beats of 43% and 50%, respectively, after which the stock rose 24%. The article highlights a 5.6% stake by Leopold Aschenbrenner’s Situational Awareness Fund, Nvidia’s planned $2 billion investment by 2030, and Nebius’s $2.6 billion Bloom Energy deal securing 250 MW of power capacity. Overall, the piece is constructive on Nebius’s AI infrastructure growth but flags long-term execution, capital intensity, and profitability risks.

Analysis

The investable signal here is not “AI demand is strong” — that is consensus — but that AI infrastructure is becoming a capacity-constrained utility business with contractable returns. NBIS’s edge is not just access to GPUs; it is being early in the queue for scarce power, rack density, and next-gen networking, which should support pricing power longer than most investors expect. That said, the market is now effectively underwriting a multi-year flawless execution path: if capital intensity rises faster than contracted utilization, today’s multiple can compress even while revenue grows.

Second-order winners are the picks-and-shovels around the buildout. NVDA benefits from being embedded at the architecture level, but the more interesting readthrough is to suppliers that reduce deployment friction — power, cooling, storage, and interconnect. BE’s role matters less as a standalone growth story and more as a de-risking mechanism for gigawatt-scale expansion; if its technology becomes the default bridge solution, it could earn a scarcity premium, but if power procurement normalizes, the market may re-rate the entire “AI power” basket lower.

The key risk is timing mismatch: capacity is being financed now, while monetization depends on sustained GPU scarcity and inference demand over the next 12-24 months. Any softening in hyperscaler capex, faster-than-expected GPU supply normalization, or evidence that new AI workloads are migrating back to larger cloud platforms would hit NBIS first and the broader neocloud trade second. The consensus is probably underestimating how quickly margins can reprice once competition arrives, because infrastructure stories often look linear until utilization slips a few points.

From a trading standpoint, this is better expressed as a relative-value and optionality trade than a naked long. The asymmetric setup is long NBIS on pullbacks versus a basket of slower-moving hyperscalers, while hedging with a short in any GPU-scarcity proxy that is most exposed to multiple compression if supply catches up. For investors already long NVDA, the cleaner incremental trade is to pair NVDA with a short in a broad AI-infra ETF or unprofitable neocloud names, since the market is likely overpaying for “capacity narrative” while still underpricing supplier leverage and execution risk.

Key Decisions for Investors

• Buy NBIS on a 3-5% pullback and size it as a 6-12 month momentum/value hybrid; target is continued multiple support if contracted utilization stays tight, but cut if utilization commentary weakens or capex begins outrunning bookings.
• Pair trade: long NVDA / short a basket of high-beta AI infrastructure names with weak balance sheets over the next 1-2 quarters; thesis is that the market will differentiate durable platform leverage from capacity speculation.
• Add BE only as a tactical trade, not a core hold; use it as a 6-9 month power-scarcity call option, but exit if customer concentration or project timing risks show up in guideposts.
• For options, consider NBIS call spreads 3-6 months out to capture upside from continued AI infrastructure repricing while limiting drawdown if the market rotates away from unprofitable growth.
• If you already own hyperscalers (AMZN/MSFT/GOOGL), use NBIS strength as a hedge signal rather than a buy signal: the better expression may be to trim some beta and rotate toward suppliers with direct pricing power in the AI stack.