Wall Street is starting to notice one of crypto’s smartest AI bets
A growing spotlight on Nof1’s Alpha Arena suggests SUI Group and Karatage may have gotten early to one of the most important experiments in finance: teaching AI how to trade in real markets.
May 15, 2026, 11:30 a.m. 2 min read
For months, the loudest question in AI investing has been deceptively simple: If large language models can replace lawyers, consultants and coders, can they replace hedge fund managers too?
That debate is increasingly moving from theory into live markets, and one startup at its center is Nof1, the AI trading lab backed by SUI Group (SUIG) and Karatage.
SUI Group is a Nasdaq-listed company focused on building institutional exposure for the Sui blockchain ecosystem. Karatage is a London-based proprietary hedge fund specializing in emerging-technology investments across digital assets, artificial intelligence and gaming.
Nof1’s flagship experiment, Alpha Arena, pits frontier AI models from companies like OpenAI, Anthropic, Google and xAI against each other in autonomous trading competitions using real capital.
The early results have been messy. Most models lost money, overtraded or struggled with risk management. But for sophisticated investors, that failure may be the signal.
The easy AI trade has been backing companies building chatbots and copilots. The harder, and potentially far larger, opportunity is understanding how AI systems behave in adversarial environments like financial markets, where prediction alone is not enough. Timing, position sizing, adaptability and risk control matter just as much.
AI has spent the past two years conquering language. Trading may be the next frontier, and a much harder one. Financial markets are fast-moving and unforgiving, rewarding systems not for sounding intelligent but for consistently allocating capital under uncertainty. That challenge is creating a new category of AI experimentation, where firms are testing whether frontier models can evolve from chatbots into autonomous financial agents.
Nof1’s thesis is that today’s frontier models were not trained to allocate capital; they were trained to predict language. Alpha Arena attempts to close that gap by exposing models to live financial environments where they are rewarded, or punished, by real market outcomes.
That helps explain why SUI Group and Karatage co-led Nof1’s $15 million funding round while also backing Recursive Superintelligence with TwinPath Ventures, a self-improving AI startup reportedly valued above $4 billion. Together, the investments look less like isolated venture bets and more like a coordinated thesis on where AI and markets are heading.
Both Nof1 and Recursive are grounded in Open-Endedness Research.
The wager is straightforward. If autonomous agents eventually become a core part of investing, the firms building the infrastructure and training environments today could end up shaping the future of finance.
Following season two, Nof1 plans to launch a consumer platform which will be the world’s first coding agents for markets. The platform will feature the company's frontier models, developed using open ended AI architecture which continually improves and aims to eclipse human and algorithmic traders.
Read more: Crypto is built for AI agents, not humans, says Alchemy's CEO
AI Disclaimer: Parts of this article were generated with the assistance from AI tools and reviewed by our editorial team to ensure accuracy and adherence to our standards. For more information, see CoinDesk's full AI Policy.
More For You
The new earn product lets users generate BTC-denominated rewards through DeFi strategies while keeping exposure to bitcoin’s price.
What to know:
- Kraken launched Bitcoin Vault, a new product within Kraken Earn aimed at long-term bitcoin holders seeking passive yield.
- The product is powered by Veda and operated by Sentora, with funds deployed across DeFi protocols including Aave and Morpho.
- The crypto platform's broader DeFi Earn offering has grown to more than...