Regime Shift
The fundamental nature of electricity has changed. Electrons cannot move globally at scale — but compute can. That is the arbitrage.
Historical constraint
Historically, electricity was priced and consumed locally — a regional utility with no global market.
The new mechanism
AI and Bitcoin mining have created a new mechanism: convert cheap local power into globally-priced digital output.
Structural arbitrage
The arbitrage is structural: wherever power is cheapest, compute economics are strongest.
Irreversible shift
Once compute infrastructure is deployed at scale, power access becomes a strategic moat.
Electricity is no longer a local utility. It is becoming a globally arbitraged compute commodity.
Power Supercycle
Electricity demand is entering a structural acceleration phase driven by AI, electrification, and industrial growth — creating a sustained supply-demand imbalance.
3.6%
CAGR (2026–2030)
+1,100
TWh per year
33,600
TWh by 2030
This growth rate is nearly double the historical average. 2026 starts the global electricity supercycle.
AI Reprices the Grid
US and EU grids face multi-year interconnection queues. New data center capacity is effectively rationed. Power access — not capital — is the binding constraint.
17%
Growth in data-center electricity use in 2025
2x
Expected global data-center power use by 2030
3x
Expected AI-focused data-center power use by 2030
Data center electricity demand is growing exponentially, making low-cost power access a scarce, strategic resource.
The Cost Curve Will Flip
As GPU depreciation costs decline and specialized silicon matures, power cost will become the dominant variable in AI economics — rewarding operators with the cheapest electricity.
Today
- GPU depreciation > power cost
- Maximize uptime at any power price
- Location follows talent and connectivity
- Hardware edge is the moat
Tomorrow
- Power cost > GPU depreciation
- Optimize power economics, accept some flexibility
- Location follows cheapest electrons
- Power access is the moat
ASICization Accelerates the Power Shift
Bitcoin mining was transformed by ASICs — single-purpose SHA-256 chips that were extremely efficient but commoditized hardware advantage entirely. Once ASICs matured, the only differentiator was electricity cost.
AI inference may follow the same curve: as text inference moves toward specialized silicon, compute prices fall faster — bringing forward the inflection point where electricity cost overtakes depreciation.
The operator who secures the cheapest long-duration power today will hold the structural cost advantage in the AI era.
Mining as Power Capture
Bitcoin mining is the optimal first-mover strategy to monetize stranded power, establish utility relationships, and create the infrastructure optionality for AI/data center conversion.
Converts under-monetized electricity into immediate USD-denominated cash flow.
Creates stable, predictable load — making Leviathan a valued utility partner.
Establishes government coordination and regulatory alignment.
Builds grid bargaining power for future capacity expansion.
Mining cash flow funds the transition to higher-value compute workloads.
Mining is not the end state. It is the cheapest way to capture power resources and win grid leverage.
Why Leviathan / Why Africa
We control the conversion layer between under-monetized power and globally priced compute.
Africa context
- <1% — Africa accounts for less than 1% of global data-center capacity.
- ~40% — Mobile data usage is growing at approximately 40% annually.
- 2GW — Uganda's Karuma Hydropower Project lifted national capacity above 2,000MW.
Leviathan's wedge
- 300MW — Uganda anchor site secured for immediate deployment.
- $0.035/kWh — Long-duration power access (vs. $0.07–0.12/kWh in US).
- Government and utility alignment secured.
- Tariff pathway established.
- Mining-first cash flow from Day 1.
In the last decade, compute followed hardware. In the next decade, compute will follow power.