US Crypto Perps Arrive, Solana Rethinks Fees

The CFTC approved a bitcoin perpetual futures contract submitted by Kalshi, with the application filed on the 28th and approval granted on the 29th, a turnaround suggesting the outcome was coordinated in advance. CFTC official Mike Selig has publicly backed onshoring perps, with the regulatory rationale being that perp trading happens regardless and bringing it onshore generates tax revenue and reduces black market activity. Coinbase separately received a no-action letter allowing it to offer perps and derivatives through its Deribit acquisition, and the practical effect is that any centralized crypto platform in the US can now effectively list perps.

Coinbase and Robinhood already offer perps internationally and could activate them for US users relatively quickly, representing a meaningful revenue bump given that perps is a highly profitable product category. The CFTC specified a case-by-case review for listings beyond bitcoin, but a rapid follow-on wave covering ETH, SOL, and other top assets is plausible given the fast approval turnaround. Approximately 80 percent of perp trading volume comes from the top three or four assets, so initial listings would capture most of the product's value. Kalshi is unlikely to be a primary beneficiary because it lacks a meaningful edge in capturing perp flows, and platforms where users already trade spot and perps are better positioned.

The central competitive question is whether US incumbents like Coinbase benefit most or whether offshore-perceived platforms like Hyperliquid can now compete on equal regulatory footing. Hyperliquid's Washington lobbying efforts through HPC appear aimed at exactly that outcome. Whether decentralized non-KYC platforms like Hyperliquid can serve US customers in a compliant way remains unresolved. The preferred structural path discussed is KYC enforcement at the frontend layer while the base layer stays permissionless, preserving unified liquidity through builder codes and compliant frontends. Creating a separate US liquidity pool would replicate the Binance International versus Binance US problem, where the US version suffered from far inferior liquidity.

On Solana, a proposal called SIMD 547 was submitted to GitHub by a contributor known as caveman lover boy, written in roughly 20 minutes on a Saturday and not yet subject to significant community discussion. The proposal introduces an additional base fee that adjusts based on resource consumption within blocks, rising during high activity and falling during low activity, charged at 0.1 lamports per requested compute unit against a current block maximum of 60 million compute units. Example fee additions shown in the proposal range from plus 31000 to plus 82000 lamports over the existing 5000 lamport base fee, representing roughly an order of magnitude increase in total base fees paid, though absolute costs remain very low.

Compute-unit-efficient transactions like oracle updates would see minimal cost increases of around 2 percent, while retail users on platforms like pump.fun could see around 60 percent increases on top of existing fees, which remains negligible in absolute terms since most transactions cost less than one cent. Programmatic users such as oracles, DeFi protocols, and routing contracts processing millions of transactions would feel a 600 percent fee increase meaningfully, and Pyth oracle teams have already been working to reduce their compute unit requirements to manage operating margins. Because the fee charges based on requested rather than actual compute units, and current blocks show consumption sometimes only one quarter of what was requested, sophisticated users would have strong incentive to optimize their CU requests.

SIMD 96 was activated in February 2025 and removed the burn on priority fees, which had previously burned roughly 10000 SOL per day in Q4 2024 and January 2025, effectively taking the burn to zero. The new resource-based fee proposal could restore a burn of approximately 2000 to 5000 SOL per day, attacking Solana emissions from the demand side through increased burn tied to network usage rather than reducing validator issuance, leaving validators unaffected. Current annualized SOL inflation is approximately 3.8 percent declining at 15 percent per year, and if the burn reduces net emissions by roughly 2 to 2.5 percent, effective net inflation would be meaningfully lower. The proposal is designed to activate only after the Alpenglow upgrade because before that it would significantly increase validator voting costs and harm smaller validators. A dissenting view from a Solana DAT team member holds that increasing transaction costs moves in the wrong direction for a network that should be getting cheaper to support an on-chain agentic and AI future requiring massive transaction scale.