The Rise of Bittensor

Bittensor ($TAO) is a decentralized, permissionless AI network that is quietly emerging as the most compelling testbed for open-source AI coordination. Recently, they reached a milestone of 100 subnets - a market they reached in just two years.

Let’s unpack how we got here, what’s changed, and why TAO is no longer just an “AI token,” but something much bigger.

From 1 to 100: A Timeline of Exponential Growth

• March 2023: First subnet goes live

• May 2024: 33 subnets launched in 14 months

• Feb 2025: 65 subnets total

• April 2025: 100 subnets and counting

At this pace, 200+ subnets by year-end is not just a mere pipedream. And we’re not even counting sub-subnets: deeper layers of specialization that stack functionality on top of existing subnet economies.

Ok, But What Exactly Are Subnets?

At its core, Bittensor is an incentive layer. It rewards useful digital outputs like machine learning models, data pipelines, sports predictions, 3D asset generation, and protein folding. The protocol doesn't care about raw compute power. It cares about how useful your output is. If it's good, you earn TAO. If not, you earn nothing. 

Usefulness is the mining function. This is enforced through subnets. 

Each subnet is a self-contained market that distributes emissions based on performance. Subnet owners define the task, miners compete to solve it, and validators score the results. The better the output, the more emissions the subnet receives.

• Subnet 18 handles math and code inference. It sees over four million queries daily with low latency, proving demand at scale. 

• Subnet 42 runs a fork of AlphaFold2 for protein folding, it processed over 1.1 million queries in 24 hours after launch, the first scientific workload on-chain. 

• Sportstensor focuses on sports betting models and supports agents like “Billy Bet” that operate autonomously. 

• 404Gen handles text-to-3D generation. 

Other subnets offer compute rental markets, competing with centralized GPU marketplaces like Vast, Akash, or Render.

As of now, over 67% of all TAO emissions are flowing into subnets other than the original language model. Bittensor is rapidly decentralizing, and value is moving toward the most useful specialized subnet economies.

Bittensor Tokenomics and dTAO

Bittensor’s tokenomics are based on a hard-capped supply of 21 million TAO. There was no pre-mine and no allocation to venture capital. Emissions follow a halving schedule similar to Bitcoin, making the token increasingly scarce over time. This scarcity supports long-term value.

However, keep in mind that TAO is not a static asset. You will need to participate in the subnet economy. Users stake TAO to access subnet-level activity, either by supporting intelligence outputs or earning emissions. Subnet 0 offers the simplest path for participation, where users stake TAO and receive emissions in TAO directly. 

In 2025, Bittensor launched Dynamic TAO, or dTAO. This upgrade replaced the old emissions system, which relied on validator votes to decide which subnets received rewards. That system created bottlenecks, allowed for insider bias, and made it difficult for new or innovative subnets to break through.

With dTAO, every subnet now has its own alpha token. TAO holders can stake into a subnet’s liquidity pool by buying and holding that alpha token. The emissions allocated to each subnet are now tied to the market price of its alpha token. The higher the price, the larger the share of emissions the subnet receives. This creates a direct feedback loop between perceived value and actual token flow.

What Bittensor Represents

If Bittensor succeeds, we could be looking at the Ethereum of AI coordination. It is important to understand that Bittensor isn’t trying to outbuild OpenAI. This is the first real-time funding model for AI that isn’t tied to corporate boards or grant committees. As one of the community members puts it:

“Bittensor isn’t a bet on AI. It’s a bet on incentivized coordination.”

If it fails? We still end up with a bunch of decentralized AI startups, some of which may 50–100x on their own.

Regardless, Bittensor definitely looks healthy and primed for a monstrous 2025.