A Carrington-Level Storm: $2T+ Damage, Near-Zero Hardening
A 1-in-150-year geomagnetic superstorm would destroy thousands of high-voltage transformers and knock out power for months to years. This risk has been known for decades. The grid has not been hardened.
$2T+Estimated damage from Carrington-level event
1-in-150Annual probability of Carrington-class storm
1859Year of the last Carrington-level event
~0%Grid hardening progress since risk identified
Choose your depth. The data doesn't change — just the explanation.
In 1859, the sun shot out a massive burst of energy that made telegraph wires catch fire. If the same thing happened today, it would fry millions of electrical transformers — the big boxes that make the power grid work. Replacing them takes 1-3 years and they're mostly made overseas. We'd lose power to large parts of the country for months. This has a roughly 1-in-150 chance of happening any given year. Scientists know about this risk. We have done almost nothing to prepare.
A Carrington Event is a geomagnetic superstorm caused by a coronal mass ejection (CME) hitting Earth's magnetic field. In 1859, the resulting geomagnetic induced currents (GICs) burned telegraph lines. Today, those currents would flow through the power grid, overwhelming high-voltage transformers (HVTs). HVTs take 12-36 months to manufacture and cost $3-10M each. The US has ~2,000 large HVTs. A 2013 Lloyd's of London report estimated Carrington-equivalent impact at $0.6-2.6T in the US alone, with 20-40 million people losing power for 1-2 years. NOAA and NASA estimate Carrington-class CME probability at ~0.67%/year (1-in-150).
Carrington Event (September 1-2, 1859): Dst index estimated -850 nT (compared to -589 nT for the 1989 Quebec blackout that affected 6 million). Faraday cage approach for HVT protection: GIC blocking devices (neutral-to-ground resistors) cost ~$150K per transformer. Full US HVT protection: ~$1-2B investment. Expected annual cost: $7-13M at 0.67%/year probability × $2T damage = $13.4B annual expected loss without action. EMP vs. GMD distinction: EMP (nuclear detonation) is faster-rise, shorter-duration; GMD (solar) is slower but affects all exposed infrastructure simultaneously. FERC Order 830 (GMD Reliability Standard, 2019) established baseline planning standards for utilities but implementation remains inconsistent. The 2024 solar maximum creates elevated risk window through 2026.
Lloyd's 2013 report "Solar Storm Risk to the North American Electric Grid": lloyds.com/solarstorm. NOAA Space Weather Prediction Center: swpc.noaa.gov. FERC Order 830: ferc.gov/industries-data/electric/electric-reliability/geomagnetic-disturbances. NIST EMP/GMD hardening guidelines: nist.gov/ems. Oak Ridge National Laboratory HVT analysis: ornl.gov/publication/large-power-transformers-vulnerabilities. NAS 2008 report "Severe Space Weather Events": nap.edu/catalog/12507. 2024 solar maximum forecast: spaceweather.gov.
The Solar Storm Risk We're Ignoring
Geomagnetic Storm Frequency vs. Intensity
Annual occurrence rate by storm class (Dst index severity)
Estimated Grid Recovery Time vs. Protection Cost
Recovery months without hardening vs. investment cost to harden
Population Affected vs. Storm Intensity
Millions of Americans losing power for 6+ months by storm severity
⚡ Why This Is Worse Than Any Nuclear Plant Failure
When a single power plant fails, neighboring grid sections reroute. When thousands of transformers fail simultaneously across a continent, there's nothing to reroute to. HVTs aren't like circuit breakers — they're custom-built, 100-ton pieces of equipment with 12-36 month lead times, mostly manufactured in Europe and Asia. Without power: water treatment fails within days, hospitals go to backup generators (typically 72-96 hour fuel supply), GPS and telecommunications degrade, and supply chains collapse. The cascading effects extend 18-24 months after the original event. The cost to harden US critical HVTs: ~$1-2 billion. The estimated loss from one Carrington event: $2 trillion. Congress has not appropriated the hardening funds.