Delegating TAO looks like a single action and is actually three decisions: which subnet, which validator, and how long. Most people only make the first one deliberately.

The mechanic, step by step

  1. You pick a subnet (by netuid) and a validator operating on it.
  2. Your TAO is staked into that subnet and you hold the subnet's alpha token as the receipt for your position.
  3. The subnet earns emissions each block. The validator you backed receives a share based on its consensus standing.
  4. The validator keeps a take rate (commonly around 18%, but it varies and is set by the operator) and distributes the rest across its delegators pro rata.
  5. Rewards accrue continuously and compound if you leave them staked.
  6. To exit, you unstake — converting alpha back to TAO at the prevailing pool rate.

The two returns you are actually taking

Return #1 is emissions: the yield figure everyone quotes. Return #2 is the alpha token's price against TAO, which is a market price and can go down a long way. Any APY number you see is only half the picture.

Choosing a validator

  • Take rate — a 10-point difference in take is a 10-point difference in your yield, permanently. Check it before you check anything else.
  • Consensus standing — validators that publish weights out of line with the network get penalised, and so do their delegators. Long-running validators with stable standing are worth a slightly higher take.
  • Uptime — a validator that misses evaluation windows earns nothing during them.
  • Concentration — spreading across two or three validators costs you nothing and removes single-operator risk.

Unstaking: the part people get wrong

Unstaking is not instant liquidity at the price on your screen. You are selling alpha into a pool, so the size of your exit relative to the pool determines your fill. On a subnet with a $5M cap and $200K of daily volume, unwinding a large position moves the price against you before you are done. This is why turnover-to-market-cap is a risk metric, not a curiosity.

Three common mistakes

  1. Chasing the highest APY. The top of the yield table is usually the thinnest, most volatile end of the subnet set. The yield is compensation for that.
  2. Ignoring compounding frequency. Rewards accruing continuously and left in place compound far more aggressively than the headline rate implies. Model it rather than guessing.
  3. Treating a delegation like a savings account. It is a leveraged position on one subnet's alpha token that happens to pay a coupon.

If you want to see the difference compounding frequency makes on your own numbers, the calculator runs the standard future-value formula and shows the reward split explicitly.

TAO Compound is an analytics tool, not financial advice. Subnet figures referenced above come from the sample data set in this build; the live TAO price shown elsewhere on the site is from CoinGecko.