TAO is now trading live on Raydium, Solana's leading DEX, confirmed by an Opentensor retweet and already being paired against new Solana memecoins. Behind that milestone, community members are pushing back on the Arbos monitoring tool over missed and misclassified subnets, while SN61 racks up enterprise pilots and one operator floats eye-catching subnet revenue numbers.
Full coverage
5 reports
01 / DAILY
TAO Goes Live on Raydium
TAO is now tradable on Raydium, a Solana-based decentralized exchange, per an Opentensor retweet. It's already showing up as a pairing token for new Solana memecoins alongside assets like ZCash and Hyperliquid. A community observer described Solana-side activity around TAO as strong.
02 / QUICK HIT
Community members flagged that the Arbos monitoring tool misses some subnets entirely -- including SN53 -- while misclassifying other legitimate subnets as problematic. Separately, subnets that fixed flagged issues say their emissions still haven't been restored. One community member added that some top-ranked subnets keep collecting emissions despite minimal GitHub activity, while newer, actively-developing subnets get none -- calling it a barrier to entry for new teams.
Arbos fails to detect some subnets, including SN53, and misclassifies others
Subnets that resolved flagged issues report no emissions restoration
Claim: low-activity top subnets keep emissions while active new subnets get none
03 / DAILY
SN61 RedTeam Racks Up Pilots
Opentensor's Novelty Search series spotlighted SN61 (RedTeam), which turns miners into a network of ethical hackers defending against bot attacks. The subnet reported five new enterprise pilots and teased an upcoming 'Immune System' feature. A livestream on SN61 RedTeam is scheduled for Thursday.
04 / DAILY
Subnet Revenue Numbers Floated
markjeffrey said several Bittensor subnets are already generating $5M-$10M in annual revenue, with some on track to hit $40M ARR by year end.
05 / DAILY
Miner Inference Gains on H200
const_reborn shared Pareton_ai results showing miners achieved a 3.5x speedup on Qwen3.8-27B-FP8 inference versus stock vLLM on a single H200 GPU. Median request latency dropped from 727ms to 190ms, throughput climbed from 58.7 to 221.8 tokens/sec, and SLA compliance jumped from 3% to 100%.