The price of electricity just became the price of admission to the future, and bitcoin mining lost the bid.

The Summary

The Signal

Paraguay sits on Itaipu Dam, the second-largest hydroelectric plant on Earth. The country generates more power than it can use, so for years it sold the surplus cheap. Bitcoin miners saw arbitrage. Pack containers with ASICs, plug into the grid at $44 per megawatt-hour, and mine coin while everyone else pays double or triple. It worked. Paraguay became home to 4% of the global bitcoin hash rate, a material slice of the network's security.

Now the state utility, ANDE, is calling time. The contracts expire at the end of 2027, and officials say the deadline is final. The reason is not ideology or carbon politics. It is economics and optionality. AI training farms need the same thing bitcoin miners need: cheap, abundant, 24/7 power. But they stay longer, sign bigger contracts, and align better with what governments want to show the world.

"Founders Fund did not back a bitcoin mine in Paraguay. They backed an AI campus with a 15-year power deal."

The Founders Fund-backed AI project is not a one-off. It represents a pattern playing out across the energy-rich developing world. Countries that used to compete for mining operations are now competing for compute. Malaysia, Kazakhstan, and parts of Africa are running the same playbook: attract AI infrastructure with tax breaks and cheap energy, position yourself as a sovereign alternative to AWS and Google Cloud regions. Bitcoin mining was a proof of concept. AI inference and training is the main event.

For bitcoin miners, the economics are brutal:

  • Mining margins are thin and volatile, tied to bitcoin's price and network difficulty
  • AI datacenters offer predictable revenue, longer contracts, and enterprise-grade SLAs
  • Governments can pitch AI as jobs and innovation, while mining gets framed as energy waste

The $350 million Paraguay earns annually from miners is not trivial, but it is also not sticky. The power purchase agreements were always short-term, often renegotiated, and carried reputational risk every time bitcoin's price crashed or energy activists showed up. A 15-year deal with a Founders Fund-backed AI campus is bankable debt. It is infrastructure governments can point to when courting foreign investment. It is optically safe.

The Implication

If you run a bitcoin mining operation in Paraguay, you have 13 months to find new power or new ground. The broader signal is starker: the era of bitcoin mining as the highest-value use of stranded energy is over. AI ate the arbitrage. The same cheap electricity that made mining viable in second-tier grids now makes those grids prime real estate for compute.

Watch for miners to move again. They will chase the next tier of undeveloped hydroelectric, geothermal, and natural gas flaring in places that have not yet signed 15-year AI deals. But the window is closing. Every government with surplus energy is now getting pitched by AI infrastructure funds. The miners who survive will be the ones who can also run inference workloads or partner with AI companies to share the grid. Pure-play mining is losing the only moat it ever had: access to the cheapest electricity on the planet.

Sources

CoinDesk