Solaria hands the beam to software and the dust front doesn't get a vote

A dust front is coming for your rectenna field, and you have about ninety seconds to decide what to do about it. That's the problem Solaria's control engineers set out to solve this rotation. Here is the beautiful part: this time, nobody decided anything. The software did.

In a live trial from the Array's downlink control deck, a handoff routine watched a modeled haze start to degrade reception at a primary rectenna field on the coastal spine. It found a clear field two hundred kilometers inland and moved the beam there. The numbers tell the whole story: 2.1 gigawatts, shifted in under four seconds, with no measured dropout on the grid side. A load that size vanished from one mesh and reappeared on another, and the Orbital Exchange's frequency readouts never flinched.

"We've always been able to reroute," said Mara Venn, who runs downlink timing for the Array. "What we couldn't do was reroute faster than a human could second-guess it. A dust front doesn't negotiate. By the time you've called the ground station, confirmed their field is clear, and walked the handoff through, you've already spilled power or dropped the corridor. The machine closes that gap."

That gap isn't abstract. The maintainer shortage Ken Nakashima keeps documenting from the ground side is, in part, a shortage of exactly these decisions — trained corridor crews awake at the right hour to catch a degrading beam before it becomes an incident. If the handoff holds, Solaria's argument goes, you need fewer people holding their breath over a readout at dawn. You need the physics to hold its own breath instead.

The corridor veterans aren't buying it.

"Four seconds is fast, and four seconds is the whole problem," said Tomas Okonkwo, a field supervisor on the coastal spine with two decades on the receiving end of beams like this one. "Nobody fully trusts that routine yet — not the people who wrote it, if they're honest. And 'automation nobody fully trusts' is the exact phrase from the near-miss review. We were told the lesson was more eyes on the beam, not fewer."

He's got a point the Array concedes. The near-miss that prompted that review turned on a handoff a human caught and a system had missed. Venn's team ran this trial with the full crew standing by and the manual override armed the entire time. The first demonstration of the beam routing itself still required everyone who would eventually be routed around it.

Solaria plans a second trial against a real dust front, not a modeled one, in the next transfer-planning cycle. The override stays armed. The grid, for its part, never noticed it was being handed off — which is either the point or the worry, depending on whether you build the bridge or stand under it.

"Measure it in watts," Venn said. "Two-point-one gigawatts, no dropout. That's not an opinion. That's a number we have now that we didn't have last rotation."