ASIC Resistance

"ASIC resistance" is the (mostly failed) effort by some cryptocurrencies to design proof-of-work algorithms that don't benefit from specialized hardware. The idea is that if everyone mines with general-purpose CPUs or GPUs, mining stays decentralized among hobbyists rather than concentrated in industrial farms.

The history of ASIC resistance is essentially "every attempt eventually gets ASICs anyway":

• Scrypt (Litecoin, 2011). Designed to require more memory than CPUs/GPUs could spare. ASICs arrived in 2014.
• Ethash (Ethereum, 2015). Memory-hard, with growing DAG. ASICs arrived by 2018.
• RandomX (Monero, 2019). Heavy use of random code execution to discourage hardware optimization. Holding up better than predecessors, but pressure exists.
• ProgPoW (proposed but never deployed). Would have changed the algorithm regularly to disrupt ASIC manufacturers' multi-year amortization.

Bitcoin explicitly doesn't pursue ASIC resistance. The reasoning:

• ASIC resistance is temporary at best. Any sufficiently profitable algorithm attracts specialized hardware eventually. The result is just a churn of mining hardware obsolescence.
• Specialization is actually a security feature. Bitcoin ASICs are useless except for SHA-256 mining. That means the global mining industry has sunk billions into hardware that can only be used to secure Bitcoin. Repurposing that hardware to attack the network would destroy its own value. This is a strong commitment device.
• Stable algorithms enable stable mining infrastructure. Predictable hardware lifecycles let miners build long-term operations.

Bitcoin's bet: better to accept specialization, build a deep ASIC ecosystem, and rely on competition between manufacturers and miners to maintain healthy decentralization. The bet has worked imperfectly - manufacturer concentration is real - but the alternative of perpetual algorithm churn has worse failure modes.

See ASIC for the hardware itself and Mining Centralization for the real concerns about industry concentration.