Infrastructure

Black Start

At 12:33 CEST on Monday 28 April, the power systems of mainland Spain and Portugal collapsed, along with a small corner of France. Around 60 million people lost electricity. Trains stopped, traffic lights went dark, mobile networks faded as backup batteries ran down, card payments failed, and people walked home from work across Madrid and Lisbon. Grid operators had to perform a black start, rebuilding the system from the ground up, and restoration continued through the night into Tuesday.

At the time of writing, the cause is still under investigation, and the European grid operators’ association (ENTSO-E) is setting up an expert panel.

Everything Assumes the Grid

Within a month we’ve had one substation fire close Heathrow and now a whole peninsula going dark. In IT resilience planning we often treat power as a given: there’s a UPS, there’s a generator, job done. Monday asks harder questions:

  • How long is your generator fuel good for, and who delivers more when the roads are jammed and the phones are down?
  • Can your people reach the site, or even reach each other?
  • Do your communications depend on the same grid? Mobile networks, VoIP and the cloud-hosted chat tool you use for incident bridges may all go down too.
  • What’s the restart order? Bringing everything back at once creates its own surge, in electrical load and in application load.

Black start is a good concept to borrow. Do you know how you would restart your service from nothing, rather than just fail over from something?

As an interesting aside, I live in Ireland and have family in Australia. Both are island grids with limited interconnection, where this kind of event has a different flavour. That’s a post for another day.

More Information

Postscript (March 2026): ENTSO-E’s final report, published on 20 March 2026, found no single cause. It pointed to a combination of many interacting factors, chiefly voltage control and oscillations, which led to over-voltage and cascading disconnections.