The Gold Inside Your Phone

The gold market is usually described in two halves: jewellery on one side, investment and official reserves on the other. Between them sits a third category that rarely gets a paragraph, because it involves quantities too small to photograph and applications too dull to romanticise. Technology consumes a few hundred tonnes of gold a year in the form of films a fraction of a micron thick and wires thinner than a hair, and it does so for reasons that have nothing to do with what gold looks like.

Engineers do not specify gold because it is precious. They specify it because of a narrow and unglamorous property: it does not build up an insulating layer. Everything else follows from that.

The oxide problem

Copper is a superior conductor to gold and costs a tiny fraction as much. In a sealed, soldered joint it is the obvious choice and it is used everywhere. The difficulty appears at any surface exposed to air, particularly one that is repeatedly separated and remated — a connector, a card edge, a switch contact, a relay.

Copper oxidises. Silver, the best conductor of all, forms sulphide films in ordinary indoor air. Both films are electrically resistive, both grow with time and temperature, and both concentrate at exactly the microscopic high spots where a contact actually touches. The result is a connection whose resistance drifts upward over years — the intermittent fault that appears on a cold morning and disappears when someone wiggles the plug. Gold does not form these films at all. A gold-plated contact behaves in year fifteen approximately as it did on the assembly line.

Where the metal actually sits

Open a phone and the gold is distributed across a handful of unremarkable places, none of them visible without magnification.

• Connector and contact plating — board-to-board connectors, SIM and card contacts, test pads and antenna springs, typically a hard gold layer under a micron thick over a nickel barrier.
• PCB surface finish — electroless nickel immersion gold, a very thin gold film that keeps copper pads solderable and flat during storage and assembly.
• Die interconnect — the wires or bumps joining a silicon chip to its package, historically gold, now substantially copper in high-volume parts.
• Specialist and legacy hardware — aerospace, medical implants, defence and long-life industrial equipment, where qualification cost dwarfs metal cost and nobody changes a proven material.

The substitution that already happened

The most consequential change in industrial gold demand this century was not a price spike but a packaging decision. Bonding wire — the fine strand connecting a chip's pads to its lead frame — was gold by default for decades, and semiconductor assembly consumed the metal by the tonne. Over the 2000s and 2010s the industry converted the bulk of high-volume packaging to copper wire, driven by cost and helped by genuine improvements in bonder control and mould compound chemistry.

That transition removed a large block of demand permanently and is not coming back. What it also demonstrated is the shape of substitution risk in this market: it is slow, it is capital-intensive, it requires requalification of every affected part, and once it happens it is irreversible. Nobody converts back to gold because the price fell.

What remains is the residue that resists substitution — surfaces that must survive the field, connections that must be separable, and applications where a failure costs more than the entire metal budget of a factory. That residue is small, but it is unusually stable.

Concentration, and what that means for recycling

The comparison that makes waste-stream recovery interesting is with the mine. Typical open-pit ore runs around one gram of gold per tonne. Populated circuit boards from consumer equipment commonly assay in the hundreds of grams per tonne, and dense telecom and computing hardware can run higher still. In grade terms, a container of scrap boards is a spectacular orebody that arrives pre-crushed and needs no drilling.

The catch is that the gold arrives bonded to twenty other elements — copper, tin, lead, palladium, brominated plastics — in an assembly designed to be permanent. Recovering it properly means integrated smelting and refining, off-gas treatment and a permit regime, which is why formal e-waste refining is concentrated in a small number of very large facilities. Where that infrastructure is absent, the same material is processed with open acid baths and burning, which recovers a fraction of the metal and poisons the people doing it.

Industrial gold, then, is a small market that behaves nothing like the rest of the metal. It does not care about the price. It does not respond to sentiment. It shrinks when an engineer finds a cheaper material that passes qualification, and it holds when nobody does. It is the one part of the gold market where demand is decided by a reliability specification rather than by a view about the world.