Bittensor’s TAO token is trading around $275, giving the network a market capitalization of roughly $2.6 billion and a fully diluted valuation near $5.8 billion. A report cited from Pine Analytics argues that this valuation is being driven far more by scarcity, staking, ETF expectations, and AI sector sentiment than by verified revenue fundamentals.
Chutes captures the largest share of token emissions
According to the figures in the report, Bittensor distributes 3,600 TAO per day across subnet owners, miners, validators, and other participants under a fixed issuance model. The top 10 subnets control about 56% of total emissions, and Chutes holds the largest single share at 14.4%. That puts its daily allocation at roughly 518 TAO, worth about $52 million annually at the cited price levels.
Chutes, developed by Rayon Labs, offers serverless inference for open-source models. The report says it has more than 400,000 users, including over 100,000 API users, processes more than 5 million requests per day, and has handled a cumulative 9.1 trillion tokens. Within the Bittensor ecosystem, it stands out as the clearest demand-side case study.
Low pricing appears to depend on token incentives
The central claim is that Chutes’ pricing is not mainly the product of operating efficiency. Its external annual revenue is estimated at just $1.3 million to $2.4 million, with the upper figure based on team-reported data that has not been independently audited. On that basis, the subnet’s subsidy-to-revenue ratio comes out to roughly 22:1 to 40:1. In practical terms, for every $1 paid by users, the network is distributing $22 to $40 in TAO inflation to support the service.
Removing that subsidy changes the cost picture. Based on reported throughput of around 101 billion tokens per day, the implied cost works out to about $1.41 per million tokens. The report compares that with centralized market pricing: Together.ai’s LLaMA 3.3 70B Turbo at about $0.88, DeepSeek V3 at roughly $0.40 to $0.80, and some smaller models as low as $0.18 per million tokens. Without incentives, Chutes would be priced around 1.6x to 3.5x above centralized alternatives.
Demand remains hard to verify across the network
Bittensor’s supply side is relatively transparent. Daily issuance, halving rules, staking levels, and token flows are visible on-chain. Demand is not. Inference calls, training jobs, and compute usage happen off-chain, so investors cannot directly observe API activity through blockchain data. They are left to infer demand from staking flows, subnet token prices, and project disclosures.
Using the figures cited in the report, confirmed annual demand-side revenue across the entire network is only around $3 million to $15 million. Chutes alone receives about $52 million a year in annualized subsidies, more than the high end of total external network revenue. Beyond Chutes, Targon is described as the highest-revenue subnet with estimated annual revenue of about $10.4 million and an implied valuation near $48 million. Templar completed Covenant-72B training, but its external revenue is listed at zero.
Valuation multiples sit far above common AI infrastructure ranges
At the reported revenue range of $3 million to $15 million, TAO is trading at roughly 175x to 200x revenue on market cap, and close to 400x on a fully diluted basis. The report contrasts that with centralized AI infrastructure companies, which it says have recently raised capital at around 15x to 25x forward revenue, while even fast-growing SaaS names rarely sustain multiples above 50x for long.
It also argues that Bittensor subnets face pressure from both sides. Open-source models and local deployment tools lower the ceiling for self-hosted inference, while Microsoft, Google, Amazon, and Meta are said to have combined for more than $200 billion in AI capital expenditures in 2025. Bittensor’s full-year incentive budget, by comparison, is about $360 million. Any subnet that builds a useful service still has to absorb token friction, validator overhead, subnet owner take rates, and network latency tied to the decentralized model.
The report does not dispute Bittensor’s appeal as a narrative asset. It points to the 21 million token cap, Bitcoin-style halvings, growth in subnet count from 32 to 128, and institutional catalysts as factors supporting market interest. Its core argument is narrower: current TAO pricing looks far more tied to scarcity and sentiment than to a proven base of AI service revenue.