Stillcore Capital Makes the Case for Bittensor as the “Third Great Ecosystem” After Bitcoin and Ethereum

The Bitcoin Mining Analogy

Bittensor takes an idea that already worked spectacularly once and applies it to AI. Bitcoin incentivized the world to donate electricity and computing power to build what became Earth’s largest supercomputer — a network 500 times more powerful than the next largest — and generated $2.5 trillion in value at its peak. Bittensor asks a simple question: what if you could make that mining process programmable?

Mark Jeffrey, who wrote one of the earliest books on Bitcoin in 2013 and participated in the Ethereum ICO, frames it plainly: imagine instead of forming a company, you post a list of tasks, and freelancers from around the world compete to finish each one. Only the best get paid; not by you, but by a blockchain. Bittensor takes that contest model and points it at AI.

Where Ethereum extended Bitcoin’s smart contract capabilities to create programmable money, DeFi, and the ICO boom, Bittensor extends the other side (the mining and reward mechanism) making it programmable. That distinction has never been done before.

Inside the Subnet Economy

Bittensor is subdivided into 128 subnets, each a separate task contest with its own owner, incentive design, and purpose. The range of what these subnets are building is striking:

Ridges (Subnet 62) is developing AI-powered coding assistants that have beaten Claude and Cursor on the SWE-Bench benchmark at certain points, at roughly one-seventieth the cost.

404 Gen (Subnet 17) generates AI-created 3D objects for gaming and design, amassing what the team describes as the world’s largest library of 3D objects — 25 million and counting — all AI-generated.

Nova (Subnet 68) is mining for new pharmaceutical molecules in drug discovery.

Templar and IOTA are building decentralized AI training networks designed to rival the massive GPU clusters owned by the likes of Elon Musk, Meta, and Google. Templar has released a 70-billion-parameter model and was cited by an Anthropic co-founder as the premier example of decentralized training architecture.

The ecosystem also includes Bitcast, a decentralized advertising network where AI verifies compliance with advertiser briefs and the blockchain handles all payments — no invoices, no contracts, no human review.

The Stillcore Thesis

Stillcore Capital, co-founded by Rob Greer, Mark Jeffrey, and prolific angel investor Jason Calacanis, is structured as a single-ecosystem fund focused entirely on Bittensor. Their framing draws from historical precedent: in the early 1990s, Microsoft, AOL, and CompuServe were the well-capitalized incumbents everyone expected to dominate the internet. Then TCP/IP, Linux, and the World Wide Web converged on an open-source substrate and changed everything.

Stillcore believes Bittensor is that open-source substrate for AI. The current cast of centralized incumbents — OpenAI, Anthropic, Google DeepMind, xAI — collectively command roughly $1.5 trillion in valuation. Bittensor sits at around $1.7 billion. The fund’s thesis rests on the conviction that this valuation gap is unsustainable.

Calacanis, historically a crypto skeptic, was persuaded by a specific idea: using financial incentives to coordinate open-source AI development at global scale. The concept that GPUs and developer talent scattered around the world could be organized through token economics, without a central company, clicked for him.

The Investment Approach

Stillcore operates more like a venture capital firm than a hedge fund, despite investing in liquid tokens. The fund benchmarks itself against TAO (the base Bittensor token) rather than the S&P 500, and the team claims to be significantly outperforming that benchmark.

Their portfolio strategy blends traditional VC analysis — total addressable market, competitive differentiation, go-to-market strategy — with subnet-specific evaluation of incentive mechanisms and mining quality. They have publicly disclosed three investments:

Ridges is their lead position, viewed as the best-productized subnet in the ecosystem and a direct competitor to Cursor at a fraction of the cost.

Hippius provides decentralized storage at prices the team describes as between one-four-hundredth and one-four-thousandth of competitors. Its go-to-market strategy leverages WordPress, which powers roughly half of all websites, through a storage plugin that offers dramatic cost savings.

Targon (Subnet 4) offers enterprise-grade private inference at roughly one-seventh to one-tenth the cost of competitors, with end-to-end encryption and a guarantee that prompts are not used to retrain models — addressing a complaint publicly voiced by Chamath Palihapitiya about centralized AI providers consuming private data.

The Cost Advantage and Compounding Effects

A recurring theme throughout the conversation is the structural cost advantage that Bittensor subnets enjoy. Because the blockchain subsidizes miners through token emissions, subnets can offer services at a fraction of centralized competitors’ prices. More importantly, these savings compound when subnets stack on top of each other.

Ridges, for example, uses inference from two other subnets — Targon and Chutes — which themselves offer inference at roughly one-tenth the market rate. The result is a compounding cost structure that is native to the Bittensor ecosystem and impossible to replicate in a centralized environment.

Jeffrey frames Bittensor’s approach to talent the same way Bitcoin approached energy: where Bitcoin incentivized the world’s stranded electricity, Bittensor incentivizes stranded talent. A developer in India, Poland, or Moscow can compete in subnet contests and earn significant income without needing a corporate job or venture backing.

The ICO Parallel and What’s Missing

Both Jeffrey and Greer draw a direct line to Ethereum’s ICO era. Jeffrey recalls watching Ethereum sit dormant for two years before the ICO phase exploded. The capital was always there; no one had invited it in.

The parallel to Bittensor’s current state is pointed. Subnet valuations range from roughly $1 million to $80 million at the top end. No subnet has reached a billion-dollar market cap. The aggregate market cap of all subnets barely touches $1 billion — less than many individual Silicon Valley startups raising their Series B.

The team identifies three reasons for this lag: wallet and exchange infrastructure has only recently become usable, subnet tokens are not available on major centralized exchanges like Coinbase, and no single subnet has yet had its breakout viral moment. They believe all three barriers are on the verge of falling.

The Bear Case

When pressed on risks, Jeffrey is direct: Bittensor’s back-end talent is world-class, but the ecosystem lacks equivalent strength in front-end design, productization, and marketing. If Bittensor fails, it will be because subnet teams couldn’t translate brilliant technology into products that consumers and enterprises actually adopt.

There is also the risk that subnets could use Bittensor emissions to subsidize R&D, achieve product-market fit, and then leave the ecosystem — launching their own tokens or raising traditional venture capital. Greer argues this is less likely than it appears because the mining activity is typically central to the product itself, making it difficult to decouple. Subnets also have strong incentives to support their token holders, since rising token prices directly increase the emissions revenue flowing into the subnet.

The Price Targets

In the rapid-fire closing, the fund co-founders offered their predictions. Greer pegged Bitcoin at $150,000 by year-end. Jeffrey predicted Ridges as the breakout subnet of 2026 and set a TAO price target of $3,000 by end of year — a figure he described as conservative based on Bitcoin’s own historical price trajectory from a similar stage.

Looking further out, Jeffrey projected Bittensor becoming a trillion-dollar ecosystem by 2030.

This article is based on a conversation from The Supercycle Podcast featuring Rob Greer and Mark Jeffrey of Stillcore Capital. The views expressed represent those of the interview subjects and should not be construed as investment advice.

WATCH THE FULL CONVERSATION BELOW: