The first installation reports came back in three batches.
The initial six mines had received their engines in the first shipment wave and had been operational long enough to produce a meaningful comparison dataset by the time the reports reached Jonas's desk — four weeks of pre-installation output figures against four weeks of post-installation output figures, the kind of clean before-and-after comparison that removed most of the variables that made performance assessment ambiguous.
He read the first report, noted the numbers, and read it again.
The projection he had given the Duke — a conservative five times increase in extraction capacity with proper pump installation — had been built on the assumption that the primary constraint in the mines was drainage. Remove the drainage problem and the existing labor force could work longer shifts, reach deeper veins, and maintain consistent output rather than the stop-start rhythm that groundwater intrusion imposed on conventional operations. Five times had been the honest conservative estimate, the number he was confident the engines would achieve without requiring any other operational change.
The first report was showing seven times.
He cross-referenced against the second report. Six times. The third: eight.
He sat with this for a moment.
The variance across sites made sense — different mines had different geology, different water table depths, different labor force sizes, different management quality. The range was expected. What he had not fully modeled was the secondary effect: removing the drainage problem did not merely allow longer shifts, it changed the entire rhythm of mine operation in ways that compounded. Shifts that had previously been interrupted by flooding now ran uninterrupted. Workers who had previously spent significant portions of their working hours bailing and clearing could redirect that labor entirely to extraction. Deeper veins that had been technically accessible but practically unreachable because the drainage cost was prohibitive became viable. The engine did not merely solve the water problem. It unlocked everything the water problem had been blocking.
He revised his projection upward, ran the new numbers against the Duke's investment ledger, and calculated the revised return timeline.
The original estimate had been thirty times investment in three years.
The revised estimate was closer to fifty, possibly more, depending on how aggressively the remaining mines were equipped and how quickly the operations adapted to the new rhythm.
He wrote a summary, attached the revised projections, and sent it to Mikayla's desk for forwarding to the Duke with a note that a meeting was warranted when the Duke's schedule permitted.
Then he went back to the facility floor, because the numbers were good and good numbers were not a reason to stop working.
The Duke read the summary that evening.
He read it twice, which was not his standard practice for documents he understood on the first pass. Then he set it on the desk in the position he used for active matters and looked at it for a while with the expression of a man whose internal model of a situation has just been revised upward by an amount that requires the model itself to be restructured rather than merely adjusted.
He had commissioned steam engines for his mines because Jonas had presented a coherent case and the first working model had been compelling. He had expected results. He had expected, in the measured way of a man who had been in Bavarian politics long enough to know that projections were optimistic by nature and should be discounted accordingly, results somewhat below projection.
The document on his desk was describing results above projection.
He thought about the other houses. He thought about the magic stone supply chain. He thought about the twelve mines currently without engines and the revised timeline for equipping them. He thought about what it meant for House Meister's position in the German Empire when the full installation was operational and the output differential between his mines and every competitor's mines became visible in the market.
He sent a message to his purchasing office: accelerate the next engine order. Whatever the facility's current production rate was, he wanted to know the maximum rate and what it would cost to reach it.
Then he sent a second message, to a different recipient, which was not about engines at all.
Jonas was in the facility's eastern workshop reviewing the tolerance corrections Klaus had implemented when the guard appeared at the workshop entrance and informed him that there was a visitor at the security post.
Not scheduled. The visitor log had nothing for this afternoon.
He noted this without visible reaction and told the guard he would be there in ten minutes, which he was.
The carriage at the security post was not the Meister banner. It was a different house color — grey and silver, the colors Jonas had identified from the Duke's map room annotations as belonging to a house in Munich's inner circle whose name he had encountered repeatedly in the political texts without ever having occasion to interact with directly.
He did not recognize the banner as Wallenstein's. He recognized it from the map room.
He stood at the security post and looked at the carriage and ran a rapid assessment.
An unscheduled visit from a house he had no direct relationship with, arriving at the facility rather than the ducal estate, using the facility's address rather than the Duke's as the point of contact. This was either a commercial inquiry — someone who had heard about the engines and wanted access — or something else, and the specific choice of approach told him more about which of those it was than the visitor's identity would.
A commercial inquiry went through the Duke's commercial office. That was the correct channel and any commercially sophisticated party knew it. A visit that bypassed that channel and came directly to the facility was either deliberately unorthodox or was about something that the visitor did not want processed through the Duke's administrative structure.
Neither option was uncomplicated.
"Send them through," he said. "Standard security procedure. Escort to the administrative building."
He walked back to the office at his usual pace.
The man who came in was not Count Wallenstein.
He was perhaps thirty-five, well-dressed in the grey and silver of his house's colors, with the composed professional manner of a senior house agent rather than a principal — the bearing of someone who carried authority without holding it personally, who spoke for someone rather than as someone. He introduced himself as Aldric Venn, commercial representative for House Wallenstein, and he sat in the chair across Jonas's desk with the ease of a man who had conducted many meetings in rooms he did not own.
"Lord Jonas." His voice was smooth without being ingratiating. "Thank you for seeing me without prior arrangement. I appreciate the flexibility."
"I had time," Jonas said, which was approximately true and entirely noncommittal.
"House Wallenstein has become aware of the engineering work being conducted under your direction," Venn said. "The steam engine, specifically. We understand it has been deployed in several mining operations with significant results." He paused. "House Wallenstein controls fourteen mining operations in the Bavarian and Upper Austrian territories. We would be interested in discussing an equipment supply arrangement."
Jonas looked at him.
The sentence was simple and the request was on its face legitimate — an interested party approaching a supplier about access to a product. In any context where Jonas did not know what he knew about House Wallenstein's intelligence operation, about the agent who had arrived through his window and left as ash, about Count Wallenstein's financial intelligence initiative directed specifically at understanding what was happening inside this facility, the conversation would be straightforward.
He knew what he knew.
He kept his expression in the register of professional consideration and ran the threads simultaneously.
What Wallenstein was doing was elegant, in a specific way that deserved acknowledgment even from the receiving end. The direct intelligence approach — the agent in the estate, the attempts to get sources inside the facility — had failed to produce useful information. The financial intelligence approach was producing information about supply chains and material costs but not about the product itself. The logical third move was commercial engagement: approach the supplier directly, establish a relationship, use the relationship to get access to what covert methods had not obtained.
Not necessarily the product itself. The knowledge. A commercial arrangement required technical discussion. Technical discussion produced documentation. Documentation, in the hands of skilled analysts, produced understanding.
The request was a Trojan horse, constructed with professional competence and delivered with a smooth face.
"House Wallenstein's mining operations," Jonas said. "Fourteen sites." He said this as though he were processing new information, which required some effort because it was not. "That's a significant footprint."
"It is," Venn agreed. "Which is why we are keen to ensure our operations remain competitive."
"I understand the interest." Jonas leaned back in his chair slightly — the posture of a man who is genuinely considering rather than stalling. "The current production capacity at this facility is allocated to House Meister's requirements for the foreseeable future. I am not in a position to take on external orders without the Duke's authorization, and I am not in a position to seek that authorization for a competing house without considerable justification."
Venn's expression remained composed. He had expected this answer or something close to it. "We are not asking for an immediate commitment. We are asking for a conversation — a preliminary discussion about whether an arrangement might be structured in a way that benefits all parties."
"Including House Wallenstein," Jonas said.
"Naturally."
Jonas looked at him for a moment. "What is House Wallenstein's current extraction rate across the fourteen sites?"
Venn blinked — a fractional hesitation, the slight lag of someone who had not expected the conversation to move in this direction. "I don't have those figures to hand."
"That's fine." Jonas smiled — not warmly, but pleasantly, which was a different thing. "I ask because any arrangement I would consider presenting to the Duke would need to demonstrate that the operational benefit to House Wallenstein is significant enough to justify the political complexity of supplying a competing house. Without the baseline figures, I cannot assess that." He paused. "If you can obtain them, I'm happy to have a substantive conversation. If not—" he spread his hands in the mild gesture of a man identifying a constraint rather than closing a door "—we are discussing something neither of us can evaluate."
Venn looked at him for a moment with an expression that had shifted fractionally from its opening position — not suspicion exactly, but the recalibration of someone who had entered the conversation with a specific sequence in mind and found the sequence had been redirected.
"I can obtain them," he said.
"Then send them to my administrative office and we will arrange a follow-up." Jonas rose, which closed the meeting with the polite inevitability of a signal that was clearly given. "I appreciate House Wallenstein's interest. The technology has significant applications and I expect the conversation to be worth having once we have the right information in front of us."
Venn stood, inclined his head with the professional courtesy of someone who recognized a graceful termination when he encountered one, and left.
Jonas sat back down and looked at the door for a moment.
What he had just done was buy time while communicating nothing useful, deflect a request for access without refusing it, and request information from House Wallenstein that they would now spend resources compiling — which was time and attention directed toward producing a document for him rather than toward whatever they had originally intended. The follow-up meeting, if it happened, would happen on his terms, in his facility, with a prepared position rather than an improvised one.
He picked up his pen and made a note.
Then he sent a message to Mikayla's desk: House Wallenstein had made direct contact at the facility. She should know.
He did not write the assessment he had formed about what the contact meant. Mikayla was capable of forming that assessment herself, and she would, and she would bring it to the Duke, and the Duke would draw the conclusions that the Duke drew. That chain of events was already in motion.
What mattered now was what he did with the time it produced.
He turned back to the tolerance review and continued where he had left off.