Chapter 12: Can't Take That Cost

She ran the food first.

Morath produced food for 8 billion people across the 11-system network. Seventy-three percent of that food moved through the distribution network managed by Dominion's agricultural bureau on Veradis Prime. With that administration gone, distribution fractured: some systems had sufficient local reserves, some had two to four months of stored food, some—particularly the industrial worlds like Kethel, where agricultural land had been converted to production—had less than six weeks.

She ran the numbers. If the Dominion administrative apparatus collapsed immediately after 3rd Fleet defection, she was looking at 400 million people in acute food insecurity within three months. Mortal starvation risk climbing for 800 million in the first year. The model got worse in year two, as the cascading effects of production shutdowns hit medical and industrial infrastructure. By year three, she was looking at a range of 1.8 to 4.2 billion deaths from compound collapse, depending on how quickly Fracture forces could establish replacement administrative structures.

She looked at that number for a long time.

The Sovari genocide had killed approximately 6 billion people over a period of ten years. The Fracture's "victory" through accelerated collapse would kill, by her best conservative estimate, 2 billion people over five years. In the worst-case model: 4 billion.

She had done enough catastrophe modeling in her career to feel the mathematics of mass death as something beyond abstraction. These were not numbers. They were food distributions not arriving, medical equipment not functioning, heating systems failing in winter cycles, the compound cruelty of infrastructure collapse in a society that had been engineered for dependence on central administration. These were the specific, material deaths of people who had survived everything the war had thrown at them and would die instead from the administrative absence that followed the Fracture's victory. People who had not committed the genocide. People who did not know the genocide had happened. People whose only crime was having been born into a system that fed them from the same administrative hub it was about to lose.

She sat with this for a while, longer than was analytically necessary. She did not have the skill of making mass death manageable by looking at it from a sufficient distance. She had never tried to develop that skill. The distance felt like cowardice, and she had other faults.

The Fracture knew this. Senn knew this. She had been clear: speed matters. The faster the Dominion fell, the fewer the deaths from the collapse—because reconstruction could begin sooner, because the Fracture could establish replacement systems faster if the transition was planned rather than chaotic.

Kael ran the accelerated-collapse scenario, assuming Senn's best-case transition planning. If the war ended in thirty days from 3rd Fleet defection, Fracture forces controlled all systems within four months, reconstruction began immediately under Fracture administration: she got somewhere between 600 million and 1.4 billion deaths. Better than the catastrophic case. Still more people than she could hold in her mind as an abstract number.

She closed the model.

The mathematical truth: there was no version of 3rd Fleet defection that did not result in hundreds of millions of deaths. Senn's offer was the fastest path to a morally justified ending—the Dominion deserved to fall—but the mechanism of that fall would kill civilian populations who were not responsible for the genocide, who had never voted for the genocide, who did not even know the genocide had happened.

She opened a blank message in the encrypted terminal—her private terminal, not flag communications. She typed: I can't take that cost.

She looked at the words for a moment.

Then she closed the message without sending it.

Because Senn would receive it as a tactical refusal, and it was—but it was also the first honest thing Kael had expressed about any of this, and expressing honesty to the enemy commander did not seem wise even in encrypted private channels. She was a fleet admiral. She had no luxury of honesty that carried operational risk.

But the thought needed to exist somewhere. Even if only in her own record.

She pulled up the Ceros Station engagement plan and looked at it from the position of someone who knew the Dominion's cause was unjust. The engagement plan was still, in its own terrible way, beautiful. Three-pronged approach, using the asteroid cluster's radar scatter as cover for the flank moves, timing the main battery engagement to coincide with the moment Fracture defensive positioning was most exposed.

Beautifully constructed. Designed to maximize Fracture casualties.

She looked at the model for the Fracture forces at Ceros Station—26 vessels, including two command-class cruisers, roughly 18,000 crew. People who had chosen the Fracture's cause, who had read the Sovari files or been told the truth by people who had read them, who were fighting a war that Kael now believed was fundamentally justified. People who had looked at the same evidence she was looking at and had made a different choice than she was making—who had decided that fighting was the answer, that the truth was worth dying for—and who would die at Ceros Station because she was better at this than Senn was and the institution she served had more ships.

Her battle plan would kill most of them.

She couldn't not fight this battle. She couldn't refuse. She'd sent the compliance signal. She had 3rd Fleet moving toward Ceros Station with the momentum of a decision that had already been executed.

But maybe, maybe, she could fight it differently than she was supposed to.

Not defection or refusal—something else.

She started making small marks in the engagement plan—adjustments to timing that would look like tactical refinements. Adjustments that would, if Senn's defensive forces read them correctly, telegraph the approach pattern two minutes earlier than Kael's standard technique. Two minutes was a long time in a fleet engagement. Two minutes was enough for ships to begin evacuation procedures. Two minutes was enough for the Fracture's best people to get off their command ships before those ships became debris.

She made the adjustments with the same precision she brought to everything—small enough to be explainable as efficiency refinements, deliberate enough to be readable by someone who had spent five years teaching her to think. If Senn saw what she was seeing, she would know. If Senn's people were competent interpreters of her tactical pattern—and they would be; Senn trained people the same way she trained Kael—they would read it.

She might not be able to avoid this battle. But she could fight it in a way that gave the other side a chance to survive it.

That was not defection. That was not betrayal in any form she could name.

She didn't know what it was. She knew only that she couldn't write the model for mass civilian death and then immediately execute the maximum-casualty version of the engagement plan as if the two things were in separate rooms in her mind.

She would go to Ceros Station. She would do what High Command had ordered. She would find the edges of the order—the margins, the interpretive spaces—and try to fill those spaces with something less catastrophic than the order at its most efficient.

It wasn't enough. She knew it wasn't enough. But it was what she had right now, at 0400 on a flag bridge, alone with numbers and a message she hadn't sent.

I can't take that cost.

She kept the thought, and turned back to the plan.