S

Synthdata

SN50

A decentralized probabilistic forecasting network built on Bittensor

Performances

RD Popularity Index

293.51% (24h)
0100200300400500
Project Rank: 440/12927

RD Growth Index

3570.00% (24h)
0400800120016002000
Project Rank: 1108/12927
Comparison:
1h24h7d30d60d90d
0.00%2.81%13.52%47.57%40.90%--

Comparison

Tokenomics

Unlock Progress

Unlocked5M SN50(26.81%)
Locked15M SN50(73.18%)
Untracked21M SN50(100%)

Data source: CoinGecko

SN50 is the token of Bittensor subnet 50 (Synth), incentivizing miners and validators via TAO emissions, and is used for synthetic financial-data API access and subnet governance.

Follow Updates

Follow Lists

People

About Synthdata

Synthdata provides probabilistic financial forecasts: rather than a single price target it outputs a full probability distribution for an asset's price over a horizon, covering Bitcoin, Ethereum, US equities and gold, updated on one-hour and 24-hour cadences and delivered by API to trading desks and quant teams.

It targets the point-estimate habit of conventional forecasting, where trading and risk need distributions rather than one number and building the model in-house is hard. The data is produced by competing model teams on a Bittensor subnet, scored on forecast accuracy, with only top performers reaching the API. The limit is the demand side: customers are few and output is heavily subsidized by network incentives.

Updated: Oct 30, 2026

Synthdata (Bittensor Subnet 50) has passed several key milestones since launch. It initially supported only BTC with 100 simulated paths and a 24-hour forecast window. In November 2025, paths were expanded to 1,000 with ETH, SOL, and XAU added, marking a shift toward high-frequency trading. In January 2026, tokenized equities were introduced; in March, XRP, HYPE, and WTIOIL were added. In June 2026, the competition was split into three tracks: Crypto 1h, Crypto 24h, and Commodities/Equities 24h. In July 2026, price feeds migrated from Pyth to Binance spot and Hyperliquid perps, with SPYX replaced by SP500. For real-world validation, Synth achieved 110% profit over four weeks on Polymarket using a $2,000 test account, executing ~12,000 trades totaling $500,000 in volume, proving its probabilistic models' market edge. Synth HFT followed, shortening the forecast horizon to 1 hour with updates every 2 minutes. The API went commercial with Standard ($49/month) and Pro ($199/month) plans. Most recently, Synth Ultra (ultra-low-latency competition) and the Synth Metamodel were introduced.

Updated: Sep 4, 2026

Buyback: The first purchase closed on July 2, buying 9,965.64 SN50 at an average of $1.50; the $100,000 commitment is spread over eight weekly buys and is funded from second-quarter profit rather than outside capital.

Revenue and progress: The subnet was profitable in Q2 and ranks in Bittensor's top 10 by revenue; on July 9 it published its first paper, fine-tuning Amazon's Chronos-2 and running it as a live miner, where it placed 12th in the 24-hour crypto competition. In October SN50 traded near $1.08 for a $6.2 million market cap, below its $1.90 cost basis.

Updated: Oct 30, 2026

Synthdata's Synth (Bittensor Subnet 50) published its first research paper on July 9, 2026, fine-tuning Amazon's Chronos-2 time-series foundation model and deploying it as a live miner scored by validators, then launched a miner analytics dashboard on July 16. On August 18 its HFT team released a study showing the predictive correlation between Kalshi's 15-minute BTC contracts and subsequent moves on Binance and other CEXs rose from 0.036 in January to 0.173 in August. In Q1 it also shipped Synth LLM for natural-language interaction and secured enterprise data partnerships with Mode and others.

Going forward, Synthdata's Synth plans to keep cutting latency across its API pipeline so miner data flows faster from the meta model to distribution, pursue academic partnerships with ML teams to raise the ceiling on forecast quality, and pitch Synth LLM as a proof of concept to major LLM partners, wrapping its probabilistic market intelligence in a conversational interface. It also aims to broaden asset coverage and strengthen trade-ready data endpoints that AI agents can call directly, cementing its on-chain probabilistic intelligence layer position.

Updated: Oct 30, 2026

James Ross — Co-founder; also founder of Mode Network, leads Synth, a price-distribution forecasting subnet.

Amber Shi — Chief Operating Officer.

Sam Hyatt — Quantitative modelling; co-author of the Synth whitepaper.

Stanislav Parkhomenko — Engineering; key contributor to the subnet codebase.

Updated: Oct 30, 2026