τ AlphaCheckCheck a subnet's alpha before you stake

Data as of (block 9307813)

SN51 lium.io: is the APY real?

Compare SN51 with another subnet

As of 2026-10-18 (block 9307813), subnet 51 (lium.io) shows an estimated 26.5–32.3% APY depending on validator, paid in its alpha token (29.1% at the stake-weighted validator take of 10.05%). The pool holds 178,522 τ, so staking 100 τ costs about 0.11% in entry slippage. Held 90 days at a flat alpha price, compounded daily, you'd get back about 106.26 τ after the swap fee and exit slippage, which beats root staking (≈101.5 τ). Change the inputs below to check your own case.

Report a wrong number (opens a prefilled GitHub issue)

TAO coverage: 61.5% · What this means

Break-even days

Days for this stake's own compounding alpha rewards, valued at today's alpha price, to cover the entry cost (stake slippage plus the 0.05% swap fee) if alpha does not move against TAO: ln(1 + entry cost ÷ stake after entry) ÷ ln(1 + daily rate). Stake-weighted validator take 10.05%. Never means zero rewards or a stake larger than the pool; over 365 days shows as > 1 year. How it's defined.

10 τ staked0.8 days
100 τ staked1.52 days
1,000 τ staked8.71 days

Across the validator APY range (26.5–32.3%), 100 τ breaks even in 1.39 days to 1.65 days.

Exit slippage, shown separately: selling the alpha from 100 τ into today's pool costs about 0.11% (0.106 τ).

Max safe stake

Largest stake that keeps entry slippage at or below each limit in the current pool.

≤ 0.5% entry slippage807.15 τ
≤ 1.0% entry slippage1,713.31 τ
≤ 2.0% entry slippage3,553.36 τ

Share image for this subnet at 100 τ, with the estimated APY, break-even days, and 30-day net TAO filled in.

What if?

Drag to see APY, slippage and break-even update below. Same estimates as the inputs above.

Share image

How it works, in plain English

Root staking (subnet 0) — you stake TAO and are paid in TAO. No alpha price risk, usually lower yield.

Subnet staking — your TAO is swapped into the subnet's alpha through a pool. Rewards are paid in alpha. When you leave, alpha is swapped back to TAO at whatever the price is then.

Slippage — each subnet pool is a weighted balancer. With equal weights it behaves as a constant-product pool (TAO × alpha = k). A swap fee applies on the way in and on the way out. Big trades in small pools move the price against you.

FAQ

Is subnet staking APY real profit?

Only in alpha terms. Your profit in TAO depends on the alpha/TAO price when you unstake. A 60% APY with a 50% alpha drop is still a loss.

What's the difference between root and subnet staking?

Root stakes TAO and earns TAO from a share of all subnets' validator emissions. Subnet staking holds alpha and earns alpha from one subnet.

How is the APY estimated?

Alpha emitted per block × 41% validator/staker share × (1 − root proportion) × (1 − validator take) × 7,200 blocks/day × 365, divided by alpha staked on the subnet. The default take is the stake-weighted average from on-chain Delegates (take ÷ 65535). You can pick a listed validator or type a custom take. Root APY uses each subnet's own average take, sells one epoch of root alpha through that subnet's pool (slippage and the swap fee), and divides the TAO by root stake. Holding-period results compound once per epoch. These are estimates; yields change every block. See Methodology.

How is slippage calculated?

Spot price is (base weight ÷ quote weight) × (TAO reserve ÷ alpha reserve), from the subnet's SwapBalancer weights. At equal weights, price = taoIn ÷ alphaIn. The swap fee is FeeRate ÷ 65535 (about 0.05%) on the TAO in and on the TAO out. Exit reserves are rebuilt at the scenario price while the weighted-pool invariant stays fixed.

Where does the data come from?

Directly from the Bittensor Finney chain via the public OpenTensor RPC — see the "as of" note above the results.

Weekly subnet changes, in plain English

Free, weekly, unsubscribe anytime. Powered by Buttondown.

All subnets

Not financial advice. Estimates only, for education. Yields, prices and liquidity change constantly. Do your own research.