Leopold Aschenbrenner's Situational Awareness Fund Bought Bloom Energy Stock Before a 176% Run. Here Is the Artificial Intelligence (AI) Stock He Owns That I Think Will Go Parabolic Next.

The article highlights Situational Awareness LP’s AI-focused positioning, including a $875 million Bloom Energy stake (10.1 million shares plus 408,500 call options) and a 20.2 million-call-option position in Intel. Bloom Energy rose from an estimated average Q4 2025 purchase price of about $105 to $290.81 by May 1, implying roughly a 176% gain, underscoring the fund’s early AI infrastructure bet. The piece is broadly bullish on Intel and the AI power/compute buildout, but it is primarily commentary rather than fresh company-specific news.

Analysis

The market is starting to price AI as a power-and-fab scarcity story rather than a pure GPU story, and that shifts the profit pool downstream. Bloom is the cleaner expression of the bottleneck trade because it monetizes urgency: data center operators can’t wait on utility interconnects, so on-site generation becomes a time-to-capacity solution, not just an efficiency play. That said, the move likely pulled forward a lot of near-term enthusiasm; if order timing slips or financing costs stay elevated, the next leg depends on evidence of repeatable backlog conversion rather than narrative momentum.

Intel is a more complex second-order beneficiary because the market still treats it as a turnaround while the option value is in domestic capacity, packaging, and inference-era CPU attachment. If AI deployment becomes more heterogeneous—more custom ASICs, more edge inference, more power-constrained racks—then the winners are not just accelerator vendors but the companies that sit in the orchestration layer and manufacturing stack. The hidden risk is that Intel’s upside is highly path-dependent: if foundry wins remain incremental or gross margin dilution persists, the stock can re-rate only on delayed evidence, which makes this a slower-moving trade than Bloom.

The underappreciated contrarian point is that NVIDIA, AMD, Broadcom, and even Alphabet are not necessarily being disrupted; they may be enabling the very bottlenecks that expand the addressable market for power, packaging, and host CPUs. The competitive pressure is more likely on utilities, legacy power suppliers, and smaller semiconductor names that lack scale in advanced manufacturing or systems-level integration. Over the next 6-18 months, the key catalyst is whether hyperscalers convert capex into physical deployments quickly enough to validate the thesis; if AI spend decelerates or power-grid easing arrives faster than expected, the scarcity premium in BE and INTC can compress sharply.

Key Decisions for Investors

• Maintain a tactical long BE into evidence of backlog conversion and data-center deployment announcements over the next 1-2 quarters; upside remains strong, but trim strength if the stock continues to price in flawless execution.
• Express the AI infrastructure thesis via a pair trade: long BE / short a basket of utility or regulated power names with limited growth optionality, targeting 3-6 months as the scarcity premium should show up first in on-site generation adoption.
• Initiate a medium-dated bullish options structure on INTC (6-12 months) rather than outright equity; the stock’s rerating is likely to be slower and more path-dependent, so convexity is preferable to reduce capital at risk.
• For higher-conviction exposure to the broader AI stack, prefer long NVDA or AVGO on pullbacks while hedging with a partial short in BE only if power-equipment order flow appears to be peaking; this isolates whether the market is overpaying for the bottleneck trade.
• Set a catalyst calendar around hyperscaler capex updates and Intel foundry/packaging disclosures; if those data points fail to confirm deployment acceleration within 1-2 quarters, reduce exposure because the trade thesis weakens materially.