Bittensor AI Subnets Are Finally Finding Paying Customers as Operator Revenue Passes 32 Million USD

Bittensor subnet operator revenue has grown from near zero to an estimated 32 million USD-plus over the past 18 months as companies have signed commercial implementation deals built on the network’s AI competitions, according to Yuma chief revenue officer Evan Malanga.

In comments shared with crypto.news, Malanga argued that customer deals, operator revenue, and subnet token purchases together offer the best available evidence of growing real-world use across the 128-market network — a counterpoint to skeptics who see Bittensor as little more than a token trading vehicle.

## From carwashes to Fortune 500 firewalls

The strongest examples Malanga cited span computer vision and cybersecurity. Score, a subnet operator, has planned carwash monitoring deployments for Avia across thousands of locations — a use case that sounds mundane but represents exactly the kind of paid, recurring commercial work that critics have long said Bittensor lacks. RedTeam, another operator, serves multiple Fortune 500 financial institutions and protects more than 1 billion transactions each week, according to Malanga.

The supplied Yuma background also described Innerworks, a London-based bot-detection company, as a business embedding Bittensor competitions in its products. Innerworks reportedly increased its detection rate from 73 percent to 99.5 percent in less than a year.

Yuma describes Bittensor’s subnets as task-specific contests with public scoring, where teams receive tokens for producing the best-performing AI work. Businesses running many of these competitions then incorporate the winning outputs into their own products.

## How the tokenomics connect

Malanga also pointed to an increasing number of subnet operators buying their own subnet tokens on the open market. Under Bittensor’s design, demand for subnet assets connects directly to demand for TAO, the network’s native token — meaning commercial adoption at the subnet level feeds back into the base asset.

Bittensor supports 128 markets, or subnets, covering tasks such as mathematical proofs, machine learning, prediction, and resource optimization. Within each subnet, miners compete to answer a defined task while validators assess the quality of the work. Malanga compared the process to Bitcoin mining, where participants compete to find a valid nonce, but stressed the broader scope.

“Bitcoin rewards one type of work, while Bittensor can reward many forms of machine intelligence,” he said.

On reward allocation, independent validators score competing solutions, and miners receive rewards according to the usefulness assigned to their outputs. Bittensor aggregates those assessments through Yuma Consensus, a mechanism designed to resist collusion while still allowing agreement — including when parts of an evaluation are subjective. Miners whose solutions receive higher utility assessments collect a larger share of newly issued tokens.

For lasting commercial value, however, Malanga was explicit that the reward system alone is not the point. “Lasting commercial value is ultimately determined by external demand for the intelligence,” he said, naming Lium for GPU computing, Chutes for AI inference, Leadpoet for AI sales intelligence, Score for computer vision, and RedTeam for cybersecurity as services available through subnets.

## Growing stakes, listed exposure

The adoption claims come against a backdrop of rising commitment to the network. Earlier coverage in March documented the value of TAO staked across subnets exceeding 620 million USD, with the subnet count growing from roughly 80 to more than 120 over the period examined.

For United States equity investors, TAO Synergies, a Nasdaq-listed company, disclosed 42,111 TAO in treasury holdings in August 2025 — a mix of purchased tokens and holdings generated through staking. The company had bought 10 million USD worth of TAO the previous month as part of a strategy concentrated on Bittensor, giving American stock-market investors an indirect route to network exposure.

Yuma itself, which is backed by Digital Currency Group, builds and invests in infrastructure for specialized open-source AI on Bittensor. The company says it operates the network’s largest owned-hardware validator, with more than 200 million USD in TAO and subnet tokens staked.

## The risks remain real

Asked about the risks to that position, Malanga did not sugarcoat them. He identified individual subnets failing to turn their competition outputs into commercial businesses, early-stage asset volatility, and competition from established AI developers — the large, well-funded platforms building outside the crypto ecosystem — as the principal threats.

The 32 million USD revenue figure is also an estimate presented by a party with a significant stake in the network’s success, and it remains small next to the multibillion-dollar valuations AI infrastructure commands elsewhere. Still, paying customers at thousands of physical locations and billions of weekly protected transactions are the kind of metrics Bittensor has rarely been able to point to before.

At the time of writing, Bitcoin trades near 83,614 USD and Ethereum near 2,689 USD, per Binance spot data, with AI-linked tokens continuing to trade as a high-beta segment of the broader market.