Bittensor subnets have reached a combined 70 million USD in annualized recurring revenue, marking a 40 percent monthly leap driven by real-world demand for artificial intelligence compute.
By Aisha Okonkwo | November 7, 2026
The Announcement
A fresh benchmark report from the SubConnect Bittensor Revenue Index reveals that decentralized artificial intelligence networks are starting to generate meaningful commercial cash flow. Across the entire Bittensor ecosystem, specialized networks known as subnets have achieved an estimated 70 million USD in annualized recurring revenue.
The milestone represents a rapid acceleration in paid customer activity. Monthly earnings jumped 40 percent from October’s baseline of roughly 50 million USD, effectively doubling the 32 million to 35 million USD run rate recorded when tracking first began in late August and early September. Analysts monitoring the index note that if current enterprise adoption continues, annualized receipts could cross the nine-figure mark before the calendar year closes.
This surge in commercial revenue comes during a tense period for digital assets. Broader crypto sentiment remains weighed down by macroeconomic caution, with Bitcoin trading at 75,758 USD, Ethereum changing hands at 2,239 USD, and Solana sitting near 89.38 USD. The daily Crypto Fear and Greed Index registered a reading of 25, pointing to Extreme Fear across retail desks. Yet while speculative trading volumes have slumped across major exchanges, enterprise customers appear increasingly willing to spend hard cash for decentralized computing power.
What It Actually Does
For months, critics have questioned whether decentralized AI protocols do anything beyond printing digital tokens for speculators. The latest data suggests a genuine commercial market is forming beneath the noise. According to the index findings, between 78 percent and 82 percent of all verified customer income is concentrated in compute and infrastructure subnets that rent out processing power to outside businesses.
Instead of relying on token subsidies, these networks function as market brokers for high-performance hardware:
- Lium (SN51) — operates as a peer-to-peer marketplace connecting graphic processing unit (GPU) owners with software developers who need short-term computing power to train machine learning models.
- Targon (SN4) — provides enterprise-grade chip clusters and confidential computing environments, allowing corporations to run sensitive artificial intelligence workloads without leaking proprietary code.
- Engy (SN53) — delivers specialized computing infrastructure tailored to real-time machine intelligence tasks and complex data processing.
- Specialized service hubs — including platforms like Bitcast, Almanac, and Score, which monetize live data feeds, translation engines, and automated code review pipelines.
Rather than swapping speculative assets back and forth, external developers and private tech firms are depositing actual funds to rent processing power, query algorithmic models, and extract structured web information.
How It Works, in Plain English
To understand why this matters for your portfolio, it helps to imagine the Bittensor network not as a single tech firm, but as an open-air shopping center. The main blockchain provides the building, the security guards, and the parking lot. The individual subnets are independent store owners leasing space inside the mall.
Under the network’s Dynamic TAO (dTAO) economic rules, every storefront issues its own specific shares, known as alpha tokens. In the past, many blockchain projects survived solely by handing out newly minted tokens to anyone who walked through the door — much like a struggling restaurant printing endless free discount coupons. Over time, that constant inflation pushes token values down.
The shift happening today changes that equation entirely. When an outside customer hires Lium or Targon to run a heavy machine learning calculation, that client pays in cash or stablecoins. The subnet management does not merely pocket the funds or sit on them. Instead, teams take those real business earnings and execute open-market buybacks of their own alpha tokens.
By buying tokens back from public markets and retiring them into reserves, the subnets create organic purchasing pressure backed by customer activity rather than promotional promises. In simple terms: when an artificial intelligence firm rents computer chips from a subnet, a portion of that payment automatically buys back the subnet’s digital tokens.
Who Benefits
The emergence of verified commercial revenue changes the risk profile for several groups across the crypto economy:
- Everyday stakers — retail investors who delegate tokens to productive subnets gain exposure to systems supported by outside cash flow rather than pure token dilution.
- Hardware providers — independent data centers and individual computer owners holding expensive graphic cards can earn steady income by leasing idle capacity, competing directly against traditional cloud monopolies.
- Software startups — small technology firms gain access to flexible computing clusters on demand without signing rigid, multi-year hosting contracts with centralized tech giants.
- Long-term token holders — open-market token buybacks remove circulating supply, helping insulate network economies against broader market downturns.
For ordinary investors, the key lesson is that decentralized software is beginning to prove its utility in the commercial world. When digital asset prices experience heavy turbulence, protocols with measurable customer revenue tend to establish stronger valuation floors than projects sustained only by narrative momentum.
The Open Questions
While reaching 70 million USD in annualized sales marks an important milestone, critical financial hurdles remain. The largest challenge is the massive gap between organic customer revenue and protocol inflation. Bittensor distributes an estimated over 300 million USD in annual token emissions to subsidize network participants.
That means verified external sales currently offset only between 9 percent and 23 percent of the network’s total yearly token dilution. In plain terms: the ecosystem still pays out significantly more in newly created tokens than it brings in from outside customers. For the network to become genuinely self-sustaining without relying on ongoing inflation, paying customer demand will need to multiply several times over.
There is also heavy concentration risk. With roughly four out of every five revenue dollars flowing into just a few compute rental networks like Lium and Targon, dozens of other subnets focused on specialized AI agents, financial forecasting, and language processing are still struggling to build sustainable client bases. If demand for general GPU compute cools or centralized tech giants slash rental rates, revenue growth could stall rapidly.
Aisha Okonkwo’s take
Seeing real paying customers generate 70 million USD proves to me that decentralized machine intelligence can function as an actual business rather than an academic simulation. My focus stays fixed on the 300 million USD annual emission gap heading into December to see whether organic client fees can outrun heavy token inflation. If top compute providers abandon these decentralized clusters to chase private enterprise contracts, the economic engine driving these token buybacks evaporates.
The cryptocurrency market remains highly volatile. This article is for informational purposes only and does not constitute financial advice.
40% jump in one month, from 50M to a 70M run rate. dTAO doomers went real quiet lately lol
And October was already up from the 32-35M run rate before it. Enterprise demand renting SN4 clusters is doing the heavy lifting here.
Targon renting confidential chip clusters to corporates is the underrated part. Actual enterprise demand instead of subnet-to-subnet ping pong.
the mall analogy in the article covers it, SN4 is the anchor tenant. rest of the stores ride that traffic