The Great Split Is Accelerating.

“Competition is no longer between companies. It is increasingly between entire economic systems.”

What if the biggest investment mistake of the next decade is assuming globalization is merely slowing down — when in reality capital is already reallocating toward competing AI ecosystems? For years, markets viewed geopolitics as background noise and technology as the primary growth story. Today those two narratives have become one.

Hormuz is no longer simply an energy story. TSMC is no longer simply a semiconductor company. NVIDIA is no longer just selling chips. CATL is no longer just building batteries. Anthropic and OpenAI are no longer simply developing models. Each headline points the same way: The Great Split is accelerating. The next cycle will not be won by those building the best applications — it will be won by those controlling the infrastructure that makes intelligence possible.

1. 01 Capital is beginning to follow The Great Split

   What we described as a geopolitical transformation is becoming a capital-allocation framework. Baidu is restructuring its listing to strengthen access to Asian capital. TSMC is committing $265 billion to expand US manufacturing. Japan is investing in sovereign AI infrastructure powered by NVIDIA’s next-generation Rubin platform. These reflect two increasingly independent ecosystems competing across AI, semiconductors, manufacturing and capital formation.

   Signal Globalization is not disappearing — it is fragmenting, and capital is fragmenting with it.

2. 02 Hormuz has become part of the AI supply chain

   Markets still interpret Hormuz through the price of oil. That framework is obsolete — the real issue is electricity. AI factories consume unprecedented power; power increasingly depends on reliable LNG; semiconductor manufacturing depends on uninterrupted electricity. The chain is unmistakable: geopolitics → energy → compute → AI capacity → productivity → earnings.

   Signal Energy security has become intelligence security — and the market hasn’t fully priced it.

3. 03 AI has entered its industrial phase

   TSMC, NVIDIA, CATL, robotics, battery storage, advanced manufacturing — no longer adjacent themes, but the foundations of the next industrial cycle. The first wave of AI created software winners; the second is creating infrastructure winners that control accelerated computing, semiconductor manufacturing, electricity generation, battery storage, robotics, AI factories, industrial automation and intelligent mobility.

   Signal Markets still reward AI; soon they will reward everything AI depends upon.

4. 04 The productivity revolution is becoming real

   Coding is becoming the first major profession fundamentally transformed by agentic AI — but the implications extend far beyond software. Law, finance, consulting, healthcare and engineering are approaching the same inflection point. Future advantage will no longer be expertise alone; it will be the ability to combine human judgment with AI at organizational scale.

   Signal Organizational learning velocity is becoming a new source of enterprise value.

5. 05 Europe’s window has not closed

   The narrative says Europe has already lost the AI race. That may prove premature. The next cycle is increasingly about industrial capability — electricity, power grids, battery storage, rail, advanced mobility, engineering, industrial robotics — areas where Europe holds world-class capabilities. The challenge is no longer innovation. It is capital allocation.

   Signal Europe need not win yesterday’s software race — it needs to lead tomorrow’s industrial AI economy.

Markets still analyse headlines separately — energy, semiconductors, AI, defense, mobility, payments. That is yesterday’s framework. Tomorrow’s economy will be built around ecosystems, and the winning ones will integrate compute, energy, semiconductors, sovereign capital, robotics, industrial manufacturing, mobility and AI deployment.

The winning ecosystems integrate

Competition between economic systems

• Compute & accelerated computing.
• Energy generation and grids.
• Semiconductors and manufacturing.
• Sovereign capital.
• Robotics and industrial automation.
• Mobility and AI deployment.

Key numbers to watch

The signals behind the split

• $265B — TSMC’s planned US investment, reinforcing semiconductors as strategic national infrastructure.
• ~20% — share of global LNG trade passing through Hormuz, linking geopolitics and AI infrastructure.
• Rubin — NVIDIA’s next-generation platform is becoming the foundation of sovereign AI strategies.
• Sodium-ion batteries — CATL’s expansion into Europe could reshape the economics of large-scale storage.
• Agentic AI — the first measurable productivity gains across knowledge work are emerging.

The next decade will not simply reward exposure to AI — it will reward exposure to the infrastructure that enables AI. The debate is shifting from software to strategic capability, from digital platforms to industrial ecosystems, from globalization to economic sovereignty. Watch closely: accelerated computing, semiconductor manufacturing, energy infrastructure, battery storage, robotics, sovereign AI, AI factories, intelligent mobility, industrial software and physical AI.

For thirty years investors asked which company would dominate. The defining question of the next decade is different: which ecosystem will dominate the global economy?