Chapter 61: The First Prototype
The first failure arrived before breakfast.
Adrian considered that a good sign.
He was standing inside Calder’s advanced engineering facility at 6:47 in the morning when a monitor on the far wall flashed red.
THERMAL INSTABILITY — TEST CELL THREE.
A warning alarm sounded once and stopped. Nobody moved with urgency. Nobody reached for a phone. The engineers around the test chamber looked at the data, made notes, and continued working.
That was one of the first things Adrian had insisted on when the new programmes were established. Failures were expected. Hidden failures were not. The difference between a team that learned and a team that stagnated was almost always whether people were willing to say what had gone wrong before they’d figured out how to fix it.
Dr. Miriam Shaw came through the laboratory doors carrying a tablet in one hand and a coffee in the other that she hadn’t had time to drink yet.
"Helios?"
"Prototype cell," one of the engineers answered without looking up. "Terminated at sixty-eight percent of target load. Thermal spike at the electrode interface."
"How contained?"
"Completely. No safety concern."
Miriam stopped beside Adrian and looked through the observation glass at the test chamber. Inside was a battery cell barely larger than a paperback book — months of materials science and chemistry and manufacturing iteration compressed into something you could hold in one hand.
"Why did it fail?" Adrian asked.
"The interface," the engineer said.
Miriam nodded. "Exactly where we predicted it might."
Adrian looked at her. "Then why are you smiling?"
She glanced at him. "I’m not."
"You are."
"Because we know which variable is wrong." She turned toward the team. "How many samples in the current batch?"
"Thirty-two."
"Survivors?"
"Seven."
"Good."
Adrian raised an eyebrow. "Seven out of thirty-two is good?"
She looked at him with the specific patience of someone who had been doing materials research long enough that other people’s intuitions about success rates had stopped surprising her.
"For research at this stage? Seven survivors tells us the pathway is viable. Zero survivors would tell us to rethink the chemistry. Seven tells us to refine the manufacturing." She turned back to the team. "Document everything from all thirty-two, including the failures. Especially the failures."
At 7:15 Adrian crossed into the automotive division.
The building was already running at full pace. Engineers gathered around chassis components arguing over tolerances with the particular productive intensity of people who were too close to a problem to be interested in anything else. Manufacturing specialists were running a simulation on a large display. Three software developers had claimed the corner workstations and were running what appeared to be competing models of the same thermal system.
At the centre of the main workshop sat Orion.
More accurately, what would eventually become Orion.
No polished bodywork. No interior. No dramatic staging. A bare aluminium-and-composite chassis resting on four temporary supports, surrounded by measurement equipment and data cables, looking more like the subject of a surgery than the object of ambition.
Adrian walked slowly around it.
Laura Bennett, head of automotive engineering, appeared at his shoulder.
"We’ve changed the architecture."
"I can see."
"The wheel-mounted motor configuration created more thermal complexity than the models suggested. We’re testing a compact in-wheel unit with a redesigned cooling pathway instead."
"Why is that better?"
"Because the original architecture would have cost us eighteen months solving a problem that doesn’t actually improve the vehicle meaningfully. The redesign achieves the same performance characteristic through a simpler manufacturing process." She paused. "I wanted to flag it before changing the technical spec."
"Is the redesign better?"
"Yes."
"Then the original spec was wrong." He kept walking around the chassis. "That’s not a criticism. That’s what development means."
Laura looked at him with the slightly surprised expression people had when they’d been prepared to defend a decision and hadn’t needed to.
"Some engineers struggle with that," she said.
"I know." He stopped and looked at the chassis. "Build the best version, not the first version."
At 8:03 the first Orion drivetrain test began.
The temporary test vehicle sat on a reinforced platform while sensors tracked torque, temperature, vibration and power consumption across forty-seven measurement points. The room went quieter in the specific way it went quiet when something was about to happen and everyone present had a stake in the result.
Laura stood beside Adrian.
"Thirty-second cycle."
The motors activated. A low mechanical hum filled the workshop.
The readout climbed.
12 kW. 28 kW. 41 kW.
The temperature remained stable. Laura’s posture was very still — the controlled tension of someone watching data rather than watching a spectacle.
63 kW.
Then — FAULT. The motors stopped.
The room held its silence for a moment.
One of the younger engineers glanced at Adrian with the expression of someone who hadn’t yet learned how to calibrate their reaction to a test failure in front of the person responsible for the budget.
Adrian looked at the data display.
"Overcurrent," Laura said. "Rear drive controller."
"Component?"
"Software. Not hardware."
Adrian looked at the young engineer who was still watching him.
"Fix the software," he said. "Run it again."
"That’s it?"
"What else would you like me to do?"
The engineer didn’t have an answer.
Adrian looked at the team.
"This is development. We don’t celebrate thirty seconds of clean operation and we don’t panic when it fails at thirty-one. Find the problem, fix it, run it again. That’s the cycle." He looked at the test platform. "How long to patch the controller software?"
"Two hours," the software lead said.
"Then we have the next test at ten."
Robert arrived at 9:20 and stopped just inside the workshop entrance.
He looked at the chassis, at the engineers, at the measurement equipment, at the complete absence of anything that resembled a finished product.
"I was expecting something more impressive."
"It’s been four days," Adrian said.
"I know." Robert walked closer. "How long until it moves under its own power?"
"Properly? Six weeks."
"That’s faster than the original estimate."
"We simplified the architecture."
"And the battery situation?"
"Still the bottleneck. Helios is still failing."
Robert nodded slowly.
"Good," he said.
Adrian looked at him.
"You too?"
Robert smiled. "I’m learning what good failures look like." He fell into step beside Adrian as they moved toward the executive offices. "There’s something else."
"Government?"
"How did you—"
"You only say ’there’s something else’ in that tone for things you think I might not want to hear immediately."
Robert produced a folder.
Inside was a letter from the Department for Business and Trade. Formal, carefully worded, congratulatory about Calder’s recent expansion into strategic technology sectors. It expressed interest in discussing the long-term implications of the programmes — domestic manufacturing, resilient communications infrastructure, energy storage.
"When?" Adrian asked.
"Tomorrow afternoon."
"Accept."
"Already done."
Adrian handed the folder back.
"Good."
The Aether laboratory was quieter than the automotive floor.
Unlike Orion, Aether had no physical structure to point at. The room contained engineers working at modular electronics benches, each one surrounded by components in various states of assembly and disassembly. The central display showed a device roughly phone-sized that had been almost entirely taken apart, its components arranged around it like the exploded diagram in an engineering manual.
Adrian stopped.
"This is uglier than the previous version."
The lead engineer, a man named Chen who had joined from a semiconductor firm that had been working on miniaturised neural processing chips, looked up without apparent concern.
"Version three."
"Why does it look worse than version one?"
"Because version one was designed to look good. Version three is designed to work." He picked up the display assembly and showed Adrian how it detached cleanly from the frame. "Full modular replacement. The user can swap battery, display, and communications module without tools and without voiding any structural integrity."
"How long does it take?"
"Under two minutes, currently."
"Target?"
"Under ninety seconds."
Adrian looked at the separated components.
"Thermal management?"
"Still the primary problem. Everything we want to do requires space and space requires a thicker device."
"How much space?"
"Eight millimetres more than the current commercial standard."
"Take it."
Chen looked at him. "The market standard is—"
"The market hasn’t seen this device yet. Make it powerful enough that eight millimetres doesn’t matter to the person holding it." He looked at the display. "What does it do that nothing else currently does?"
Chen began listing. Satellite connectivity for emergency communications. On-device AI processing without cloud dependency. Secure communications architecture that didn’t route through third-party infrastructure. The modular hardware system.
Adrian was listening, but his attention had moved to the secondary display, where a network diagram was running.
"What’s that?"
The telecommunications engineer, a woman named Priya who had joined from a satellite communications startup, turned from her bench.
"Hybrid network architecture test. We’re modelling what happens when the device has no terrestrial coverage at all — it switches to satellite. Latency is manageable for messaging and emergency voice."
"What about peer-to-peer? Two devices communicating directly without any infrastructure between them?"
Priya frowned slightly. "We’ve looked at mesh networking."
"Look harder." Adrian pointed at the diagram. "If two Aether devices can communicate directly over a short range, and each device can relay that signal to the next device, what does that create?"
"A distributed network," she said slowly.
"Without towers. Without satellites. Without any external infrastructure at all." He looked at the team. "Think about what that means when conventional infrastructure fails. Flooding. Earthquake. Conflict. Remote communities where building permanent infrastructure isn’t commercially viable." He looked at the diagram. "The device shouldn’t just connect people to a network. In the right conditions, it should help become the network."
The room was quiet for a moment.
Priya looked at Chen. Then back at Adrian.
"That’s considerably harder than what we’re currently building."
"I know," Adrian said. "Start working on it."
At 12:40 his phone vibrated.
Daniel.
We recovered the damaged name.
He stopped walking in the corridor between Aether and the executive offices.
A second message.
It’s not Ashcroft.
Then a third.
It’s Sterling.
He stood still for a moment. Around him the corridor continued its ordinary business — someone passing with a folder, a door opening somewhere, the sound of a meeting in progress through the glass wall to his left.
He called Daniel.
"What did you find?"
"We reconstructed most of the first name from the damaged section of the document." A pause. "It’s William."
Adrian’s expression didn’t change externally.
His father’s name.
William Sterling.
"The document is dated 1978," Daniel continued. "Which puts your father at sixteen years old."
"So it’s not him."
"No. There’s another Sterling in the document."
"Who?"
"I don’t have it yet. But Adrian—" Daniel’s voice was careful in the way it got when he was presenting something he hadn’t fully processed himself. "There’s something wrong with the framing."
"What framing?"
"I’ll send you what we have. Call me when you’ve read it."
He put the phone in his pocket and kept walking.
He forced himself back into the work at 1:30.
He could have spent the afternoon in Daniel’s direction — calling back, pressing for details, trying to reconstruct whatever history R.A.C. had been building before he existed. The pieces were starting to accumulate in a way that suggested something significant was approaching legibility.
But the people who had designed that trail had placed it precisely where they knew he would find it, at a pace that kept his attention oriented toward the past rather than the future.
He went back to the engineering floor.
At 2:05 Miriam’s voice came through the internal system.
"We’ve got something."
He walked into the Helios laboratory.
The failed cell from the morning had been sectioned. Its internal structure filled the main monitor in cross-section, the electrode interface enlarged to the point where the individual material layers were visible.
"We were wrong about the cause," Miriam said without preamble.
"The electrolyte wasn’t failing?"
"The electrolyte was reacting to the manufacturing pressure. We assumed the interface needed to be flat — perfectly uniform contact across the full electrode surface. That assumption was wrong."
She changed the model on the display. A microscopic patterned surface appeared — not random texture but deliberate geometry, controlled variation across the contact area.
"Controlled micro-texturing. It increases the effective contact area without changing the physical dimensions of the cell. The interface doesn’t need to be flat. It needs to be correctly structured."
Adrian looked at the model.
"How much improvement?"
"Unknown yet. We’re running the first modified batch now."
"When do you know?"
"Four hours."
"I’ll be here."
At 4:30 the first modified batch results came in.
The modified interface had survived longer under load. Not dramatically — not a result that would appear in a press release. But the thermal instability had decreased. The energy transfer efficiency had improved. The failure point had moved further along the cycle.
The laboratory didn’t celebrate.
They documented everything, set up the next experiment, and kept working.
Adrian looked at the numbers.
The question had changed. It was no longer whether solid-state batteries could work at the chemistry level. It was whether this manufacturing technique could be scaled to commercial volumes. That was a different kind of problem — one that responded to capital, to time, to engineering iteration, to the kind of sustained industrial effort that he was now in a position to fund.
That mattered more than the result itself.
Robert came back at 5:12.
"The investors have moved again. One of them is now above five percent."
Adrian looked up.
"Which one?"
Robert named it. One of the three that had started buying before the Dominion acquisition had settled — the same institution that had appeared in Ashcroft’s folder, connected to the same settlement bank that had appeared across all three trades.
"They’re not passive," Adrian said.
"No. The analysts are starting to notice the portfolio structure. Someone published a note this afternoon calling it an integrated technology group." Robert placed the document on the desk. "They think you’re building something."
"For once," Adrian said, "the market is right."
At 6:03 the System appeared.
No reward. No mission.
A single notification.
STRATEGIC DEVELOPMENT UPDATE
Independent technological progress detected.
ORION: Prototype development active. AETHER: Architecture development active. HELIOS: Breakthrough pathway identified.
Portfolio value creation: ORGANIC
System reward dependency: DECREASING
Adrian read the final line twice.
Then set the phone face-down on the desk.
That was the most significant notification the System had produced in weeks. Not because it promised capital or strategic access or classified rewards.
Because it confirmed something he had needed to prove to himself.
He could build without it.
At 7:18 he walked through the nearly empty engineering floor.
The Orion chassis sat under the workshop lights. Aether’s team were still at their benches. Helios was running its third modified batch experiment of the day. The Atlas network simulation was cycling through scenarios on twelve terminals.
None of it was finished. None of it was anywhere near finished.
But Adrian could see for the first time the beginning of something that hadn’t existed when he took on Calder. Not a portfolio. Not a collection of acquired capabilities. An industrial ecosystem — companies that could design technology, develop it, manufacture it, deploy it, and support it, all within a structure he controlled and could direct.
His phone vibrated.
Daniel.
We found the full name. The Sterling in the 1978 document was your grandfather.
He stopped walking.
A second message arrived.
And Adrian — the document doesn’t say he was being observed.
He waited.
It says he was the observer.
He stood in the quiet of the engineering floor and read the message three times.
The hidden corridor. The Origin Records Division. R.A.C. and its predecessor. The Royal Archival Commission. The fifty-year trail that had been reconstructed piece by piece from documents that were never supposed to be found.
His grandfather.
Not a subject.
An observer.
He put the phone in his pocket and looked at the Orion chassis under the lights.
Tomorrow he had a government meeting. A battery breakthrough to develop. An automotive prototype to refine. A communications architecture to redesign.
And now a family history that had apparently been hidden in plain sight for fifty years, waiting for someone in his family to look in exactly the right direction.
He turned off the office light.
The work would continue without him overnight.
That, more than any System reward, was what actually made the difference.
He walked toward the exit.
Behind him, the engineering floor kept running.
End of Chapter 61