Three stories landed this week that, read together, sketch the emerging geopolitics of energy scarcity: Sam Altman's fusion startup Helion is reportedly negotiating to sell 12.5% of its power output directly to OpenAI; Bloomberg mapped the Strait of Hormuz as the chokepoint defining global energy flows; and EQT CEO Toby Rice told CERAWeek that data centers are now the single biggest variable shaping energy demand. The thread connecting them: AI doesn't just consume culture, it consumes electricity at civilizational scale.
The Vertical Integration of AI Energy Demand
Helion's reported deal with OpenAI is less about fusion's viability than about a tech company deciding to own its energy supply chain end-to-end. When the biggest AI lab in the world negotiates directly with a fusion startup for dedicated power, it's not an energy story — it's a vertical integration play. This is the kind of move that reshapes markets before regulators have names for it. For founders building in the AI infrastructure space, understanding how seed-stage AI investors are thinking about energy dependencies will matter more and more in pitch rooms. The capital is already connecting these dots.
Chokepoints, Data Centers, and the New Resource War
Meanwhile, the Strait of Hormuz isn't just a shipping lane — it's a variable in the AI compute equation. Fast Company noted that Trump's Iran deadline extension affects energy prices globally, and Toby Rice's CERAWeek remarks made explicit what the industry already knows: LNG demand and data center proliferation are now coupled systems. A 2026 arXiv paper on Hyperagents and self-improving AI systems implicitly raises the stakes — if AI systems become more autonomous and more energy-hungry simultaneously, the geopolitical pressure on energy chokepoints intensifies. The Strait of Hormuz is, in a strange and very 2026 way, an AI infrastructure problem.