One Production Run, Three Price Points, One Piece of Silicon
Chips come off a wafer with varying quality, and manufacturers test each one and sort it into performance grades. That sorting turns an unavoidable manufacturing variance into a product line and a pricing strategy.
Identical Designs, Different Outcomes
Semiconductor manufacturing patterns hundreds of identical circuits onto a silicon wafer through dozens of process steps at atomic scale. Despite extraordinary process control, the results are not identical.
Contamination, lithographic variation, and thermal differences across the wafer mean individual dies differ in how fast they can switch, how much power they draw, and whether every functional block works at all. Dies near the wafer edge typically perform worse than those near the centre.
Every die is tested. The results determine what the die becomes.
Binning Turns Variance Into Product
Binning is the practice of sorting tested dies into grades and selling each grade as a different product. Two forms operate.
Speed binning sorts by achievable clock frequency at a given power and voltage. The fastest dies are sold as the premium part, slower ones as lower tier parts in the same family.
Functional binning disables defective blocks. A processor designed with eight cores, of which one fails testing, can have two cores disabled and be sold as a six core product. A graphics processor with defective compute units can be sold with fewer units enabled.
| Test Result | Without Binning | With Binning |
|---|---|---|
| Meets top specification | Premium product | Premium product |
| Slightly slower | Scrap | Mid tier product |
| One block defective | Scrap | Reduced configuration product |
Binning converts a yield problem into a product portfolio. Without it, a die that missed the specification is waste. With it, the same die is a sale at a lower price, which turns the entire distribution of manufacturing outcomes into revenue.
Why the Economics Are So Favourable
Semiconductor fabrication has enormous fixed costs and very low marginal cost per die. A leading edge fabrication facility costs billions and the incremental cost of processing one more wafer is small relative to that.
Under that structure, any die sold at any price above the marginal cost of packaging and testing contributes to fixed cost recovery. Scrapping a die that could be sold at half price destroys real contribution.
It also stabilises supply. Yield is uncertain and varies with process maturity, so a manufacturer that could only sell top grade parts would face wildly variable output. Binning lets total output remain stable while the mix across grades varies, which is a far more manageable business.
The Practice That Confuses People
The counterintuitive part is that binning is not purely reactive. Manufacturers frequently sell fully functional dies as lower tier products, deliberately disabling capability that works.
The reason is demand mix. If yields improve and most dies now meet the top specification, but customers want a range of price points, the manufacturer would rather sell a fully functional die as a mid tier part than not sell it at all. Producing a separate physical design for the lower tier would require a second mask set and a second production line, which is far more expensive than disabling a block.
This is versioning, a standard price discrimination technique, and semiconductors are the cleanest large scale example of it. The cheaper product costs the same to make as the expensive one and is deliberately degraded so that customers willing to pay more will do so.
The classic historical illustration outside chips is the same printer sold at two prices with the cheaper model containing a chip that slows it down. Both models cost the same to manufacture and the slower one costs more to design.
Unlocking and Why It Stopped
Because disabled capability physically exists, enthusiasts historically found ways to re enable it, most famously unlocking additional cores on certain processors. Manufacturers responded by making the disabling permanent, through fuses blown at test that cannot be reversed.
That response is economically rational and worth noting for what it reveals: the manufacturer is spending effort to prevent customers from obtaining capability that costs nothing to provide, because permitting it would collapse the price discrimination the product line depends on.
What It Means for Analysis
Several implications follow for anyone reading a semiconductor company.
Gross margin depends heavily on bin mix, meaning the share of output meeting the top specification, which improves as a process matures. A company launching a new node has poor mix and improving mix over the product life, which produces margin expansion unrelated to pricing.
Yield commentary should be read carefully, because a company can have low yield of top bin parts and high total usable yield, and those are very different situations.
And supply constraints on a specific high end product frequently reflect bin availability rather than total capacity. A shortage of the premium part alongside ample supply of the mid tier part is the signature of a binning constraint, not a fabrication one.
The Bottom Line
Binning is the semiconductor industry answer to the fact that identical manufacturing produces non identical results, and it converts an unavoidable variance into a product ladder. It is genuine cost recovery when the lower tier part is a die that failed the top specification, and it is deliberate price discrimination when the part was fully functional and disabled anyway. Both happen simultaneously in the same product line, which is why the cheapest and most expensive versions of a chip are so frequently the same piece of silicon.