Key Takeaways

  • The United States runs on roughly 500 gigawatts of total power consumption across the entire country.
  • SpaceX plans to bring online 10 gigawatts of dedicated power capacity, which Gavin Baker notes equals 25% of all incremental power added to the US grid.
  • Elon Musk proposes a direct economic ratio: a 1% expansion in total national power usage yields roughly a 1% increase in US gross domestic product.
  • Jensen Huang's hardware improvements and algorithmic gains compound this effect by driving up the useful intelligence delivered per watt of electricity.

The 500 Gigawatt Ceiling

Tech founders often treat computing capacity as an abstract cloud expense. At the American.gov Summit, Elon Musk anchored the artificial intelligence race to hard physics. “Total power consumption in the United States is about 500 gawatt,” Musk said. “So 500 g remember that number that's how much power we got in America.”

Against that baseline, the power requirements of frontier artificial intelligence systems stop looking like ordinary commercial real estate projects. They start looking like national infrastructure programs. Gavin Baker pointed directly to the sheer scale of private data center plans: “Elon, if you bring on 10 gigawatts, which SpaceX has has discussed publicly, that'll be about 25% of all the power added in America.”

When a single company plans to build out a quarter of a country's entire incremental electrical grid additions, computing constraints cease to be software problems. They become power generation and transmission problems. If you cannot secure substations, turbines, and grid interconnects, your software roadmap stalls regardless of how much venture capital sits in your bank account.

Why Watts Predict Gross Domestic Product

Musk tied power generation directly to macroeconomic output. “I would bet anyone that 1% increase in power usage corresponds to roughly 1% increase in GDP,” Musk argued, “and so 10 10 gawatt would be a 2% increase in in GDP.”

Historically, industrial economies expanded in lockstep with raw primary energy consumption. As Jordi Hays observed, “The more energy a country is producing, the higher their GDP.” Musk's thesis asserts that artificial intelligence revives this historical link rather than breaking it. If machine intelligence handles real-world labor, economic velocity scales with the electricity feeding the clusters.

There is a second variable in the equation: computational efficiency. John Coogan pointed out that raw power input multiplies against hardware and software progress. “Elon makes the point that useful intelligence per watt is increasing through more more efficient chips from Jensen and also better algorithms effectively.” If total energy increases while intelligence per watt multiplies, economic output accelerates faster than standard macroeconomic models project.

What to Do With This

Audit your product's unit economics against compute and power bottlenecks this week. If your startup relies on frontier model API calls or dedicated GPU clusters, calculate your cost per task under a scenario where data center power prices double over the next 24 months, then rewrite your inference pipelines to prioritize intelligence per watt.