The Velocity Edge
The Close · 21:00

The Bottleneck Is the Asset

The AI Supercycle is no longer constrained by intelligence — it is constrained by the ability to turn intelligence into reliable, financeable and deployable physical capacity.

Francesco De Leo Kaufmann September 16, 2026 5 min read Capital
≈5%U.S. 10-year Treasury yield around today's session
>$105Brent crude even after today's decline
+25 bptoday's Federal Reserve rate increase
72 GPUsNvidia's Vera Rubin NVL72 as an integrated accelerated-computing system

“The bottleneck is the asset.”

The AI Supercycle is no longer constrained primarily by the availability of intelligence. It is increasingly constrained by the ability to turn intelligence into reliable, financeable and deployable physical capacity.

Today provided unusually strong evidence. The Federal Reserve raised rates for the first time in three years, while the U.S. 10-year Treasury yield remained around 5% and Brent crude stayed above $105 even after retreating from recent highs. Yet technology stocks proved comparatively resilient.

At almost exactly the same moment, Nvidia, Google and Emerald AI launched the AI Energy Management Alliance, designed around a radically different idea: AI data centers should become flexible participants in electricity systems rather than passive consumers of power. These are not separate stories. They describe the same structural transition.

01

From Intelligence Scarcity to Physical Scarcity

Part I — The Big Picture

02

The Great Split Is Becoming a Cost-of-Capital Test

Part II — Capital Allocation

03

Tomorrow & Beyond

Part III — the new scarce asset

  1. The AI Supercycle is moving from CAPITAL → ACCELERATED COMPUTING toward a longer chain: CAPITAL → PERMISSION → ENERGY → GRID ACCESS → ACCELERATED COMPUTING → INFERENCE → DEPLOYMENT → PRODUCTIVITY → CASH FLOW. The new scarce asset is assured permission to convert capital and energy into intelligence. signal: The scarce resource is no longer the model — it is permission to produce.

  2. Recent damage to Gulf cloud infrastructure has highlighted the physical vulnerability of digital infrastructure, while insurers examine new approaches to covering increasingly concentrated AI infrastructure risk. Geopolitical risk drives redundancy, higher capital intensity and distributed architectures — and therefore higher value of assured capacity. signal: Resilience itself begins to acquire economic value.

  3. Winning AI locations will not be those offering the cheapest land. They will be those capable of combining Energy × Grid × Capital × Connectivity × Resilience × Permission × Talent — with enormous implications for the United States, Europe, the Gulf and China. signal: Advantage accrues to whoever can guarantee intelligence is actually produced.

04

The Velocity Equation

joint optimization

Key takeaways

The Velocity Insight

  • The extraordinary abundance of intelligence does not eliminate scarcity — it relocates it.
  • Scarcity migrated from models to GPUs, then to electricity, then to grid connections, and now to assured permission to convert capital and energy into intelligence.
  • Flexibility of electricity demand is a new strategic variable — potentially the most consequential number of today's session.
  • The most valuable assets of the Next Economy may not be the companies with the most powerful models, but those able to guarantee intelligence can be produced continuously, economically and at scale.

Francesco De Leo Kaufmann · The Velocity Edge

More from The Velocity Edge

Jump straight into another note — no need to head back.

Keep reading

More field notes, from dawn to close.

Day signals and night briefings, published like a journal for board-level AI decisions.

Back to The Velocity Edge