Discussions of technology and the environment gravitate toward carbon, because carbon is the metric that reporting frameworks require and the one most people have a rough intuition for. It is a reasonable proxy for a great deal, and it misses two impacts that are arguably more locally severe: what device manufacturing does to water, and what it does to land.
Neither of these shows up in a device’s specifications, and neither is felt by the person using it. They are borne entirely by the regions where extraction and fabrication occur, which are usually far from where the finished product is sold.
Considering the e-waste environmental impact through this lens changes the emphasis. The material inside a discarded device does not merely represent energy already spent. It represents watersheds altered, aquifers drawn down, and landscapes that will not recover within any timeframe that matters to the people living near them.
Water as the Constraint Nobody Sees
Semiconductor fabrication is one of the most water-intensive manufacturing processes in existence. Chips are washed repeatedly during production with ultrapure water, which itself requires substantial processing to produce, and a single large fabrication plant can consume millions of litres per day.
That consumption is concentrated. Fabrication capacity clusters in a small number of regions, several of which face genuine water stress, and competition between industrial and agricultural use during drought periods has become a recurring political issue in more than one of them.
Mining adds its own demand. Ore processing, dust suppression, and tailings management all consume water at scale, and copper and lithium extraction in particular are located frequently in arid regions where the water has alternative uses.
The quality dimension matters as much as the quantity. Acid mine drainage from sulphide ore bodies acidifies watercourses and mobilizes heavy metals, and it continues for decades or centuries after a mine closes. It is among the most persistent forms of industrial pollution there is.
What Extraction Does to Landscape
The land impact is easier to picture and just as durable.
Open pit mining removes overburden across large areas to reach ore bodies that may represent a fraction of a percent of the extracted mass. The ratio of moved rock to recovered metal in copper and gold mining is extreme, and the moved rock does not go back.
Tailings, meaning the processed waste left after metal extraction, are stored in impoundments that must be maintained indefinitely. Tailings dam failures are rare but catastrophic when they occur, releasing enormous volumes of slurry into river systems.
Rare earth processing generates radioactive residues as a routine byproduct, because the ore bodies contain thorium and uranium alongside the target elements. Managing those residues is a permanent obligation.
Land disturbance also affects the people already there. Extraction frequently occurs on or near land used by communities with limited ability to negotiate terms, and the resulting displacement and loss of use is a cost that appears in no product footprint.
Why Recovery Changes This Arithmetic
The argument for recovering material from retired equipment is usually framed in energy terms, and the energy case is strong. The water and land case is arguably stronger, because those impacts are local, concentrated, and largely irreversible.
Metal recovered from a circuit board requires no new pit, no new tailings impoundment, and no new water allocation in a stressed basin. It arrives at a refinery already concentrated, at grades that natural ore bodies cannot approach, and the processing required to refine it is a fraction of what primary production demands.
The comparison for gold is the most dramatic. Primary gold production moves enormous quantities of rock and uses cyanide leaching to recover grams per tonne. Recovery from electronics works at concentrations orders of magnitude higher in a contained industrial setting, with no watershed exposed to drainage in perpetuity.
Copper follows the same logic at greater volume. Every tonne recovered is a tonne that does not require the moved rock, the drainage, or the water allocation.
Where This Leaves Individual Decisions
The honest position is that no individual disposal decision measurably affects a watershed in Chile or a fabrication plant’s water draw in Taiwan. That is not the useful frame.
The useful frame is that these impacts are the reason device longevity matters more than any other single choice. A device kept in service for six years instead of three halves the extraction and fabrication burden attributable to it, and that is a larger effect than anything that happens at the disposal stage.
Reuse comes next, because passing working equipment on delivers the same deferral to someone else’s replacement purchase.
Recovery comes third, and it matters because the alternative, meaning burial, guarantees that the extraction cycle runs again for the replacement material.
The pattern across all three is the same. The environmental cost of a device is overwhelmingly front-loaded, incurred in places the buyer will never see, and the only way to reduce it is to require fewer devices to be made. Everything else is secondary, though the secondary part is still worth doing properly.
