NEAR’s Polosukhin sees shrinking need for centralized exchanges as onchain tools expand

No. Centralized exchanges still process ~75% of crypto volume as of October 2026. Polosukhin's comment targets specific use cases—staking and low-fee swaps—where onchain tools excel. Market structure data shows CEXes maintain dominance in order-book depth, derivatives, and price discovery. Both models coexist and will likely continue operating in parallel for the foreseeable future.

Native DEXs, yield protocols, and self-custody wallets with integrated trading. Examples include Jupiter (Solana ecosystem) and similar NEAR-based venues. These platforms enable direct token swaps without intermediaries. Growth metrics show 30-40% YoY expansion in onchain trading volumes, though absolute CEX volumes remain significantly larger. Functionality improvements have reduced friction for self-directed traders.

DEX volumes have grown substantially, but CEXes retain structural advantages. Order-book depth, liquidity aggregation, and regulatory compliance frameworks favor centralized venues for large orders and derivatives trading. October 2026 data indicates onchain spot trading represents ~20-25% of total crypto transaction volume, with CEXes maintaining the majority. This reflects segmentation rather than replacement.

Traders gain optionality. Decentralized tools offer lower fees, non-custodial settlement, and censorship resistance for simple swaps. CEXes provide superior liquidity, lower slippage, derivatives access, and fiat on/off-ramps. Sophisticated traders typically use both—onchain for specific needs, CEXes for volume trades or leveraged positions. Market fragmentation continues, not consolidation toward either model exclusively.

Onchain trading carries smart contract risk, front-running exposure, and lower liquidity for large orders, resulting in higher slippage. Self-custody requires user diligence; mistakes are irreversible. CEXes offer custodial insurance and regulatory protection but introduce counterparty risk and regulatory uncertainty. Each model presents distinct tradeoffs. Users should evaluate their specific requirements and risk tolerance independently.