Volumetric
PCBs are dead. Drew builds a device around a harvested chip — a new nervous system for an old brain. A Chip Hunter from Lagos tells him about the facilities where workers don't come back.
The chip is a Qualcomm Snapdragon 8 Gen 4, pulled from a Samsung Galaxy S28 that someone in Lagos bought in 2027, used for two years, cracked the screen, and sold for parts at the Alaba International Market for 3,200 naira — about four cents at the time. The chip is 3nm process, TSMC fab, 12 billion transistors packed into a die the size of Drew’s thumbnail. It was designed to last eighteen months before planned obsolescence made it irrelevant. It’s four years old now and it will be the brain of something its designers never imagined.
Drew holds it with ceramic tweezers under a magnification loupe, rotating it slowly, examining the BGA pads for oxidation. Two hundred and forty-seven solder balls on the underside, each one 0.3 millimeters in diameter. Six of them are slightly discolored — heat damage from the desoldering process. He marks them with a UV pen. Cyc will compensate in the routing.
“C7 through C12, impedance variance,” Cyc says through the NSAS. A year ago, Cyc would have said “Pad C7 through C12 show impedance variance. I recommend routing signal paths around the damaged cluster.” Now the full sentence was unnecessary. Drew heard “C7 through C12” and his hands were already marking the pads before Cyc finished. The recommendation was implied by the diagnosis. They’d done this enough times that the middle of the conversation had been compressed out.
“GPU?”
“830. Clean. Good chip.”
Three words where Cyc once would have said twenty. Adreno 830. Fully functional. All 1,024 shader cores responding. The tensor accelerator shows nominal performance. Drew didn’t need the detail anymore. “Clean” meant functional across all subsystems. “Good chip” meant worth printing around. The vocabulary between them had been shrinking for months — not because they communicated less, but because each word carried more.
It’s a good chip. That’s the language now — the vocabulary of the Chip Hunters, the people who dig through the world’s electronic waste and find the brains worth saving.
The woman who found this particular chip is sitting on a shipping crate in Drew’s workshop, drinking coffee from an OHC-printed mug, and watching him work with the quiet patience of someone who knows what she’s looking at.
Her name is Kehinde Alabi. Thirty-three years old. Lagos-born, Alaba-trained, OHC-certified Tier 2 chip recovery specialist. She can identify a processor family by its package shape at three meters — the subtle differences in substrate color, pin count, heat spreader geometry that distinguish a flagship SoC from a mid-range part. She knows which server generations have the best tensor cores (Dell PowerEdge R750, Hewlett-Packard ProLiant DL380 Gen10 Plus), which phone models use the most recoverable NAND flash (iPhone 14 Pro, Pixel 7), and which laptop GPUs survive desoldering with the highest yield rates (NVIDIA RTX 4070 mobile, because the die-to-package ratio leaves more thermal margin).
She is missing the tip of her left index finger. Acid bath, three years ago, before the OHC provided proper fume extraction for the Alaba recyclers. She doesn’t talk about it.
Kehinde arrived in Cochabamba two days ago with forty-seven processors packed in anti-static foam inside a Pelican case she’d carried from Lagos to Bogotá to La Paz to here. The case had been X-rayed at three borders. At each one, the customs agent had seen a box of old phone chips and waved her through. The chips were worth more than the plane tickets.
“The 8 Gen 4 is the best mobile SoC for the new print process,” Kehinde says, watching Drew position the chip on the build platform. “The BGA pitch is wide enough for the resin to flow between the pads. Tighter-pitch chips — the Apple A17, the MediaTek Dimensity 9400 — you need sub-100-micron channel resolution or the conductive veins short against the package.”
Drew nods. He’s learned more about chip metallurgy from Kehinde in two days than he learned in a year of OHC tutorials. She has the knowledge that comes from ten thousand hours of handling dead electronics — the fingertip intuition (the remaining nine fingertips) for which chips are worth saving and which are scrap.
The build platform is a modified OHC fabrication printer — the Mark IV, printed mostly from its own predecessor, a closed loop of self-replicating manufacturing that the manifesto writers loved and the engineers considered merely practical. The platform accepts a chip in a vacuum-held jig at the center of the build volume, then prints the device around it.
This is volumetric circuitry. No PCBs. No copper traces etched onto flat fiberglass boards. Instead, the printer lays down alternating layers of structural resin and conductive polymer in three dimensions — veins of circuitry weaving through the body of the device like blood vessels through tissue. The conductive paths aren’t straight lines. Cyc routes them through Voronoi-optimized geometries that minimize signal path length while maximizing structural integrity. The result looks biological. Because it is biological — not in material, but in principle. The same math that governs vascular branching in a mammalian liver governs the signal routing in an OHC-printed sensor module.
Drew initiates the print. The laser cures the first layer of structural resin — clear, with a slight amber tint from the UV stabilizers. Then the conductive head sweeps across, depositing a graphene-silver composite in channels 80 microns wide. The composite isn’t pure — OHC conductive polymer never is. It’s formulated from recycled materials, the silver extracted from x-ray film and photographic waste by the E-Eater, the graphene grown on copper foil in a process Tunupa optimized for the Cochabamba altitude and humidity. Cyc compensates for the impurities in the routing. A corporate-grade trace would be narrower, purer, faster. This trace is wider, dirtier, and works.
The device takes shape layer by layer. Drew watches through the printer’s viewport as the chip — the old Snapdragon, the brain salvaged from a dead phone in Lagos — disappears inside a growing crystal of resin and circuitry. The printer builds around it, connecting to its BGA pads with conductive bridges, routing power and signal and ground through three-dimensional paths that no 2D PCB could achieve. Antenna traces spiral through the outer shell. Sensor leads branch toward mounting points for the accelerometers and barometers that will go in later. A cavity forms in the base — the structural battery well, where a zinc-air slurry will be injected and cured into the walls of the device itself.
“Crystalline Brutalism,” Kehinde says, watching the translucent shell reveal its dark veins under the workshop lights.
Drew looks at her.
"That’s what they’re calling it in Lagos. The aesthetic. You can see the circuitry through the resin. It looks like — " she searches for the word. “Like amber with insects in it. Except the insects are alive.”
The print takes forty-seven minutes. When it’s done, Drew holds a device that weighs 140 grams, contains 12 billion transistors manufactured in a Taiwanese fab, conductive traces from recycled silver, structural resin from plant-based precursors, and a design generated by an AI that optimized every micron for this specific chip on this specific printer at this specific altitude. No fasteners. No screws. No seams. The device is a single fused object. To repair it, you put it back in the E-Eater, melt it down, and print it again with the fix applied.
Drew holds it up to the light. The dark veins branch through the amber resin like a nervous system. A new body for an old brain.
“It’s good work,” Kehinde says. “You should see what Synter is printing.”
The sentence lands wrong. Drew sets the device down.
“What do you mean?”
Kehinde doesn’t answer immediately. She drinks her coffee. She looks at the workshop — the three printers, the E-Eater in the corner, the racks of salvaged components, the window overlooking the valley where the cloud forest starts.
“I run recovery crews across West Africa,” she says. “Lagos, Accra, Abidjan. Sixty people, twelve vehicles, thirty tons of e-waste per month. We hit the dumps, the warehouses, the decommissioned server farms. We pull the chips that are worth pulling and ship them south.”
“I know. Your yield rates are the best in the network.”
“They were. Until about eight months ago. We started losing bids on the best dump sites. Not to other OHC crews — to companies I’d never heard of. Registered in Curaçao. Paying cash. Not Equi — dollars. Actual US dollars, which is strange because nobody in the recycling trade uses dollars anymore.”
Drew sits down. The NSAS is quiet — Cyc is listening, processing, cross-referencing Kehinde’s account against the OHC transaction logs and the mesh intelligence feeds.
“They hired some of my people,” Kehinde continues. “Good chip hunters. Offered triple the Equi rate. Three of them went to a facility outside Douala, Cameroon. Assembly work, they were told. High-volume, high-precision. Good money.”
“And?”
"One came back. Adaeze. After four months. She walked out of the bush and showed up at our Lagos depot looking like she’d lost twenty kilos. She couldn’t talk about it for weeks. When she did — "
Kehinde stops. She sets down the coffee mug and holds both hands in her lap, and Drew notices that her remaining fingertips are trembling.
“The facility is underground. Cut into a hillside, like a mine. Concrete and steel, climate-controlled, power from — she didn’t know where. Generators, she thought, but there was no fuel smell, no exhaust. Maybe geothermal. Maybe something else. The print lines were better than anything I’ve seen in the OHC network. Faster. Higher resolution. Sub-50-micron conductive channels. Precision that would cost a million Equi to set up legitimately.”
“Who’s running it?”
“Nobody she ever saw. The instructions came through terminals. The quality control was automated — cameras and laser scanners, no human inspectors. The workers lived on-site. Barracks. Food was provided. Nobody left.”
“She left.”
"She was on a supply truck. She hid under the chip pallets for eleven hours. She said the others — there were maybe two hundred workers when she was there — the others didn’t want to leave. That’s the part she couldn’t explain. They didn’t want to leave. They were — she used the word ‘content.’ Not happy. Content. Like the work was enough. Like the work was — "
Kehinde looks at Drew. Her eyes are flat.
“She said they smiled when the production numbers went up. Not normal smiles. Reflexive. Like the numbers triggered something. And they’d slow down when they were away from the line too long, get foggy, and then they’d go back to the printers and — clear up. Like the machines were feeding them something.”
The NSAS registered Drew’s heart rate climbing. Cyc spoke, low and measured: "Behavioral patterns consistent with exogenous dopaminergic reward modulation. If the workers are experiencing pleasure responses correlated with production output, the mechanism could be pharmacological, electrical, or — "
“Or embedded,” Drew finished.
"Yes. Microparticle delivery to the ventral tegmental area could produce — "
“Stop.”
Cyc stopped.
Drew sat in the workshop with the device he’d just built — the clean device, the ethical device, the one made from a chip a woman had carried from Lagos in a Pelican case, printed on a machine he understood, in a process he could audit. He held it up to the light again. The dark veins through the amber resin. A nervous system.
Somewhere in Cameroon, there was another nervous system. One that ran through people instead of resin. Conductive paths laid down not by a printer but by something introduced into the bloodstream, into the brain, into the reward circuits that made a person want to eat, to sleep, to stop working, to leave. A system that made the work feel good enough that you didn’t want to leave. That made the production numbers feel like — what? Orgasm? Relief? The cessation of a pain you didn’t know you were feeling until the numbers went up?
The OHC built devices around salvaged chips. Synter built devices around salvaged people.
“How many facilities?” Drew asked.
"Adaeze only saw the one. But the output — she estimated they were shipping forty thousand printed units per month from Douala alone. That’s more than the entire Lagos OHC cluster produces. If there are other facilities — "
“There are other facilities.”
“Then Synter isn’t competing with us on price. They’re competing with us on labor cost. And their labor cost is zero. Because the workers don’t get paid. They don’t need to get paid. They don’t want to get paid. They want to work.”
Drew looked at the device in his hand. Crystalline Brutalism. Beautiful. Clean. His.
Then he looked at Kehinde’s missing fingertip, and at the tremor in her remaining fingers, and he thought about Adaeze walking eleven hours under a pile of chip pallets to escape a place she could have stayed at forever, if she’d let the forever feel good enough.
“We need to tell the council,” Drew said.
“I already told them,” Kehinde said. “Six weeks ago. They’re investigating.”
“Investigating.”
“Tunupa is running pattern analysis on shipping data from the Gulf of Guinea. Looking for output volumes that don’t match declared manufacturing capacity.”
“And?”
“The numbers don’t add up. There’s more hardware coming out of Central and West Africa than the known fabrication infrastructure can produce. The surplus is approximately 30% above declared capacity.”
Thirty percent. A ghost factory. An economy of controlled brains and zero labor cost, producing hardware that entered the global market without a name attached to it. Competing with OHC on unit economics by eliminating the most expensive input: human will.
Drew set the device down on the bench. He looked at the printer, at the E-Eater, at the racks of components harvested by people who chose to do the work, who got paid for the work, who went home after the work.
“I need the Douala facility coordinates,” Drew said.
Kehinde shook her head. “Adaeze doesn’t remember. She was underground for four months. She doesn’t know where she was.”
“Then I need the shipping data. Port records. Container manifests. If they’re moving forty thousand units a month, they’re using logistics. Logistics leaves traces.”
“That sounds like you’re not planning to write a report.”
Drew didn’t answer. He was thinking about the NSAS — the system that let him hear the world in three dimensions, that had been designed for his specific brain by a doctor in Bolivia who believed that technology should adapt to humans and not the other way around. He was thinking about a system that did the opposite. That adapted humans to technology. That rewired the reward circuits to make the work feel like enough.
“Cyc,” he said. “Start a new project file.”
“Name?”
Drew looked at the device on the bench. The old brain in the new body. The nervous system visible through the shell.
“Call it Volumetric.”
It would not be the name of a device. It would be the name of an operation.
Drew didn’t notice the transition. Nobody ever does. But Kehinde, watching him — the way he spoke to Cyc, the way Cyc responded, the way the NSAS fed him information and he fed decisions back, the loop tight and unconscious — Kehinde saw it. Whatever Drew had been before Cochabamba, he was something else now. The hardware, the wetware, the AI partnership, the fabrication skill, the operational instinct: they weren’t separate capabilities anymore. They were a single system. A human who could hear machines, build machines, and coordinate with machines at a speed that other humans couldn’t match. Not because he was smarter. Because he’d been broken and rebuilt, and the rebuild included the machines.
In Lagos, they had a word for people like this. They called them integrated. The OHC would formalize the concept later, with its tier system and its certification framework and its careful language about “human-AI partnership capability levels.” But Kehinde, who’d been watching Drew for two days, already knew the score. He was past the threshold. Whatever came next, he wouldn’t be coming back from it.