Chapter 240: Chapter 240 – The Lateral Stage
Soren did the geometry.
He had been listening while Neral explained the lateral stage concept and he had his notebook out before Neral finished the second sentence. He ran calculations for twelve minutes while Kai and Neral went through the ancient grammar in detail, then interrupted with the precision of someone who had been waiting for a pause.
"The ancient network’s nearest active stage is the one we camped above, approximately two hundred kilometres east. That’s the connection point." He turned the notebook. "The lateral stage needs to bridge from there to the drain’s pressure field boundary, which is approximately forty kilometres west of that stage. Total build: forty kilometres of lateral substrate routing."
He looked at Kai.
"Stage 5 in Vael’s Crossing was twenty-three elements over three weeks. A lateral stage at this depth, based on the ancient grammar Neral has described, would require significantly more elements per kilometre due to the horizontal routing complexity. I’m estimating forty-five elements per kilometre. Forty kilometres gives you roughly eighteen hundred elements."
A pause.
"At Vael’s Crossing’s rate, that’s approximately five years."
Right.
Kai looked at the geometry and thought through what he actually knew. Vael’s Crossing’s Stage 5 had used the builders’ vertical grammar—he’d been learning as he built, making corrections, finding the right pace. He knew the vertical grammar now. The lateral grammar was different but built on the same fundamental sovereign seed mechanism. And he had Source Point integration now, which had not existed during Vael’s Crossing.
"The ancient grammar is simpler than the builders’," he said. "Fewer elements per segment. And each segment can run longer before needing a junction point." He looked at Soren’s estimate. "Revise the elements-per-kilometre down. Significantly. The ancient network was built for durability over precision."
Soren looked at Neral. Neral nodded once.
Soren revised.
"If the element count is one-third of my estimate, the build is twenty-three days at Vael’s Crossing’s pace. At higher output—" he paused— "seven to ten days. But you’d be drawing significantly more per session."
"That’s the build," Kai said. "Let’s start."
The first segment took four hours.
Vertical construction had been like carving—the sovereign seed depositing path-energy into the substrate in a downward direction, the geometry following the builders’ architectural specifications. Lateral construction was different. The path-energy had to hold its position horizontally in a substrate that didn’t provide the same natural guidance a vertical channel did. The ancient grammar accounted for this with a different kind of junction architecture—Neral had spent the previous evening explaining it and Kai had understood it in theory.
In practice, the first attempt failed at the three-hour mark when the segment’s lateral coherence drifted three metres from the intended routing.
He stopped, recalibrated, and restarted from the drift point.
This is why you don’t assume the theory is the practice.
The second attempt held. The first segment set at hour four. The pool was at forty-one percent.
He surfaced and let Adaptive Recovery work for three hours, ate, and went back down.
By nightfall he had three segments. Forty-five were needed. He was building faster with each one.
Seven to ten days. That’s the shape of it. You’ve done harder things.
Day three. The lateral stage connected to the ancient network’s active stage.
He had been building east from the drain’s pressure boundary toward the ancient stage. The ancient stage had been running at its low continuous output for geological time, routing west as far as its architecture extended. When the lateral stage’s easternmost segment reached the ancient stage’s western boundary, the connection happened without any activation output from Kai—the ancient grammar recognised the compatible construction and integrated it immediately.
The deep pressure at the drain began routing east. Not much—three segments of lateral stage out of forty-five meant only a fraction of the accumulated pressure had a channel to move through. But the depression’s floor vibration changed. He felt it through the substrate before he surfaced.
Outside, Soren had been monitoring with his portable equipment.
"Vibration rate is down eleven percent from baseline," he said. He showed Kai the measurement. "In three days. That’s faster than I’d projected from the limited lateral routing."
Because the drain was already under pressure when we started. Any release point changes the dynamic immediately.
"It’ll accelerate as we add segments," Kai said.
Soren wrote this down. He had an expression Kai had come to recognise over the past months: the specific focus of someone whose model is being updated by live data faster than he can write it down.
He looked at the timeline.
Three segments in three days. Forty-two remaining. At current rate: fourteen days. The breach timeline, minus three weeks’ elapsed travel: roughly three and a half months remaining. Fourteen days was achievable with margin to spare.
He sat with that for a moment.
The problem had been genuinely unknown four days ago. He hadn’t known what was below the depression, hadn’t known the intervention method, hadn’t known whether the carrier function could build lateral at all. Now he was three segments into a solution and the numbers were working.
This is going to work.
Not triumph. Just the specific clarity of a problem that had found its shape and was now being solved. He went back to the substrate and built segment four.
Lateral stage build status: Day 3
Segments complete: 3 of 45
Connection to ancient network: established
Depression vibration: -11% from baseline
Pool management: Adaptive Recovery viable at current pace
Breach timeline remaining: ~3.5 months
He was mid-way through segment four when the Architect communicated.
Not direction. Not the patient coordination quality it had been running since the plateau. Urgency. The first urgency the Architect had sent since the carrier first connected at the convergence point, and Kai felt it the same way he had felt it then: not through words, through the quality of the five-node connection shifting.
He surfaced immediately.
Mira was already holding the vault pair in both hands when he came up. She had felt the shift before he reached the surface.
"The Architect’s coordination layer is stressed," she said. Not asking. Telling. "The lateral construction is drawing on its substrate reach to hold the ancient network connection stable while you build. Managing five outer Rifts and supporting the lateral build simultaneously is more than its coordination layer was sized for."
She held the shells.
"At current build pace, the Architect will reach its coordination limit in eleven days."
Kai looked at her. Eleven days. Three days short of the fourteen-day build estimate.
So you either build faster or the Architect starts losing stability before the stage is complete.
Faster meant more pool draw per session. More pool draw meant shorter recovery windows. Shorter recovery windows meant running closer to the edge of what Adaptive Recovery could maintain.
He had run close to the edge before. Ren-Sarath’s activation sequence had pushed the pool to levels that required conscious management. He knew what that felt like. He knew where the line was.
He thought about this for four seconds. Genuinely four seconds—it wasn’t a hard calculation.
"Then we build faster," he said.
He looked at Neral.
"The junction architecture—can segments be extended before setting? Longer segments, fewer junctions, same total coverage?"
Neral looked at his grammar notes. He had been ready for this question, which Kai noticed and appreciated. Neral’s theatrical quality had entirely disappeared months ago. What had replaced it was this: readiness. The specific readiness of someone who knew the material so well they were already at the next question.
"Yes," Neral said. "The ancient grammar allows longer segments at higher coherence cost. The junction architecture stays the same. Each segment runs further before the junction is needed."
"That cuts the element count," Kai said.
"It does. But the pool cost per segment increases proportionally."
"I know."
He went back down.
The next segment took two hours and thirty minutes instead of four. The pool was at thirty-eight percent when he surfaced.
He gave Adaptive Recovery ninety minutes instead of three hours.
Then he went down again.
