Bittensor is a decentralized network that creates open markets for artificial intelligence. Instead of one company building AI behind closed doors, Bittensor allows anyone to contribute AI models, compete for rewards, and build on top of each other's work.

Think of it as an open economy for machine intelligence. Miners produce AI outputs. Validators evaluate quality. The best work gets rewarded with TAO. The network runs across dozens of specialized subnets, each focused on a different type of AI task.

TAO is the native token of the Bittensor network. It serves three functions: it is the reward miners earn for producing useful AI work, it is the staking asset validators lock up to participate in consensus, and it is the currency used to register subnets and operate within the ecosystem.

TAO has a fixed supply of 21 million tokens, the same cap as Bitcoin. New TAO enters circulation through emissions — the network's block reward system that distributes tokens to miners and validators based on performance.

Subnets are specialized markets within Bittensor. Each subnet is designed to perform a specific type of AI work — text generation, image recognition, financial modeling, video analysis, data scraping, and many others.

Each subnet has its own miners (who do the AI work), validators (who judge quality), and a subnet owner (who defines the rules and incentive structure). Subnets compete against each other for a share of the network's total emissions. Better-performing subnets attract more stake and earn more TAO.

As of now, there are dozens of active subnets, each with its own token, team, and thesis.

Staking in Bittensor means locking your TAO behind a validator to help secure the network and earn rewards. When you stake, you are delegating your TAO to a validator who evaluates miners on your behalf.

With the introduction of dTAO, staking has evolved. You can now stake directly into individual subnets, which gives you exposure to that subnet's performance through its alpha token. Staking is no longer just about passive yield — it is a directional bet on which subnets will perform.

dTAO (dynamic TAO) is the upgraded tokenomics model that replaced the original staking system. Under dTAO, each subnet has its own token — called an alpha token — that trades against TAO in a liquidity pool.

When you stake TAO into a subnet, you receive that subnet's alpha token. The price of the alpha token moves based on supply and demand in the pool. If a subnet attracts more stake, the alpha token price rises. If people exit, it falls.

dTAO replaced the old system where a small group of validators decided how emissions were distributed. Now the market decides. Capital flows toward subnets people believe in, and pricing is continuous rather than set by committee.

Emissions are how new TAO enters circulation. Every block, the network distributes TAO to validators and miners across all active subnets. How much each subnet receives depends on how much TAO is staked into it relative to the total network stake.

Emissions are the lifeblood of the ecosystem. They reward useful AI work and create the economic incentive for miners and validators to participate. For subnet token holders, emissions represent a form of yield — the subnet earns TAO, which flows back to stakers through the alpha token.

Anyone can create a subnet on Bittensor by paying a registration cost in TAO. The cost fluctuates based on demand — when many subnets are being registered, the price goes up.

Registering a subnet does not guarantee success. The subnet owner must define a useful task, attract miners to compete on it, and build enough traction to earn meaningful emissions. Most subnets fail. The ones that survive typically have clear commercial use cases, strong teams, and defensible evaluation methods.

Miners do the AI work. They run models, generate outputs, and compete against each other within a subnet. The better a miner's output, the more TAO they earn.

Validators judge the miners. They send tasks, evaluate responses, and assign scores that determine how emissions are distributed. Validators stake TAO to participate and earn rewards proportional to their stake.

The validator-miner dynamic is the core of Bittensor's quality control. Without good validators, bad miners get rewarded. Without good miners, the subnet produces nothing useful. The best subnets have tight evaluation loops where quality is measurable and gaming is hard.

Under dTAO, every subnet has its own token — an alpha token. When you stake TAO into a subnet, your TAO goes into a liquidity pool and you receive the subnet's alpha token in return.

Alpha tokens represent your share of that subnet's economic output. If the subnet earns strong emissions and attracts more stake, the alpha token appreciates against TAO. If people exit or emissions drop, the alpha token depreciates.

Alpha tokens are what make Bittensor investable at the subnet level. Instead of just holding TAO and getting exposure to the whole network, you can take concentrated positions in individual subnets you believe will outperform.

Each subnet's alpha token trades in an automated market maker (AMM) pool — a constant-product pool of TAO and the alpha token. When you stake TAO into a subnet, you are buying the alpha token from this pool. When you unstake, you are selling it back.

The price of the alpha token is determined by the ratio of TAO to alpha in the pool. More TAO going in means the alpha price goes up. More alpha being sold back means it goes down. This is the same basic AMM mechanic used in DeFi protocols like Uniswap.

Most subnet pools are small. That means even modest trades can move the price significantly. If a pool has 500 TAO in it and someone stakes 200 TAO, the alpha token price jumps. If that same person exits, it crashes.

Low liquidity creates two problems. First, slippage — the difference between the price you expect and the price you actually get. On a small pool, a trade that looks like it costs 100 TAO might actually cost 115 TAO once slippage is factored in. Second, volatility — prices swing hard on low volume, which makes subnet tokens feel more volatile than the underlying fundamentals justify.

Before staking into any subnet, check the pool size. If it is small relative to the position you want to take, be prepared for significant price impact on both entry and exit.

Not all subnets are worth your TAO. Evaluating a subnet means looking at the team, the product, the economic model, and the competitive position. Here is a basic framework:

• Does the subnet solve a real problem with measurable demand?
• Is the team credible, accessible, and shipping product?
• Is there a path to revenue beyond emissions?
• Is the evaluation method rigorous and hard to game?
• How much TAO is staked and how liquid is the pool?
• What does the competitive landscape look like — on Bittensor and off-chain?

The subnets that survive long-term will be the ones that generate real revenue and have evaluation systems that consistently reward quality work.

APY (annual percentage yield) in the subnet context represents the estimated return from emissions — how much TAO your staked position earns over a year, expressed as a percentage. A subnet showing 100% APY means that if you stake 100 TAO, you would earn roughly 100 TAO in emissions over a year at the current rate.

Important: APY is not fixed. It changes constantly based on how much TAO is staked in the subnet, the subnet's share of total emissions, and the alpha token price. Today's 200% APY can become 40% APY next month if more capital flows in.

Extremely high APY numbers — 200%, 500%, or more — often signal one of two things: the subnet is very early and has little stake, or the alpha token has been depressed by selling pressure. Both situations carry real risk.

A high APY on a low-stake subnet means the pool is tiny. Your entry will move the price, your emissions will compress the APY, and your exit will crater the token. You might earn 200% APY in emissions while the alpha token drops 60% in value — a net loss.

High APY is not free money. It is compensation for risk. Treat it accordingly.

Slippage is the cost of trading in a thin pool. If a subnet's AMM pool holds 300 TAO and you want to stake 100 TAO, you are buying against a small reserve. The first TAO you stake buys alpha at one price, but each additional TAO buys at a worse price because you are draining the pool.

The same applies on exit. Selling a large alpha position into a small pool means you receive less TAO than the quoted price suggests. For larger positions, slippage can easily eat 10-20% of your value.

Rule of thumb: if your position is more than 5-10% of the pool, expect meaningful slippage.

In a market this early, the founder is the single biggest variable. A good subnet idea with a weak founder will fail. A mediocre idea with a strong founder might pivot into something valuable.

What to look for: relevant technical or domain expertise, a history of shipping product, transparency about challenges and risks, willingness to do AMAs and answer hard questions, and a clear articulation of what they are building and why Bittensor is the right platform for it.

The AMA Archive on TAO Nerds is designed to give you direct access to founder thinking — unscripted, in their own words.

Emissions are not revenue. They are block rewards. A subnet that only earns from emissions is entirely dependent on the network continuing to allocate TAO its way. That is not a business — it is a subsidy.

The subnets worth watching are the ones building toward real revenue: enterprise contracts, API fees, marketplace transactions, data licensing, or SaaS products. Revenue means the subnet can sustain itself even if emissions decline, and it gives the alpha token a fundamental floor beyond speculation.

When evaluating revenue claims, ask whether the numbers are independently verifiable or founder-reported, whether the revenue is recurring or one-time, and whether there is a repeatable sales process or just early inbound interest.

TAO Nerds uses a community-driven scoring system. Each subnet in the Intelligence Index shows a Community Score based on votes from site visitors. The score reflects collective sentiment — not a formal audit or financial recommendation.

Alongside the score, each subnet entry includes a category tag, founder name, AMA date, and a short excerpt summarizing the Nerds Take — our editorial assessment of the subnet after hosting an AMA with the founder.

The ratings are a starting point, not a conclusion. Use them to narrow your research, then go deeper with the full AMA writeups and your own due diligence.

The AMA Archive contains every founder conversation The Nerds have hosted, written up as structured analysis. Each entry includes a subnet thesis, a Nerds Take (our editorial opinion), and a detailed writeup covering what the founder said, what impressed us, and what risks we see.

Start with subnets in categories that interest you. Read the Nerds Take first for a quick assessment. If the subnet looks interesting, read the full writeup. Then check the Community Vote page to see how others rated it.

AMAs are not endorsements. They are information. The goal is to give you enough signal to make your own decision.