- Trending
- Comments
- Latest
Blockchain analytics firm Nansen has launched AI-powered trading tools that allow users to buy and sell cryptocurrencies on Base and Solana by typing conversational commands into a mobile app, expanding beyond its core data analytics business into direct transaction execution.
The feature uses natural language processing to interpret trading instructions like “buy $100 of ETH on Base” and executes them through an embedded self-custodial wallet, with Nansen claiming its AI agents outperform general-purpose chatbots by drawing on a proprietary database of hundreds of millions of labeled blockchain addresses.
The trading feature initially supports activity on Base and Solana with plans to expand to additional networks.
According to Nansen, its AI interface draws on a proprietary onchain database containing hundreds of millions of labeled blockchain addresses, designed to deliver more accurate insights than general purpose bots used for Ai crypto trading.
“For years, our priority has been delivering high quality onchain signals to investors,” said Nansen co-founder and CEO Alex Svanevik.
“This launch lets users act on those insights directly.”
He added that the update “closes the loop” by combining analytics and execution through an AI native mobile experience and a web trading terminal, further streamlining Ai crypto trading workflows.
To enable cross chain execution across Solana and Base, Nansen partnered with decentralized exchange Jupiter, crypto exchange OKX, and cross chain protocol LI.FI.
Trades are executed via the embedded Nansen Wallet, powered by Privy supporting a self custodied experience tailored for Ai crypto trading.
Autonomous trading became available Wednesday, excluding users in restricted jurisdictions such as Singapore, Cuba, Iran, North Korea, Syria, Russia, and parts of Ukraine due to regulatory limits.
Svanevik claimed Nansen’s AI agents outperform mainstream chatbots in onchain analysis tasks critical to Ai crypto trading, including wallet identification, token discovery, due diligence, and portfolio analysis.
He said Nansen AI achieved over 85% quality scores in expert evaluations compared with roughly 20% for general large language models noting that standard systems lack deep wallet labeling and trading pattern awareness.
While Nansen relies on foundational models from firms like Google and OpenAI, Svanevik emphasized that these models must be heavily fine tuned and augmented with onchain intelligence to be effective for Ai crypto trading.
The release underscores growing interest in Ai crypto trading across the industry as platforms explore automation and conversational interfaces to lower entry barriers for retail users.
Recent testing suggests some lower cost models can even outperform larger systems in trading scenarios.
In an autonomous trading competition reported by Cointelegraph in November 2025, models such as QWEN3 MAX and DeepSeek posted stronger results than several high profile chatbots highlighting persistent challenges in real time execution for general purpose AI and reinforcing the need for specialized Ai crypto trading tools.
