Nvidia Earnings This Week; Biggest Power Deal in History | Bloomberg Tech 5/18/2026

Key Concepts

• AI Infrastructure & Compute: The physical and financial resources (GPUs, data centers, power) required to sustain the AI boom.
• Agentic AI: AI systems capable of autonomous reasoning and task execution, driving increased demand for compute and CPUs.
• Compute as an Asset Class: The emergence of financial instruments (futures, indices) to hedge against volatility in GPU pricing and availability.
• Supply Chain Bottlenecks: The transition from silicon shortages to physical infrastructure constraints, specifically power, steel, and concrete.
• 13F Filings: Regulatory disclosures revealing hedge fund positioning in major tech stocks.

---

1. Market Overview and Geopolitical Tensions

The NASDAQ 100 and semiconductor indices have faced significant pressure, with the latter losing approximately 6% over two trading days. This downturn is attributed to a combination of "risk-off" sentiment due to US-Iran geopolitical tensions and investor anxiety regarding the sustainability of the AI "melt-up."

2. Nvidia and the AI Hardware Ecosystem

• Earnings Anticipation: Investors are focused on Nvidia’s upcoming earnings, specifically looking for confirmation of top-line growth (which exceeded 80% in the previous print) and the impact of rising High Bandwidth Memory (HBM) costs on gross margins (currently maintained at mid-70%).
• Inference vs. Training: A major shift is occurring from training workloads to inference and reasoning. While Nvidia dominated training, the rise of companies like Anthropic (coding agents) has created competition in the inference space.
• Strategic Partnerships: Nvidia is increasingly working through partners like Dell Technologies to deliver "AI factories" (data centers optimized for AI workloads) directly to enterprises.

3. The Power and Infrastructure Boom

• NextEra/Dominion Deal: The $67 billion acquisition of Dominion Energy by NextEra represents the largest power deal in history, directly triggered by the massive energy requirements of AI data centers.
• Regulatory Outlook: Analysts suggest the deal faces a non-trivial but manageable regulatory path, potentially closing within 12–18 months, as it focuses on specific regional tech hubs like Virginia’s "Data Center Alley."
• The "Steel and Concrete" Constraint: Industry leaders, including Iron (AI cloud provider), emphasize that while the tech industry moves fast, it cannot "code its way" to power. Building a gigawatt-scale AI factory currently requires a 4–6 year lead time due to utility grid constraints and physical construction requirements.

4. Financializing Compute

• CME & Silicon Data Partnership: Silicon Data and the CME Group are launching the world’s first futures market for AI computing power.
• Purpose: This provides a mechanism for data centers, cloud providers, and AI startups to hedge against the 40% daily volatility in on-demand GPU pricing.
• Normalization: Silicon Data has launched GPU indices on the Bloomberg Terminal to normalize heterogeneous assets (e.g., different H100 configurations and locations), allowing for trackable, representative pricing.

5. Corporate Developments

• SpaceX IPO: Elon Musk confirmed he is in Texas working on plans for a potential SpaceX IPO, which could become the largest listing in history. SpaceX is also executing a 5-for-1 stock split, adjusting the fair market value per share to approximately $105.
• Apple’s Privacy-Focused AI: Apple is preparing to debut a revamped Siri at WWDC, emphasizing "differential privacy" and on-device processing. A key feature will be the auto-deletion of chatbot conversations (30-day or 1-year options) to differentiate itself from competitors who retain user memory indefinitely.
• Samsung Labor Tensions: Samsung faces potential labor strikes that threaten to disrupt the supply of memory chips, a critical component for the global AI industry. The South Korean government is simultaneously debating a "citizen dividend" to redistribute AI-driven corporate profits to the public.

6. Hedge Fund Positioning (13F Filings)

Recent filings show a divergence in institutional sentiment toward Big Tech:

• Bullish: Bill Ackman (Pershing Square) initiated a $2 billion stake in Microsoft, citing the resilience of the 365 suite.
• Bearish: Tiger Global and Coatue slashed their Microsoft stakes by over 50%, while Lone Pine exited the position entirely. Similar selling pressure was observed in Amazon, which was also exited by Berkshire Hathaway.

---

Synthesis

The AI industry is currently undergoing a transition from a "silicon-first" phase to a "physical-infrastructure-first" phase. While software innovation remains rapid, the primary bottlenecks have shifted to energy availability, grid capacity, and the physical construction of data centers. Consequently, compute is being treated as a volatile commodity, necessitating the creation of new financial hedging tools. Investors are increasingly cautious, balancing the massive growth potential of AI against geopolitical risks, labor unrest in the supply chain, and the high capital expenditure required to sustain the "AI factory" buildout.