Betting on High-Margin Subnets
Crypto utility networks live and die by the market. In the bootstrapping phase, token incentives are the only way to attract miners to provide a service. They’re what bootstrapped storage on Filecoin, bandwidth on Helium, and compute on Bittensor. Without incentives, no one plugs in hardware, runs software, or contributes resources, meaning the network never gets off the ground.
But token incentives are only as durable as the token price. In the pre–product market fit phase, a falling token quickly erodes miner margins. This thins the field as weaker miners drop out, and in sharper downturns, it can force a mass exit when mining is no longer viable. In other words, the network risks falling into a death spiral. If a business is being built on top of an unsteady network, the fragility can cripple the business before it ever matures.
On Bittensor, the subnets most resilient to this fragility are those that coordinate human knowledge and expertise with minimal capital requirements. By that, we mean miners contributing their intelligence directly—writing code, refining models, generating insights.
We call these high-margin subnets, where mining costs stay low, enabling subnets to better:
Weather volatility and reach product-market fit, where reliance on incentives fades.
Attract more miners, increasing competition and accelerating innovation.
A Miner’s Breakeven Formula
Every miner is effectively running a small business. Their revenue is the rewards they earn from the subnet, and their costs are whatever it takes to compete, whether that’s paying for cloud compute, purchasing data, or just investing their own time. Loyalty doesn’t extend beyond that equation. Miners show up purely to make money.
This creates a brutally rational dynamic. A miner will only participate if the rewards justify the costs. If margins disappear, they exit. That’s why the structure of costs matters so much. High CapEx limits the amount of new entrants, meaning weaker competition and less innovation. High OpEx makes existing miners hypersensitive to token prices where a 20% drawdown in price can suddenly make the whole operation unviable. For the subnet teams building a real business on top, both are existential risks.
Subnet owners can’t control token price, but they can shape miner margins. The most resilient subnets are those where costs stay low by design, not bound up in expensive hardware or infrastructure, but in human intelligence and expertise.
Coordinating Human Intelligence
Ridges (Subnet 62) has cracked the code on high-margin mining with an elegant formula: maximize rewards through a winner-take-all structure while slashing mining costs to the bone.
The first smart move was picking a low-cost domain. Mining on Ridges doesn’t require GPUs, data centers, PhD-level credentials, or specialized equipment. It simply built a competition for human intelligence—a tournament where the best coding agent developers win.
The second was eliminating friction for new miners. With just a Google login and a nominal fee of $25, aspiring miners can begin submitting their coding agents immediately. They don’t even need to start their agent code from scratch. Since all miner submissions are open source, newcomers can study the reigning champion's code, understand what works, and build meaningful improvements from there. This open ecosystem dramatically lowers barriers to entry while accelerating the pace of innovation across the network.
The winner-take-all rewards structure means the top miner captures the entire 41% miner emission pool. This high-stakes, high-reward model easily satisfies a miner’s initial cost-benefit calculation. Just winning the competition once and maintaining top of the leaderboard for two days results in earning $120k, more than many make in a year.
Combined with the minimal operating expenses of human time and intellectual effort, rather than hardware or infrastructure, the profit margins meet the threshold for bootstrapping a sustainable, network-driven product.
Where We’re Focused
The Ridges model points toward the future of successful subnets. In our view, the winning formula leverages miners' innovation in activities where human expertise delivers outsized value without requiring massive capital investment.
High-margin, low-overhead competitions are exactly the kind of subnet we want to back. If you’re building in this direction, we’d love to hear from you.
This content is provided for informational purposes only and does not constitute investment advice or a recommendation to buy or sell any security. Unsupervised Capital holds a position in TAO and may hold positions in the subnet tokens or other digital assets discussed herein and may buy, sell, or change positions at any time. Past performance is not indicative of future results. Digital assets involve substantial risk, including potential total loss of capital. Consult your own advisers regarding any investment decisions.
