Apple is increasingly utilizing defective chips to produce budget-friendly laptops. Although this may sound counterintuitive, it highlights a widely accepted technique known as “binning,” which minimizes production costs and environmental waste in smartphones and laptops.

The term “binning” originates from agriculture, where high-quality fruits and vegetables are sold to premium markets, while those in poor condition are allocated to other customers, and the least desirable items may be recycled as animal feed. This segregation ensures minimal waste, a practice echoed in semiconductor manufacturing.

For instance, take the new MacBook Neo, which incorporates the A18 Pro system featuring five GPU cores, offering users a more cost-effective Apple laptop option. Previously, the A18 Pro was found in the iPhone 16 Pro with six functional GPU cores. Reports suggest that Apple is utilizing A18 Pro chips stored in a bin with one defective core, thereby optimizing the use of discarded components. Although Apple hasn’t commented on this, industry experts like New Scientist indicate that manufacturers across various sectors, from electronics to automotive, routinely adopt this practice.

According to Owen Guy, a researcher from Swansea University in the UK, semiconductor chips are produced in large quantities on 300-mm silicon wafers housing trillions of individual transistors. Sophisticated machinery performs countless operations on these wafers, generating layers of circuits, insulators, and chemicals just nanometers thick. In truth, the complexity of this procedure often makes it astonishing that chips work at all, rather than the occurrence of defects.

“At each stage of the process, there is a small chance that something may go awry,” says Guy.

The proportion of errors on a single wafer dictates the yield, or the quantity of chips that meet the required specifications. Yields can reach up to 99 percent for standard silicon chips, a material employed since the 1960s, but often improve for advanced chip designs and newer substrate materials like silicon carbide and gallium nitride.

“The critical question is the number and severity of defects. Unless there’s a so-called fatal flaw, functioning chips can still exist even when they have imperfections,” Guy explains.

Imagine achieving a yield of 90%, where 9 out of 10 chips perform as expected. In such a scenario, one chip is destined for the bin. If one core is defective, it may be classified as a different product, featuring five cores instead of six. Alternatively, it might be regulated to operate at lower voltages or frequencies or rated for higher power consumption or temperature. There will always be customers eager to use these chips.

Researcher Tony Kenyon from University College London states that users often remain unaware of any issues. Error-correction software is employed to isolate faulty transistors in memory chips, avoiding potential data loss or rerouting computations around flawed processor cores to prevent software crashes.

“A deeper look under the hood reveals that certain components of the chip may not function. This is widely prevalent. Many believe all chips are uniform, but the reality is far more complex,” Kenyon asserts.