Most costs in a marketplace get cheaper per unit as volume grows. Authentication is the exception, and the exception shapes the whole industry.

Software scales because copying is free. Warehouses scale because bigger buildings cost less per shelf. Authentication resists both, because at its core it is expert judgment applied to a physical object, one object at a time. The expert who can distinguish a genuine stitch pattern from a good counterfeit took years to train, commands a real salary, and can only look at so many items per day. Doubling volume means doubling experts, and the supply of people who can reliably authenticate a rare handbag is not deep.

It gets worse, because the adversary improves. Counterfeiting is a business with margins that fund its own research. Every tell that authenticators rely on eventually gets studied and replicated, which forces authentication deeper into materials, hardware, stitching density, and provenance checks that take more time per item, not less. The cost curve bends the wrong way: as the category grows and attracts better fakes, the marginal item becomes more expensive to clear, not cheaper.

Platforms respond in three ways, and you can read any platform's strategy by which one it leans on. The first is tiering: full inspection for high-value items, lighter checks or spot checks below a price line, which manages cost but concentrates risk exactly where volume lives. The second is technology: imaging, machine learning on known-fake databases, hardware verification. This genuinely helps at the pattern level, but it assists experts more than it replaces them, and its training data comes from the expensive human process it hopes to reduce. The third is pushing the cost onto someone else: buyer protection programs, third party authentication services, or simply tolerating disputes as a cost line.

Here is the reframe that matters for anyone watching this industry: authentication looks like a cost center and behaves like an asset. Every authenticated item teaches the platform what real items look like, what they sell for, how condition maps to price, and how fakes evolve. That accumulated judgment, whether it lives in experts or models, is proprietary and compounds. The platforms that treat authentication as a tax to minimize will spend the next decade buying that judgment from the platforms that treated it as an investment.