3 years ago we only had 2 humanoid robots…Never thought I’d say this, but we’re now building so many robots that we need storage space for them https://t.co/NXIikLOmCV
The cheapest places for electricity production depend on the energy source, local resource quality (e.g., solar irradiance, wind speeds, hydro head, geothermal heat), construction/operational costs, existing infrastructure, labor, regulations, and financing. Levelized Cost of Electricity (LCOE) is the key metric here—it accounts for lifetime costs divided by energy produced. Renewables like solar PV and onshore wind now often have the lowest LCOE globally in good locations, frequently under $30–50/MWh (and sometimes much lower in auctions or optimal sites). Fossil fuels (especially subsidized coal/gas) and existing hydro/geothermal can also be very cheap. Untapped potential is highest in sunny/windy regions of Africa, the Middle East, South Asia, and parts of Latin America, where resources are abundant but development is limited by infrastructure, financing, or policy. Top Places by Energy Source (Cheapest Current or Potential Production)
Solar PV: Middle East (Saudi Arabia, UAE, Oman, Kuwait), North Africa (Egypt, Morocco), and parts of Australia/China/India
Why cheapest? Exceptional solar irradiance (high capacity factors >25–30%), low land costs in deserts, cheap labor/construction in some areas, and massive scale/manufacturing (especially China). Record auction prices have hit ~$10–20/MWh. Saudi Arabia has achieved some of the world's lowest LCOEs. China benefits from the world's lowest build costs due to domestic manufacturing.
Untapped potential: Huge in the Sahara, Arabian Desert, and sunbelt regions of Africa/Asia. Many areas have minimal current deployment relative to resources.
Onshore Wind: China (northern/northwestern regions), parts of the US Great Plains, Patagonia (Argentina/Chile), and high-wind areas in Africa/Europe
Why? Strong, consistent winds yield high capacity factors. China has very low costs and leads global deployment. Onshore wind often has LCOE competitive with or below solar in prime spots.
Untapped: Vast potential in offshore (higher cost) and inland high-wind zones in developing regions.
Hydropower: Existing large-scale in China, Brazil, Canada, Norway, Iceland; untapped in Africa (Congo River basin), Himalayas (Nepal, India, Pakistan), and parts of South America
Why? Very low operating costs once built; some of the cheapest electricity overall (especially existing plants). Countries like Iceland, Norway, and parts of Canada/Brazil have abundant hydro leading to low prices.
Untapped potential: Enormous in sub-Saharan Africa and Asia—economically viable potential is about 50% untapped globally, higher in developing regions. New plants can be cost-effective where sites exist.
Geothermal: Iceland, parts of Indonesia, Philippines, New Zealand, Kenya (East African Rift), and western US
Untapped: Significant in volcanic/rift zones; enhanced geothermal systems (EGS) could expand this dramatically, though upfront drilling costs are high.
Fossil Fuels (with caveats): Subsidized natural gas/coal in Iran, parts of Russia, and some Middle East/Gulf producers
Why? Extremely low consumer prices due to subsidies and abundant local resources (e.g., Iran at ~$0.003/kWh residential). These are not always the lowest production costs in a market sense but enable very cheap electricity. Not sustainable long-term due to emissions and volatility.
Overall standout countries/regions for low-cost production today:
China: Lowest generation costs studied for renewables; massive scale, manufacturing advantages, and rapid deployment drive costs down.
Middle East/Gulf (Saudi Arabia, UAE): Record-low solar costs + gas backups.
Iran and subsidized producers: Artificially low prices via fossils.
Key Explanations and Considerations
Renewables dominance in new builds: Solar and wind have seen dramatic cost drops (solar ~90% since 2010). In optimal locations, they undercut new coal/gas. Hybrids with storage or hydro for firming improve reliability.
Geography matters most: Resource quality (e.g., sun/wind/hydro head) drives ~50%+ of economics. Deserts, mountains, coasts, and rift zones win.
Untapped potential: Highest in Africa (solar, hydro, wind, geothermal), Middle East (solar), South Asia (solar/wind/hydro), and Latin America. Barriers include grid infrastructure, financing, and policy.
Other factors: Transmission costs, storage for intermittency, regulations, and subsidies affect delivered prices. Existing plants (hydro, old coal/nuclear) often have the absolute lowest marginal costs.
Trends: Costs continue falling for renewables. China leads deployment and cost reduction. Emerging markets could see the biggest gains from untapped resources.
For specific projects, always check recent auctions and local LCOE studies, as values vary widely even within countries. Data draws from IEA, IRENA, BNEF, Lazard, and national reports (as of recent years). Actual "cheapest" depends on whether you're measuring new-build LCOE, marginal cost, or retail prices.
Powering Intelligence: Why the AI Era Demands Energy Abundance The artificial intelligence revolution is redefining human progress, but it comes with an insatiable appetite for electricity. Data centers powering AI workloads consumed roughly 415 TWh globally in 2024, about 1.5% of world electricity use. Projections show this doubling to around 945 TWh by 2030, with AI-optimized servers driving much of the surge at 30% annual growth. In the United States alone, data centers could claim 7-12% of electricity by the late 2020s or early 2030s, up from ~4% recently. This is not a temporary spike. AI demand—spanning training, inference, and expanding applications—appears bottomless. As models grow more capable and ubiquitous, compute needs will explode. The logical response is not rationing or hesitation, but aggressive expansion of electricity generation, especially in the world's cheapest locations. Building abundant, low-cost power unlocks economic growth, technological leadership, and human flourishing.The Economics of Cheap Power in Prime LocationsElectricity production costs have plummeted for renewables. Onshore wind and solar PV now offer some of the lowest levelized costs of electricity (LCOE) worldwide, often under $30–50/MWh in optimal sites—frequently beating new fossil fuel plants. Key hotspots for ultra-cheap generation include:
Solar PV in the Middle East and North Africa (Saudi Arabia, UAE, Egypt, Morocco): World-class irradiance yields high capacity factors. Record projects have achieved prices as low as $10–20/MWh. Vast deserts provide low-cost land.
Onshore Wind in China, Patagonia, US Great Plains, and African high-wind zones: Strong, consistent resources plus manufacturing scale (especially in China) drive costs down dramatically.
Hydropower in China, Brazil, Canada, Norway, and untapped basins like the Congo River: Extremely low marginal costs once built. Existing fleets deliver some of the cheapest power on earth.
Geothermal in Iceland, East African Rift, Indonesia, and the western US: Near-constant output with minimal fuel costs.
Hybrid and existing systems: Gas in resource-rich subsidized markets (e.g., Gulf states, Iran) provides firming, while China's integrated renewables manufacturing ecosystem achieves the world's lowest build costs.
These locations share advantages: exceptional resource quality, available land, and often lower labor or regulatory hurdles for large-scale deployment. Untapped potential remains enormous across Africa, the Middle East, South Asia, and Latin America—regions where solar, wind, and hydro resources far exceed current infrastructure. Why Overbuild for AI? The Case for AbundanceAI compute is not a zero-sum game with other sectors; it is a multiplier. Cheap, abundant electricity enables:
Explosive productivity gains: AI accelerates scientific discovery, drug development, materials science, energy optimization, and automation. Every marginal kWh invested in compute can yield outsized economic returns.
Lower system costs: In high-resource areas, adding solar and wind is now the cheapest way to expand supply. Overbuilding generation (paired with storage, transmission, or flexible loads like data centers) smooths intermittency and reduces curtailment. Data centers can locate near cheap power sources or use PPAs to finance new builds, as hyperscalers already do.
Economic development: Countries and regions with cheap power become magnets for AI infrastructure investment, jobs, and related industries. This creates a virtuous cycle: more power attracts more compute, which funds more generation.
Energy security and resilience: Diverse, decentralized renewables plus firm sources (hydro, geothermal, gas, nuclear) reduce dependence on volatile fuels. AI itself can optimize grids, predict demand, and improve efficiency.
Global scalability: Demand is borderless. Tech companies can site facilities wherever power is cheapest and most reliable—exporting intelligence while importing electrons indirectly. This democratizes access to AI benefits.
Critics worry about grid strain, water use for cooling, or emissions. These are real challenges but solvable through innovation: advanced cooling, water recycling, co-location with renewables, and hybrid systems. AI-driven efficiency improvements in chips, algorithms, and cooling are already mitigating per-query energy use. The alternative—constraining supply—risks higher prices, slower innovation, and ceding leadership to nations that embrace abundance. Historical precedent supports boldness. Past energy expansions (electrification, oil, gas) fueled unprecedented growth. Today, solar and wind deployment rates demonstrate that rapid scaling is feasible when economics align. With AI as a catalyst, investment in generation becomes self-reinforcing.A Strategic ImperativePolicymakers and investors should prioritize:
Streamlining permitting and transmission in high-resource regions.
Encouraging corporate offtake agreements and hybrid projects.
Investing in storage, grid modernization, and firm low-carbon sources.
Fostering international partnerships to develop untapped potential in emerging markets.
The AI era does not require energy trade-offs—it rewards abundance. By generating as much cheap electricity as possible in the sunniest, windiest, and most hydro-rich places on Earth, we remove the primary bottleneck to progress. The compute will come. The question is whether the power will be ready to fuel it.Abundant energy has always been the foundation of advancement. In the age of intelligence, it is the ultimate enabler. Let's build it.
Nepal’s Hydro Powerhouse: Why Harnessing 50,000 MW Is a Compelling Bet for the AI Era Nepal sits on one of the world’s greatest untapped energy resources. With rivers cascading from the Himalayas, the country boasts a theoretical hydropower potential of around 83,000 MW, of which roughly 42,000–43,000 MW is considered economically viable. Yet as of mid-2025, installed capacity hovers around 3,400–4,000 MW—less than 10% of the viable total. Harnessing 50,000 MW—ambitious but within striking distance of the upper estimates—is not just feasible; it is a strategic imperative, especially for fueling the exploding demand of AI compute. This scale of development can be accelerated through capital-intensive execution and openness to global vendors and partnerships.A Perfect Match for AI’s Insatiable AppetiteAI data centers and training clusters demand reliable, low-cost, high-capacity-factor power. Hydropower excels here: once built, it offers some of the cheapest electricity available, with levelized costs of energy (LCOE) often ranging from $0.02–0.10/kWh for well-sited projects—highly competitive with or superior to new solar/wind in many contexts, and far more dispatchable. Nepal’s run-of-river and storage projects can deliver firm power with capacity factors that complement intermittent renewables. In an era where global data center electricity demand could surge toward 1,000 TWh annually by 2030, nations and companies that secure abundant, cheap baseload-like power gain a decisive edge. Nepal could become a green compute hub or a major exporter of electrons (via transmission links to India and beyond), turning its natural endowment into economic multiplier effects: revenue, jobs, infrastructure, and technological leapfrogging.Speed Is Possible—with the Right ApproachTraditional hydropower timelines are long, but acceleration is achievable through aggressive capital deployment. Large storage and cascade projects require substantial upfront investment, but this is precisely where capital intensity pays off: modern engineering, prefabrication, and experienced contractors can compress schedules. Nepal has already shown momentum. Projects under construction or in pipeline exceed 10,000 MW, with thousands more in planning. Recent additions like Upper Tamakoshi (456 MW) demonstrate technical capability, while international collaborations—such as Upper Trishuli-1 (216 MW) involving Korean and IFC partners—highlight successful models. To reach 50 GW at speed:
Prioritize bankable mega-projects: Sites like Karnali Chisapani (10,800 MW potential), Pancheshwar (multipurpose, ~5 GW), and others in the Karnali, Gandaki, and Koshi basins offer massive scale.
Blend run-of-river with storage: Storage projects mitigate seasonal variability (monsoon surplus vs. dry-season shortfall) and provide grid stability ideal for AI loads.
Capital-intensive execution: Deploy tunnel boring machines (TBMs), advanced electro-mechanical equipment, and modular construction to cut timelines. Financing via multilateral banks, sovereign funds, green bonds, and corporate PPAs from hyperscalers can de-risk and accelerate.
Global Vendors and Partnerships Are EssentialNepal cannot achieve this scale in isolation. Local expertise has grown impressively in smaller projects, but mega-developments benefit enormously from international know-how in financing, engineering, environmental management, and supply chains. Openness to global vendors means:
Engaging Chinese, Indian, Korean, European, and North American firms for turbines, generators, tunneling, and EPC (engineering, procurement, construction) contracts.
Public-private partnerships (PPPs) and joint ventures that bring capital, technology transfer, and risk-sharing.
Export-oriented models: Power purchase agreements with India (already importing Nepali power) and potentially Bangladesh, combined with cross-border transmission, create revenue streams that repay investments rapidly.
Successful precedents exist. Indian companies like SJVN are developing projects such as Arun-3 and Lower Arun. Broader international involvement can replicate and scale these wins while addressing environmental and social safeguards through best-practice standards. Overcoming Challenges Head-OnGeology, sedimentation, terrain, seismic risks, and community impacts are real. Solutions include rigorous feasibility studies, modern sediment management, benefit-sharing mechanisms (equity for locals, irrigation synergies, rural electrification), and diversified renewables. Transmission infrastructure must expand in parallel—another area ripe for global investment. Political stability and investor-friendly policies have improved; further streamlining of approvals, transparent PPAs, and risk-mitigation instruments will unlock capital.The Payoff: Energy Abundance and National TransformationDeveloping 50,000 MW would generate tens of billions in annual economic value through domestic supply, exports, and induced growth. It positions Nepal as a renewable energy leader in South Asia, powers its own digital economy (including potential AI/data center clusters), and contributes to global decarbonization. In the AI age, cheap, abundant electricity is the new oil. Nepal’s hydro resources represent a rare opportunity to convert geography into enduring prosperity. With capital intensity, speed-focused execution, and global collaboration, 50 GW is not a distant dream—it is an achievable, high-return bet. The rivers are flowing. The demand is surging. The time to harness this potential at scale is now.
