Electric Vehicles and the US Electric Grid
Tesla is coming out with a cyber car: a robotaxi (a self-driving car thing). And also with a new model. And also with a battery-operated sixteen wheeler. There are already about 7 million EVs on US roads. I was wondering if our electric grid could charge double that number. Here is what Grok had to say about it:
Quick numbers on total energy demand
• US electricity generation in 2025 reached a record ~4,430 TWh (terawatt-hours).
• A typical Tesla (e.g., Model 3/Y) uses roughly 4,000–6,000 kWh per year for average US driving (~13,000–14,000 miles/year at ~3.5–4 miles/kWh efficiency).
• Even assuming self-driving/robotaxis drive much more (e.g., 5x the miles for fleet use), the per-vehicle annual use might rise to ~20,000–30,000 kWh.
• For 1 million vehicles:
• Standard personal use: ~4–6 TWh/year total.
• High-utilization robotaxi scenario: ~20–30 TWh/year total.
• This is just 0.1–0.7% of current US generation—tiny compared to existing demand or growth from data centers, manufacturing, and electrification.
By comparison, the US already has ~7 million EVs on the road as of 2025, and their charging load is well under 1% of total grid demand in high-adoption states like California.
Peak power and timing (the real grid concern)
The bigger issue for grids is usually when cars charge, not total energy. US summer peak demand hit records of ~759 GW in 2025. Even if all 1 million cars charged simultaneously at a typical Level 2 rate (~7–11 kW each), that’s only ~7–11 GW—about 1% of peak demand, and easily spread out.
In reality:
• Most charging happens at home or depots overnight/off-peak, when demand is much lower.
• Tesla vehicles (and future robotaxis) support smart/scheduled charging and can integrate with vehicle-to-grid (V2G) or virtual power plants (VPPs) to help stabilize the grid rather than strain it.
• Utilities already use managed charging programs that shift EV load away from peaks.
Studies from the Department of Energy, EPA, and others confirm the grid can support tens of millions of EVs (projections of 30–42 million light-duty EVs by ~2030) with proper planning—no collapse expected.
Grid is already expanding fast
• Record renewable + battery additions are happening (dozens of GW of solar/batteries annually).
• Peak demand is growing (forecasts up to ~850–930 GW by 2030 from data centers, EVs, etc.), but new capacity is outpacing it.
• Distribution-level upgrades are needed in some neighborhoods, but that’s local—not a national barrier for just 1 million vehicles.
Tesla has long argued (and data supports) that EVs, including high-mileage robotaxis, are net-positive for the grid when paired with their energy products (Powerwalls, Megapacks, VPPs). Self-driving cars could actually improve utilization by charging strategically at depots or during low-demand periods.
Bottom line: 1 million Tesla self-driving cars would add a trivially small load that the grid is already equipped (and rapidly expanding) to handle. The bigger long-term story is how EVs + storage can make the grid more reliable and renewable-friendly. No major upgrades beyond ongoing trends would be required for this scale.
Thank you for reading. Feel free to share.