NOVA (SN68) Plans Biologics Push | Weekly Bittensor Update
This week, NOVA (SN68) stepped into a new chapter on Bittensor. Jacob Steeves, known as Const, hosted a Novelty Search call with Metanova Labs to outline the path to production drug candidates, biologics, and a new nanobody mechanism. Manifold Labs launched the Targon (SN4) Supply Portal, opening permissionless and managed earning paths for GPU operators on Bittensor. Synth (SN50) shipped its LLM for Traders across Polymarket, Hyperliquid, Limitless, and Deribit. Furthermore, Chutes (SN64) signed a non-binding framework with Nasdaq-listed AIxCrypto and rolled out Chutes Chat Plus at half the price of ChatGPT Plus. Hippius (SN75) crossed 100 TB of free egress. Affine (SN120) spun up a dedicated H200 cluster. Meanwhile, IOTA (SN9) outlined its Train at Home vision for consumer compute. Finally, Ridges AI (SN62) dropped a two-week roadmap for its coding agent subnet. Here is everything that happened.
Big Events in the Bittensor Ecosystem
Const and Metanova Labs Discuss NOVA (SN68) Drug Candidate Pipeline, Biologics, and New Nanobody Mechanism
On May 7, Jacob Steeves, known as Const, hosted the latest Novelty Search call on the Bittensor Discord. Special guests from Metanova Labs joined the conversation. The session focused on NOVA (Subnet 68), the decentralized drug discovery platform on Bittensor. The discussion covered three core topics. First, the team explored how to transition subnet submissions into production drug candidates. Second, they outlined the expansion from small molecules into biologics. Furthermore, they presented a new nanobody mechanism.
Overall, the roadmap signals a meaningful step forward for NOVA. The subnet launched on March 1, 2025 as the first decentralized drug screening platform on Bittensor. According to Metanova Labs, NOVA has already processed more than 4.6 million molecule submissions against 7,000 protein targets. The compound library exceeds 65 billion molecules.
Moving from virtual hit identification to production drug candidates would extend NOVA’s scope beyond pure screening. At the same time, the expansion into biologics opens a new therapeutic class alongside small molecules. Furthermore, the new nanobody mechanism introduces competitive on-chain incentives for one of the most active areas in modern antibody research.
Subnet Updates
Chutes (SN64) Enters Collaboration Framework with Nasdaq-Listed AIxCrypto for AI Agent Scaling
AIxCrypto Holdings, Inc. (NASDAQ: AIXC), a technology company developing AI agent and Embodied AI infrastructure, entered a non-binding collaboration framework with Chutes AI, the team behind Bittensor subnet SN64. The parties announced the arrangement on April 30, 2026. Under the framework, Chutes plans to provide AIxCrypto with access to its decentralized AI inference infrastructure, with integration targeting three core technical areas: model routing, real-time processing, and workload distribution across concurrent agent interactions.
Importantly, the framework is structured as an exploratory arrangement rather than a commercial deployment. Both sides will use it to test whether decentralized compute can support high-frequency real-time AI agent interactions, multi-agent concurrency, and latency-sensitive user-facing applications. Definitive commercial terms will follow separate agreements. A Nasdaq-listed counterparty evaluating subnet-level inference is a meaningful data point for the Bittensor ecosystem, regardless of how the framework develops.
Read the full breakdown of the Chutes (SN64) and AIxCrypto collaboration
Targon (SN4) Supply Portal Goes Live with Two Earning Paths for GPU Operators
Manifold Labs launched the Targon (SN4) Supply Portal, a unified onboarding interface for compute suppliers on Bittensor’s Subnet 4. The portal, accessible at supply.targon.com, lets datacenters and individual operators monetize idle GPU and CPU capacity. Operators choose between two routes: a permissionless path connecting directly to the Bittensor blockchain for SN4 alpha rewards, and a Targon Managed path with weekly fiat payouts for those who prefer abstracted token mechanics.
Both paths funnel hardware into the same confidential compute marketplace that powers Targon’s enterprise workloads. Supported configurations span AMD EPYC 9xx4 with SEV-SNP, Intel TDX paired with NVIDIA H100 and H200 GPUs, and Intel TDX with NVIDIA B200 in multi-GPU setups. Each setup runs the Targon Virtual Machine, providing end-to-end attestation for both host and GPU. According to Manifold Labs, the network now serves over 20 billion paid inference tokens daily across more than 1,500 H200 GPUs, with annual compute incentives exceeding $60 million.
Read the full breakdown of the Targon (SN4) Supply Portal launch
Synth (SN50) Launches LLM for Traders Across Polymarket, Hyperliquid, Limitless, and Deribit
Synth (SN50) launched Synth LLM for Traders, a conversational interface that converts live Monte Carlo forecasts into custom statistics, charts, and example trades. Initial coverage spans Polymarket, Hyperliquid, Limitless, and Deribit, with more venues planned. According to the team, this is the first conversational interface that translates probabilistic price-path forecasts directly into trader-ready outputs.
On the technical side, Synth runs on Bittensor Subnet 50, where over 200 machine learning models compete to produce 1,000 simulated price paths for each asset across 1-hour and 24-hour horizons. Validators score miners using the Continuous Ranked Probability Score (CRPS). Forecast coverage spans Bitcoin, Ethereum, Solana, and gold (XAU). The launch builds on the team’s public @SynthDataDotCo account on Polymarket, which has been trading the BTC Hourly Up/Down market since October 3, 2025.
Hippius (SN75) Serves 100 TB of Free Egress in a Single Month
Hippius (SN75) announced that the network served 100 TB of egress during the past month at no bandwidth cost to users. According to the team, the same volume would have generated roughly $8,000 in AWS S3 bandwidth fees. Hippius offers free egress as a permanent feature of its pricing model, positioning the subnet as a lower-cost alternative to centralized cloud storage providers.
Affine (SN120) Builds Dedicated H200 GPU Cluster on Targon Compute
Affine (SN120) announced the launch of a dedicated GPU cluster built on Targon Compute to improve the mining experience. The cluster runs NVIDIA H200 GPUs, with two H200s per miner across eight slots at launch and additional capacity planned. According to the team, inferencing on the Affine arena now runs through the new cluster, and a full migration will follow.
Chutes (SN64) Launches Chutes Chat Plus, a Multi-Model Bittensor Alternative to ChatGPT
Chutes (SN64) launched Chutes Chat Plus, a $10-per-month plan offering access to 60+ open-source models in a single Bittensor-powered interface. Notably, the new plan undercuts ChatGPT Plus at $20 per month while offering model selection across tasks rather than a single default. According to the team, users can swap models inside one conversation as they switch between reasoning, coding, and multimodal tasks.
The lineup includes Mimo V2 Flash for general reasoning at $0.09/$0.29 per million tokens. MiniMax M2.5 handles code at $0.15/$1.20, while Kimi K2.6 covers image input and a 262K context window at $0.95/$4.00. Furthermore, Chutes offers TEE models for hardware-level security and privacy, with no prompt logging and no training on user data. The product runs on the Bittensor decentralized inference network through Chutes Subnet 64, available at chutes.ai/chat.
IOTA (SN9) Targets Mac Minis and Consumer Devices for Distributed Bittensor Training
IOTA (SN9), the distributed AI training subnet on Bittensor developed by Macrocosmos AI, is exploring how idle consumer hardware can plug directly into its training architecture. In a discussion with EyeOn_AI, @macrocrux from Macrocosmos highlighted the role of Mac Minis in IOTA and the Train at Home initiative. According to @macrocrux, Moltbot users can plug their underutilized hardware into the network and earn passive income through training contributions.
The framing connects two trends. First, consumers are stockpiling Mac Minis to run agents privately around the clock. Furthermore, IOTA can harvest that compute when it sits idle, plugging it into a distributed training pipeline. Effectively, Bittensor’s training subnet positions itself to absorb a wave of consumer hardware bought for AI inference but rarely run at full utilization.
Ridges AI (SN62) Outlines Q2 2026 Roadmap for Bittensor’s Coding Agent Subnet
Ridges AI (SN62) published a two-week roadmap covering four planned upgrades to its Bittensor-powered coding agent subnet. First, BYOK (Bring Your Own Key) will let miners supply their own inference key and cover their own costs during evaluations. According to the team, the change improves cost accuracy and puts more skin in the game.
Furthermore, screener auto-scaling will let screener nodes expand and contract with demand, improving throughput and removing a known bottleneck. Variance reduction is the third upgrade. The team plans to increase problem counts and adjust scoring between runners, producing more accurate rankings with less noise. Finally, Infinite SWEBench introduces a synthetic benchmark pipeline built on real GitHub pull requests, replacing public datasets that miners can game. Ridges plans to ship all four within a tight two-week window, with more updates to follow on the Bittensor ecosystem’s coding agent infrastructure.
TAO Market Update
Price: $265.78 – $318.81
Weekly: +15.84%
Ranking: #30
Market Cap: $3.39B
24h Volume: $275.987M
Top Gainer Subnet: MVTRX (SN79)
MVTRX (SN79) is a Bittensor subnet building a decentralized trading exchange for dTAO alpha tokens. Notably, the project pairs the exchange with a large-scale agent-based market simulation written in C++. The simulation produces Level 3 limit-order book data that mimics real-world asset markets. Miners act as trading agents, while validators score them on the intraday Kappa-3 ratio. On the exchange side, MVTRX applies a Dynamic Incentive Structure (DIS). Furthermore, the structure rewards liquidity provision through zero-sum rebates and dynamically adjusting fees. Together, these layers position the subnet at the intersection of HFT research and on-chain dTAO trading on Bittensor.
Weekly change: +47.62%
Price: $3.670698
Market Cap: $14.42M
Volume (24h): $1.18M
Most of subnets featured in this update are listed on SimplyTao. The platform supports fiat and crypto purchases with no wallet configuration needed.
Disclaimer: Digital asset prices are subject to high market risk and price volatility. The value of your investment may go down or up, and you may not get back the amount invested. You are solely responsible for your investment decisions and SimplyTao is not liable for any losses you may incur. Past performance is not a reliable predictor of future performance. You should only invest in products you are familiar with and where you understand the risks.
FAQ:
On May 7, Jacob Steeves, known as Const, hosted a Novelty Search call on the Bittensor Discord with special guests from Metanova Labs. The session covered three core topics for NOVA (Subnet 68): the path from subnet submissions to production drug candidates, expansion from small molecules into biologics, and a new nanobody mechanism. The roadmap signals a meaningful step forward for the decentralized drug discovery subnet.
The Targon (SN4) Supply Portal is a new onboarding interface launched by Manifold Labs that lets compute suppliers monetize idle GPU and CPU capacity on Bittensor’s Subnet 4. Operators choose between two earning paths. The permissionless path connects directly to the Bittensor blockchain for SN4 alpha rewards, while the Targon Managed path issues weekly fiat payouts for those who prefer abstracted token mechanics.
Synth (SN50) LLM for Traders is a conversational interface built on Bittensor Subnet 50 that converts live Monte Carlo forecasts into custom statistics, charts, and example trades. Initial coverage spans Polymarket, Hyperliquid, Limitless, and Deribit, with more venues planned. According to the team, this is the first conversational interface that translates probabilistic price-path forecasts directly into trader-ready outputs.
Chutes (SN64) announced two key updates on Bittensor this week. First, the team entered a non-binding collaboration framework with Nasdaq-listed AIxCrypto Holdings to support AI agent infrastructure scaling through decentralized inference. Furthermore, Chutes launched Chutes Chat Plus, a $10-per-month consumer plan offering access to over 60 open-source models through a single Bittensor-powered interface.
MVTRX (SN79) is a Bittensor subnet building a decentralized trading exchange for dTAO alpha tokens, paired with a large-scale agent-based market simulation written in C++. Miners act as trading agents while validators score them using the intraday Kappa-3 ratio. The exchange applies a Dynamic Incentive Structure (DIS) that rewards liquidity provision through zero-sum rebates and dynamically adjusting fees.
