Investors should prioritize Anthropic as it nears its first profitable quarter with a massive $47 billion revenue run rate, signaling a shift from speculative growth to fundamental value. Consider SpaceX as a primary infrastructure play ahead of its anticipated IPO, as its Colossus supercomputer clusters position the company as a critical "Czar of Compute" for the AI industry. Accumulate high-conviction memory leaders SK Hynix and Micron, which remain essential "picks and shovels" providers during the current structural token shortage. Shift focus from raw model providers to "harness" platforms like Replit and Claude Code, which capture value by enabling autonomous agentic workflows rather than simple chat interfaces. Monitor the transition to usage-based pricing and hedge against "AI sticker shock" by investing in cost-management tools as the era of flat-rate subsidies ends.

Anthropic

• Massive Revenue Growth: Anthropic’s annualized revenue run rate (ARR) surged from $3 billion at the start of 2025 to $47 billion by May 2026.
• Profitability Milestone: The company is anticipating its first profitable quarter, a significant psychological shift for the foundation model sector.
• Valuation: Recently closed a $65 million fundraising round, valuing the company at just under $1 trillion.
• Strategic Partnerships:
  • Partnered with Blackstone, Hellman & Friedman, and Goldman Sachs to launch an enterprise AI consulting firm.
  • Formed a major infrastructure alliance with SpaceX/xAI to use the "Colossus" supercomputer clusters to solve compute constraints.
• Model Updates: Released Claude Opus 4.8; however, the market focus has shifted from raw model power to the "harnesses" (like Claude Code) they sit in.

Takeaways

• Enterprise Shift: Anthropic is currently outpacing OpenAI in business adoption according to data from Ramp.
• Consulting Revenue: The move into consulting suggests that "capability overhang" (the gap between what AI can do and what companies know how to do) is a major new revenue stream.
• Token Pricing: Anthropic is shifting toward per-token billing for third-party tools, ending the "subsidy era" for power users.

OpenAI

• Revenue Performance: Reached a $30 billion ARR, driven by the shift from seat-based subscriptions to API token usage.
• Enterprise Strategy: Launched a "deployment company," a separate venture to place engineers directly inside large client organizations to facilitate AI integration.
• Model Cycle: Following the release of GPT-5.2, the market is anticipating a new model release in the near future.

Takeaways

• Business Model Evolution: OpenAI is moving away from relying on $20/month consumer seats toward high-volume API revenue.
• Investment Sentiment: The massive revenue numbers have largely silenced "AI Bubble" concerns from late 2025, as the labs are now proving they can realize value from their infrastructure spend.

SpaceX / xAI

• The "NeoCloud" Pivot: SpaceX is transitioning into a massive AI infrastructure provider. Its "SpaceX AI" division is leasing its Colossus 1 and Colossus 2 supercomputers to Anthropic.
• Upcoming IPO: The SpaceX IPO is increasingly viewed through an AI lens. Investors are eyeing it as a play on the "AI supply chain" and physical infrastructure rather than just aerospace.
• Future Tech: Elon Musk is signaling a move toward Orbital Data Centers within the next 2–3 years to solve terrestrial power and cooling constraints.

Takeaways

• Infrastructure Play: SpaceX is positioned as a "Czar of Compute," providing the essential hardware (tokens) that foundation labs currently lack.
• Valuation Driver: The ability to sell compute at a premium de-risks SpaceX’s massive capital expenditure and makes the IPO highly attractive to AI-focused institutional investors.

AI Infrastructure & Memory Stocks

• SK Hynix & Micron: Both companies have surged to become trillion-dollar companies due to the insatiable demand for AI memory.
• Base 10: The inference provider is raising $1 billion at an $11 billion valuation, doubling its value in just one quarter.
• OpenRouter: Recently raised $113 million (Series B) to become an AI unicorn; the platform allows developers to toggle between models to optimize for cost and performance.
• Meta (META): Sentiment has shifted from "fear of overspending" to "optimism." Investors now see Meta’s $130 billion compute investment as a potential cloud business where they can sell excess capacity.

Takeaways

• The "Token Shortage" Era: We have entered a period of structural shortage where there is not enough compute to meet demand. This keeps the "picks and shovels" providers (chips, memory, data centers) in a high-growth phase.
• Vertical Integration: Infrastructure companies are moving up the stack, and model companies are moving down into hardware/consulting.

Emerging Investment Themes & Sectors

The End of the "AI Subsidy"

• Context: Previously, companies like GitHub (Copilot) and Google (Gemini) offered flat-rate monthly seats that allowed unlimited usage.
• The Shift: Usage-based billing is becoming the standard. Power users who were getting $5,000 of value for a $200 seat are now being transitioned to per-token pricing.
• Insight: Look for companies that provide AI Cost Management tools, as "AI sticker shock" is hitting corporate budgets.

Agentic AI & "Harnesses"

• Context: The "Agent Era" began in early 2026. The focus is no longer on chatbots but on systems that do work (coding, autonomous workflows).
• Key Players: Replit, Lovable, Zencoder (Zenflow Work), and OutSystems are highlighted as platforms enabling these agentic systems.
• Insight: Investment value is migrating from the "Model" (the brain) to the "Harness" (the interface/tools the brain uses to execute tasks).

Competitive Pricing Wars (China)

• Context: Chinese firm DeepSeek made a permanent 75% price cut on its V4 model.
• Insight: As Western tokens become expensive and scarce, expect a "race to the bottom" in pricing from Chinese providers looking to capture global market share from OpenAI and Anthropic.