--- title: "Bittensor" publisher: Khala Research canonical_url: https://www.khala.io/bittensor-network source_reports: [https://www.khala.io/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow, https://www.khala.io/bittensor-the-intelligence-olympics] --- # Bittensor *An open network where independent subnets compete to produce AI work for $TAO, a token with a 21M hard cap that was mined from genesis with no pre-mine.* Company coverage from "Bittensor: The Investment History" (June 11, 2026): https://www.khala.io/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow#era-2-dtao-2023-2025 Cite as: Khala Research, "Bittensor", company coverage. https://www.khala.io/bittensor-network ## At a glance - **Category:** Open network for AI subnets - **Founded:** 2019, by Jacob Steeves and Ala Shaabana - **Launched:** January 2021 (Kusanagi), no pre-mine - **Supply:** 21M hard cap; first halving December 2025, 7,200 to 3,600 $TAO a day - **dTAO:** Live February 13, 2025; one alpha token per subnet - **Subnets:** 128, the cap, reached September 2025 - **Institutional products:** Grayscale GTAO (OTCQX, Dec 2025; spot ETF S-1 pending), Safello STAO ETP - **Largest subnets:** Chutes (SN64), Targon (SN4), lium (SN51) by market cap (Jun 2026) ## 01 What is Bittensor? Bittensor is an open-source network for AI work, launched in January 2021 as an alternative to closed-lab AI infrastructure. Independent subnets each run a competition: miners produce outputs, validators score them, and the best performers earn tokens. Every $TAO was mined from genesis. There was no pre-mine, no team allocation and no venture round at launch. Between 2021 and 2023 the founding team sold roughly 600,000 mined $TAO over the counter to Firstmark, DCG and Polychain at about $18 per token, raising $10.8M. [The investment case in the report →](https://www.khala.io/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow#executive-summary) ## 02 How did Bittensor get here? - **March 2020:** the Bittensor whitepaper is published. - **January 2021:** Kusanagi mainnet launches; block production stops in May 2021 after a consensus failure. - **November 2021:** the network relaunches as Nakamoto from block 0, migrating 546,113 mined $TAO. - **March 2023:** Nakamoto forks to Finney, still the live chain; subnets activate that October. - **February 2025:** dTAO goes live; Coinbase lists $TAO on February 20. - **December 2025:** the first halving and the start of GTAO trading. - **May 2026:** the Conviction Upgrade auto-locks subnet owner emissions. [Full timeline →](https://www.khala.io/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow#timeline) ## 03 How do subnets and dTAO work? Before dTAO, the top 64 validators by stake decided how 7,200 daily $TAO in emissions were split, and they often favored subnets they ran or held. Dynamic TAO, proposed by Jacob Steeves in January 2024 and live on February 13, 2025, gave every subnet its own liquidity pool and alpha token. Staking $TAO into a subnet is a swap into its alpha, so emissions follow market demand rather than validator votes. Subnets grew from 82 in April 2025 to the 128 cap by September 2025. When the cap is full, a new registration removes the subnet with the lowest moving-average price, once its four-month immunity period has passed. [dTAO in the report →](https://www.khala.io/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow#era-2-dtao-2023-2025) ## 04 How are $TAO emissions split? The protocol emits 0.5 $TAO per block, and each subnet mints up to 1 alpha per block. A subnet's share of $TAO emissions is its smoothed net $TAO inflow relative to all subnets; negative net flow earns nothing. Alpha rewards are paid every 360 blocks (72 minutes): 41% to miners, 41% to validators and 18% to the subnet owner. Supply is capped at 21M on a Bitcoin-style schedule. The first halving, at block 7,103,976 in December 2025, cut daily issuance from 7,200 to 3,600 $TAO. [The dTAO emission model →](https://www.khala.io/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow#era-2-dtao-2023-2025) ## 05 Which subnets have real revenue? - **Targon (SN4):** a confidential GPU marketplace generating $10.4M in ARR; Venice AI trained a model on it and Dippy AI routes roughly half its text inference through it. - **Chutes (SN64):** serverless inference running at $5.5M+ annualized as of February 2026. - **Score (SN44):** vision AI with signed customers including Reading FC, AVIA (3,000 petrol stations in 15 countries) and Lavance, plus a PwC France alliance from April 2026. - **Vanta (SN8):** a decentralized prop trading network that reported $100K in monthly fees against $52K in miner emissions in March 2026, $48K net. The reports selected these as standouts. The February 2026 report profiled 5 of 128 subnets and said not all will succeed. [Subnets with external revenue →](https://www.khala.io/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow#era-2-dtao-2023-2025) ## 06 How has $TAO performed? $TAO opened at about $91 on its first exchange listings in March 2023, fell to about $30 that May, peaked at $757 in March 2024 and traded near $210 in June 2026, a $2B market cap. On October 17, 2026 it traded at $274.93, a fully diluted valuation of $5.77B. From March 2023 to June 2026 it returned +187%, behind NVDA, SMH and BTC, and it was the most volatile asset in the comparison. What a holder earned depended mostly on when they bought. **Exhibit: Asset returns comparison** *Cumulative return from March 2023, when $TAO listed on exchanges, to June 2026.* - NVDA, +889% - SMH, +414% - BTC, +202% - $TAO, Peak near +700%, +187% - QQQ, +160% - ARKK, +112% - BOTZ, +74% - ETH, +20% - PYPL, −38% - ENPH, −70% $TAO was the most volatile asset shown. It peaked near +700% in early 2024 before falling back to +187%. *Data: [CSV](https://www.khala.io/data/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow/returns.csv) · [JSON](https://www.khala.io/data/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow/returns.json)* [Returns in the report →](https://www.khala.io/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow#closing) ## 07 What institutional products exist? The Grayscale Bittensor Trust (GTAO) began as a private placement in June 2024, started trading on OTCQX on December 11, 2025 and filed an S-1 for spot ETF conversion on December 30, 2025. Approval is pending. It held $11.7M at March 31, 2026, traded at a 59.9% premium to NAV, and was down to $9.4M by June. Other vehicles include the Safello STAO staked ETP (about $3M), TrustedStake's non-custodial subnet indexes ($4.7M across 2,232 nominators), and Yuma, Stillcore and Unsupervised Capital, which do not disclose AUM. Disclosed assets are small against a $2B market cap. **Exhibit: GTAO: the road to a spot ETF** *Grayscale Bittensor Trust net assets, from first report through its OTCQX debut and the spot ETF S-1 filing, as of May 2026.* - $1.5M, Aug 2024 - $4.4M, Dec 2024 - $7.97M, Dec 2025 - $11.3M, May 2026 NYSE Arca conversion is pending. The trust ran as a private placement until its OTCQX debut, then traded OTC alongside the reopened private placement. *Source: [Grayscale](https://www.grayscale.com/funds/grayscale-bittensor-trust), [SEC filings (S-1)](https://www.sec.gov/Archives/edgar/data/2029297/000119312525335992/tao-20251230.htm). Data: [CSV](https://www.khala.io/data/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow/bth2-f7.csv) · [JSON](https://www.khala.io/data/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow/bth2-f7.json)* [Structured capital →](https://www.khala.io/bittensor-an-investment-history-from-genesis-to-dtao-tao-flow#4-1-structured-capital) ## 08 What are the main risks? - **Builder exits:** Covenant AI, the team behind Templar (SN3), Basilica and Grail, left in April 2026, accused Jacob Steeves of centralized control and sold about 37,000 $TAO, setting off an 18 to 27% price drop. - **Security:** a compromised PyPI package drained about 32,000 $TAO (about $8M) from operator wallets in July 2024; the chain itself was not compromised. - **Volatility:** $TAO peaked near +700% before falling back to +187% over the same period. - **Inflation and fragmentation:** $TAO issuance may not be offset by subnet growth, and expanding from 128 to 256 subnets would spread liquidity thinner. - **Competition:** centralized AI labs compete with every subnet. [Catalysts and risk factors →](https://www.khala.io/bittensor-the-intelligence-olympics#catalysts-and-risk-factors) ## 09 What should we watch next? - **GTAO spot ETF:** NYSE Arca conversion is pending. - **Subnet revenue:** Chutes' revenue and its organic-to-sponsored ratio, and contract renewals and ARR disclosures at Score and MIID. - **Nova (SN68) lab results:** wet-lab validation of compounds found by miners. - **Subnet owner exits:** unlocks under the Conviction Upgrade are now public onchain transactions. [Catalysts →](https://www.khala.io/bittensor-the-intelligence-olympics#catalysts-and-risk-factors) ## 10 Primary sources - [Bittensor whitepaper (arXiv)](https://arxiv.org/abs/2003.03917) - [dTAO whitepaper](https://bittensor.com/dtao-whitepaper) - [BIT001: Dynamic TAO](https://medium.com/@unconst/bit001-dynamic-tao-8ddc7a26dd62) - [TAO token economy explained (Bittensor blog)](https://blog.bittensor.com/tao-token-economy-explained-17a3a90cd44e) - [Emissions documentation](https://docs.learnbittensor.org/learn/emissions) - [Taostats subnets](https://taostats.io/subnets) - [Grayscale Bittensor Trust](https://www.grayscale.com/funds/grayscale-bittensor-trust) - [GTAO S-1 (SEC)](https://www.sec.gov/Archives/edgar/data/2029297/000119312525335992/tao-20251230.htm)