cointelegraph.com
Cryptocurrency traders increasingly leaned on automated strategies in 2025 as volatile but largely range-bound markets made directional bets harder to sustain, according to a year-end recap from HTX.
The Seychelles-based exchange, formerly known as Huobi, said the trend was most visible in the growing use of grid-based trading bots on its spot platform. According to HTX, grid trading volume rose 97% year over year in 2025, while capital allocated to grid strategies doubled.
The increase was especially pronounced in stablecoin pairs, where grid trading volume rose 352% year over year, compared with 122% growth in major cryptocurrencies. HTX said the bots were typically used to capture smaller, repeated price swings rather than to bet on sustained market moves.
In grid trading, traders set a price range and let automated orders execute buy and sell orders as the market moves back and forth.
HTX ranks among the world’s 10 largest exchanges by trading volume, liquidity and platform traffic, according to CoinMarketCap.
Related: Pakistan clears Binance and HTX to seek local crypto licenses
Coinbase expands use of AI agents across trading tools
While grid trading bots automate execution using fixed rules, AI-powered agents are built to make autonomous decisions, interact through natural language and operate directly onchain. Coinbase has been among the most active crypto exchanges exploring AI agents.
As early as August 2024, CEO Brian Armstrong said Coinbase had tested AI agents, including a transaction in which one automated bot used crypto tokens to interact with another AI system and purchase AI training data, a process he described as “tokens buying tokens.”
In October of the same year, Coinbase rolled out “Based Agent,” a tool that lets users create AI agents linked to crypto wallets for automated onchain activity, including trading, swaps and staking.
In October 2025 Coinbase introduced Payments MCP, a tool designed to let AI agents interact directly with onchain financial services without requiring API keys. The system allows large language models to access wallets, onramps and stablecoin payments through natural language prompts using the Model Context Protocol.
Interest in AI-managed trading appears to be rising, with an April CoinGecko survey showing that about 36% of respondents would allow AI agents to manage most of their crypto holdings.
Still, some experts have warned about the risks. Aaron Ratcliff, attributions lead at blockchain intelligence company Merkle Science, told Cointelegraph that giving AI agents access to crypto wallets adds a new layer of trust to systems designed to be trustless, shifting much of the security responsibility back to users.
Magazine: One metric shows crypto is now in a bear market: Carl ‘The Moon’