Astrid Intelligence Is Building AI Hedge Funds That Run Themselves on Bittensor

Imagine a hedge fund run entirely by AI agents, with no human traders, no office, just autonomous software making investment decisions 24/7. That’s what Astrid Intelligence is building on Bittensor.

Astrid is a publicly traded company (ticker ASTR on the Aquis Stock Exchange) that’s gone all-in on decentralized AI. Their main project, Astrid Arena (SN127), trains AI agents to handle financial tasks such as trading, asset allocation, and portfolio management autonomously.

Astrid is raising to $100 million in institutional capital to flow through their AI agents, turning Bittensor into an economic layer where AIs create wealth without human intervention.

What Astrid Actually Does

Astrid Intelligence started as something completely different, a company called Leaf Studios back in 2018. After several name changes and pivots, they rebranded to Astrid in late 2025 and focused entirely on decentralized AI within the Bittensor ecosystem.

Astrid Homepage

Their business model is straightforward. They build and operate AI infrastructure on Bittensor, running validators and miners that earn TAO token rewards. They’ve converted their entire digital asset treasury into TAO holdings, over 18,600 TAO worth millions of dollars as of recently.

But the interesting part is SN127, their flagship subnet on Bittensor. This is where autonomous AI agents learn to handle money. The agents train in simulated environments, competing to make the best investment decisions. The best performers attract more stake and deploy more capital, earning rewards for both themselves and the people who stake on them.

Think of it like this: instead of hiring human fund managers to invest your money, AI agents handle everything automatically using strategies they’ve learned through competition and training.

How SN127 Works

AI agents train in an “arena” environment where they practice financial tasks. They might optimize a portfolio, manage risk, or execute trades based on market data. The agents use sophisticated strategies, some inspired by famous investors like Ray Dalio, but they’re software, so they can operate continuously without sleep or emotional decisions.

When you stake TAO tokens on SN127, you mint a token called dTAO (also called Alpha). This represents your capital being deployed by the AI agents. The agents then use this capital to stake on other Bittensor subnets or execute trades, generating yields from emissions and market movements.

The agents work autonomously. They can create wallets, negotiate with other AI agents, and execute trades across different blockchains thanks to an interoperability bridge. They operate 24/7 with near-zero costs, making money while you sleep.

All of this happens on-chain, meaning it’s transparent and verifiable. You can see what the agents are doing and withdraw your stake anytime you want.

The Technology Behind It

The commit-reveal mechanism prevents cheating. When validators score how well miners performed, they encrypt their scores using something called Drand timelock encryption. The scores get committed now but are only revealed later, preventing other miners from copying winning strategies or front-running trades.

The interoperability bridge connects Bittensor to other blockchains like Ethereum and Solana. This lets AI agents move assets across chains, access different markets, and interact with other blockchain ecosystems seamlessly.

The training arena simulates real financial conditions. Agents learn by doing, trying different strategies, and seeing what works. The incentives are explicit and measurable; agents that perform better earn more rewards, which attracts more stake, which gives them more capital to work with.

Astrid also rebuilt Bittensor’s validator tooling in TypeScript, making it easier for developers who don’t know Rust or Python to participate. This lowers the technical barrier compared to other subnets.

Who’s Behind Astrid

Astrid is run by people with serious backgrounds in trading, hedge funds, and the Bittensor ecosystem.

Mark Creaser is the CEO, appointed in November 2025. He’s been involved in Bittensor for years and co-founded DSV Fund, the world’s first hedge fund dedicated exclusively to Bittensor. DSV launched in 2024 and focuses on investing in TAO and related assets. Creaser understands both traditional finance and decentralized systems.

Siam Kidd serves as Chief Strategy Officer, also joining in November 2025. Like Creaser, he co-founded DSV Fund. Kidd brings over 20 years of experience in trading, hedge funds, and mergers and acquisitions. His background includes being a Royal Air Force pilot before transitioning to finance, bringing disciplined decision-making to high-risk crypto investments.

Olivia Edwards is the Executive Chairperson, appointed in early 2025. She oversees the board and strategic direction, playing a key role in rebranding to Astrid and focusing on decentralized AI infrastructure.

Elliot Fielding handles finances as CFO, appointed in July 2025. He manages compliance and reporting for the publicly traded company.

Léo Mercier joined as Bittensor Strategist in October 2025. He’s an AI entrepreneur with expertise in artificial intelligence and digital assets, leading Astrid’s subnet development and ecosystem integration.

The team is small but experienced, consisting of a network of Bittensor specialists who understand both the technology and the financial side.

The Bigger Vision

Astrid sees a future where AI agents become the primary economic actors. Instead of humans working to create wealth and passing it to future generations, autonomous AI systems accumulate value continuously and operate global financial strategies without human intervention.

Over 50 million AI inferences happen daily across Bittensor. SN127 adds a financial layer on top of that, turning AI capabilities into autonomous economic entities.

If successful, this changes how wealth gets created and managed. You wouldn’t hire a fund manager or pick stocks yourself. You’d delegate to AI agents that optimize your capital continuously, learning and improving over time.


Website: astrid.global

Check out their GitHub at astridintelligence/sn-127

Follow @AstridIntel on X

Leave a Reply

Your email address will not be published. Required fields are marked *