Investors should monitor OpenAI and SpaceX closely as both companies are aggressively scaling enterprise revenue and infrastructure deals to prepare for potential IPO filings. NVIDIA (NVDA) remains a high-conviction play as it locks down global supply chains and faces "enormous demand" for its next-generation chips. Consider SK Hynix (000660.KS) as a critical "pick and shovel" investment, as it has secured a multi-year deal to be the primary memory provider for NVIDIA’s AI hardware. Google (GOOGL) is strategically hedging its AI future by renting massive compute capacity from SpaceX, a move that boosts the valuation of its own 6% stake in the aerospace giant. To capitalize on the next phase of AI, shift focus from simple chatbots to "agentic" workflows and "loops," which will drive massive usage-based revenue for cloud and infrastructure providers.

OpenAI (Private)

• The company is planning a major overhaul of ChatGPT, shifting it from a simple chatbot to a "super app" that integrates coding tools, AI agents, and external partner apps. • OpenAI is reportedly pitching the U.S. government on a plan to donate equity to a public sovereign wealth fund, potentially allowing citizens to receive "AI dividends." • The company is shifting resources toward enterprise customers, who now account for 40% of revenue, with a target of 50% by year-end. • Codex has emerged as their most successful product, driving a "vibe shift" toward agentic workflows rather than simple chat.

Takeaways

• Monitor the IPO Path: The shift toward a "super app" and enterprise revenue is a clear signal that OpenAI is grooming its financials and product structure for an IPO (potentially this year). • The "Advantage Gap": Investors and professionals should focus on "agentic" use cases (using AI to perform tasks) rather than "chat" use cases (summarization) to maintain a competitive edge. • Usage-Based Revenue: Expect a shift from flat monthly subscriptions to usage-based pricing as "loops" and agents consume significantly more tokens than standard queries.

SpaceX (Private) / XAI

• Google has signed a three-year, $920 million per month deal with SpaceX to rent compute capacity (at least 110,000 NVIDIA GPUs). • This deal mirrors a previous landmark deal with Anthropic, positioning SpaceX/XAI as a major "neocloud" provider. • Analysts suggest XAI could generate $26 billion per year in revenue from licensing compute, potentially paying back its $40 billion data center investment in just 18 months.

Takeaways

• Infrastructure as a Business: SpaceX is no longer just a space company; it is becoming one of the largest GPU cloud providers on Earth, rivaling specialized firms like CoreWeave. • Google's Strategic Hedge: Google owns a 6% stake in SpaceX. This compute deal helps ensure they have "bridge capacity" for their Gemini platform while simultaneously boosting the valuation of their own investment. • IPO Watch: With a SpaceX IPO rumored to be imminent, these massive revenue-generating compute deals are likely intended to "puff up" the company's valuation.

NVIDIA (NVDA)

• CEO Jensen Huang secured a multi-year deal with SK Hynix to ensure a steady supply of High-Bandwidth Memory (HBM). • NVIDIA will also act as a design partner for new memory chips tailored for physical AI and next-generation Vera Rubin chips. • Huang noted that the entire industry supply chain—from wafers to cable connectors—remains in a state of "supply shortage" due to enormous demand.

Takeaways

• Supply Chain Dominance: NVIDIA’s strategy involves aggressive, face-to-face "diplomacy" to lock down limited global manufacturing capacity, creating a moat against competitors. • Bullish Demand Signal: The CEO’s comments on "enormous demand" and persistent shortages suggest that the AI infrastructure build-out is not yet slowing down.

SK Hynix (000660.KS)

• The company has solidified its position as NVIDIA’s primary memory supplier through a new multi-year agreement. • The partnership extends beyond supply into co-designing chips for AI infrastructure.

Takeaways

• The "Pick and Shovel" Play: As the primary provider of the specialized memory required for high-end GPUs, SK Hynix is a critical beneficiary of the NVIDIA growth story.

Investment Themes & Sectors

Sovereign Wealth & Nationalization

• There is growing bipartisan discussion (Trump and Sanders) regarding the U.S. government taking equity stakes in major AI labs. • Insight: This suggests the government views AI labs as "too big to fail" and critical to national security, similar to the banking or defense sectors.

The Shift to "Loops" and Agents

• The industry is moving away from manual prompting toward "loops"—automated systems where AI agents work continuously to solve complex problems without human intervention. • Insight: Companies that provide the infrastructure for these high-token-consumption "loops" (like Google Cloud, AWS, and SpaceX) stand to benefit from a massive increase in compute demand compared to standard chatbot usage.

AI Implementation Gap

• Research from KPMG and Section shows a massive gap between companies that have AI tools and those that derive actual business value (only 12%). • Insight: There is a significant opportunity for consulting and platform firms (like OutSystems or Section) that help enterprises move beyond simple "meeting summaries" to high-value agentic workflows.