Yuma Launches Subnet Index (YCX) | Weekly Bittensor Update
Yuma launched the Yuma Composite Index (YCX), the first benchmark tracking the full universe of Bittensor subnets. Chutes (SN64) restructured its subscription model, Synth (SN50) released a 1-hour forecast endpoint for Polymarket traders, and Bitcast (SN93) reported a 55% surge in subscribers. IOTA (SN9) upgraded its Train at Home scoring system, Ridges AI (SN62) outlined its agentic expansion roadmap, and GroundLayer launched on SN20 as a new OTC capital layer. TAO closed the week at $167-$184, up 4.5%.
Big Events in the Bittensor Ecosystem
Yuma Launches Yuma Composite Index (YCX) to Track Subnets on Bittensor
Yuma has launched the Yuma Composite Index (YCX), a new benchmark designed to track the full universe of subnets on Bittensor. The index gives investors a clear and structured way to monitor aggregate subnet market price performance. As the Bittensor ecosystem expands, demand for reliable market indicators continues to grow.
Bittensor operates as a decentralized machine learning network where independent subnets function as specialized AI marketplaces. Each subnet has its own economic dynamics and token performance, making broader market tracking complex and time-consuming. YCX simplifies this — instead of analyzing each subnet separately, users can follow one consolidated benchmark for overall subnet market trends.
Why the Yuma Composite Index Matters
As more subnets launch, analytical complexity increases. Investors must compare multiple tokens, monitor price volatility, and assess performance across sectors. YCX acts as a sector-wide performance indicator, similar to how composite indices in traditional finance track the health of a market segment.
The index serves both institutional and retail participants. Institutional investors require benchmarks to evaluate capital allocation, while retail investors seek simplified metrics to understand market direction. In both cases, YCX provides a single reference point.
Strengthening the Financial Infrastructure of Bittensor
The launch of YCX signals further maturation of the Bittensor ecosystem. As decentralized AI networks attract more capital, structured financial tools become essential. YCX enhances comparability across subnets, allowing investors to measure performance trends without relying solely on individual token movements.
Over time, benchmarks like YCX may play a central role in how investors evaluate decentralized AI markets. The Yuma Composite Index positions itself as a key tool for tracking the evolving subnet economy on Bittensor.
HashiChain (SN115) Whale Stake Exposes dTAO Risks on Bittensor
On February 25, 2026, a single wallet staked roughly 12,000 TAO, worth about $2 million at the time, into Bittensor Subnet 115, known as HashiChain. The transaction triggered one of the largest single-day moves in dTAO history. Within hours, the HashiChain alpha token surged more than 1,000%, briefly climbing into the top subnets by emission share and capturing around 11–12% of total Bittensor emissions.
HashiChain positions itself as a Layer 1 blockchain built for AI agents, aiming to provide infrastructure where autonomous programs can transact and coordinate. However, at the time of the spike, the project had no public demo, no working testnet, and no independent audit. Because SN115 operated with very low liquidity, the large stake caused extreme slippage and rapid price acceleration. The move quickly attracted attention across the Bittensor community, with on-chain analysts flagging it as one of the biggest single entries into any dTAO subnet.
The reaction split into two camps. Some users highlighted SN115 as the top-performing subnet of the day, pointing to gains exceeding 1,000%. Others raised serious concerns about transparency and manipulation risk. Several experienced community members described the event as a stress test for the dTAO system, arguing that low-liquidity subnets remain vulnerable to concentrated capital moves. The spike also created collateral pressure, as capital rotated away from other subnets, leading to short-term price drops elsewhere in the ecosystem.
Two days later, volatility continued when SN115 saw roughly 13,000 TAO exit in a single transaction. The alpha token corrected sharply, reinforcing concerns about speculative flows and liquidity fragility. Community voices responded with both criticism and constructive analysis. Some framed the episode as a warning to focus on fundamentals such as active GitHub development, transparent leadership, miner and validator participation, sustainable on-chain revenue, and delivered roadmaps rather than price momentum alone.
The HashiChain event also highlighted the importance of TAOFLOW, Bittensor’s updated emission model introduced in November 2025. Under the previous system, subnet emissions closely tracked alpha token price, which incentivized price pumping. TAOFLOW changed that structure. Emissions now depend on net staking inflows rather than price alone, and the system applies an approximately 86.8-day exponential moving average to smooth staking data. As a result, a single large stake does not guarantee sustained emissions. Subnets must maintain continuous net positive staking over time to secure meaningful rewards.
In practical terms, the SN115 episode demonstrated both the flexibility and the limits of the dTAO design. A whale can still move price dramatically in a low-liquidity subnet. However, under TAOFLOW, long-term emissions require durable capital commitment, not short-term speculation. For the Bittensor ecosystem, HashiChain became a real-time case study in liquidity dynamics, incentive alignment, and the evolving resilience of the network’s economic layer.
Subnet Updates
Ridges AI (SN62) Shares Roadmap Update
Ridges AI (SN62), a subnet operating within the Bittensor ecosystem, has outlined its next phase of development after a joint conversation with SubnetSummerTAO. The discussion highlighted four strategic priorities that define where SN62 is heading, with a strong focus on agentic expansion, tighter product alignment, continuous performance improvements, and a developer-first open beta approach.
- First, SN62 plans to expand beyond coding pull request competitions into broader agentic challenges. The roadmap includes developer-adjacent tasks such as smart contract auditing, followed by more advanced benchmarks like personal assistant-style agents. This evolution reflects a wider trend within Bittensor, where subnets increasingly compete on complex, real-world AI capabilities rather than narrow technical tasks.
- Second, the team is strengthening alignment between subnet incentives and real product performance. The ongoing open beta is designed to tighten this feedback loop. Upcoming improvements include synthetic evaluations to reduce overfitting, scoring models that measure both cost and time efficiency, and additional quality controls such as linting and formatting. As a result, SN62 aims to reward performance that translates directly into production-grade outcomes.
- Third, synthetic evaluations will enable continuous performance gains. Because they provide effectively unlimited test coverage, they create a constantly rising performance bar. This approach reduces stagnation and expands room for innovation. In addition, SN62 is integrating Harbor and other benchmark suites to diversify evaluation frameworks and improve robustness within the competitive Bittensor environment.
Finally, Ridges AI is emphasizing a developer-first open beta experience. Access is streamlined through Google sign-in and repository authorization, allowing participants to begin testing quickly. The current beta phase focuses on aligning subnet agents with product goals, improving code quality, optimizing inference cost, increasing execution speed, and refining the user interface and overall developer experience.
Taken together, these updates signal a more mature and product-oriented direction for SN62 within Bittensor. By combining agentic expansion with structured evaluation and incentive alignment, Ridges AI is positioning its subnet for sustained competitiveness in the decentralized AI ecosystem.
Chutes (SN64) Announces Key Updates to Public
Chutes has announced several important updates to its public offerings as part of a broader effort to improve platform stability, sustainability, and long-term scalability. The changes reflect growing infrastructure demands and the increasing computational intensity of modern AI models. As a result, the team is restructuring elements of its access and subscription framework to ensure consistent performance for users.
One of the main changes involves the retirement of the Early Access program. Previously, the program offered a limited number of free daily requests. Access to certain high-security models has already been discontinued, while remaining free usage will be phased out. Eligible users will be able to transition to a Base subscription for one month at no cost or receive account credit as part of the transition process.
In addition, Chutes is updating its subscription usage structure. The previous request-based, token-agnostic model has become increasingly difficult to sustain due to rising compute costs. Moving forward, subscription tiers will include defined usage limits aligned with Pay-As-You-Go value thresholds. These caps are designed to ensure fair resource allocation while maintaining overall platform reliability. According to the company, most users are unlikely to experience significant changes, although high-volume accounts may face tighter usage constraints.
The platform is also adjusting model availability within certain subscription tiers. Some of the most resource-intensive models will no longer be included in lower-tier plans. This change aims to improve system responsiveness and reduce performance bottlenecks for the broader user base.
Chutes emphasized that these updates are part of an ongoing evolution focused on strengthening infrastructure and improving user experience. As AI models continue to grow in size and complexity, the company is adapting its operational model to meet those demands while preserving platform quality. Further refinements are expected as the platform continues to scale.
Synth (SN50) Introduces 1-Hour Forecast for Minute-Level Polymarket Edge
Synth (SN50) has launched a new 1-hour forecast endpoint that delivers minute-by-minute probability distributions for short-dated crypto contracts on Polymarket. The update is designed to improve precision in 15-minute and hourly binary markets, where outcomes resolve to $1 or $0 and profitability depends on estimating probabilities more accurately than the market price.
Unlike the standard 24-hour forecast, which returns price distributions at five-minute intervals, the 1-hour model provides 61 timesteps, covering every minute from 0 to 60. This higher resolution is critical for short-duration contracts. In a 15-minute market, a 24-hour forecast produces only three data points. By contrast, the new 1-hour endpoint allows traders to evaluate probability changes at each minute of the contract’s lifecycle.
The forecasting engine operates on Bittensor, where miners submit competing crypto price prediction models. These models are scored on accuracy, and the strongest performers are combined into a meta-forecast. To date, Synth has distributed more than $2.8 million to miners. Importantly, the 1-hour forecast is scored specifically on one-hour accuracy using a 1,000-path simulation, which means it is not simply a zoomed-in version of the 24-hour model. Different time horizons reward different modeling strategies, and those differences can produce meaningful probability gaps.
The endpoint returns nine percentile bands, ranging from the 0.5th to the 99.5th percentile, at each timestep. Traders can extract the relevant probability at minute 15 for 15-minute contracts or minute 60 for hourly contracts. In addition, Synth provides direct comparisons between its probability estimates and Polymarket pricing, making it easier to identify discrepancies that may represent trading edge.
Beyond terminal probabilities, the full distribution also describes the expected price path. Bots and agent frameworks can track how the distribution shifts as expiry approaches, analyze band widths to assess volatility, or use percentile spacing to estimate tail risk. Moreover, combining the 1-hour and 24-hour forecasts can offer layered insight, since each model is optimized for a different evaluation horizon.
With this release, Synth strengthens its role within Bittensor by translating competitive miner models into actionable, high-resolution probability intelligence. For traders operating in short-dated crypto markets, the 1-hour forecast introduces a measurable structural advantage through increased granularity and time-horizon-specific optimization.
Bitcast (SN93) Reports Rapid Growth as Subscriber Base Surges 55% Month-over-Month
Bitcast (SN93) continues to accelerate its expansion within the Bittensor ecosystem, reporting strong month-over-month growth across key performance metrics. Compared to January, the network has reached 1.4 million subscribers, marking a 55% increase. At the same time, total hours watched climbed to 47,000, up 32%, while views rose to 765,000, reflecting a 42% gain.
The growth trend signals more than incremental traction. Subscriber expansion remains the primary driver, reinforcing a compounding effect across reach and engagement. As the audience base scales, content visibility strengthens, and interaction rates increase accordingly. This dynamic creates a reinforcing loop where attention attracts creators, and creators further expand audience reach.
In parallel, Bitcast has published 611 videos to date, with production steadily increasing. Engagement metrics also remain strong, including 34,000 likes and 8,200 shares, both showing consistent upward momentum. The network now reaches audiences across 54 countries and operates in seven languages, highlighting growing international penetration.
According to the team, the current trajectory positions Bitcast to support multiple enterprise brands simultaneously. As scale improves, so does the network’s capacity to manage larger marketing budgets and parallel brand campaigns. This shift indicates a transition from early growth to operational maturity.
Within the broader Bittensor landscape, SN93 demonstrates how creator-driven subnets can translate attention into scalable infrastructure. If current trends persist, Bitcast may emerge as a key media distribution layer capable of servicing enterprise-level demand while maintaining rapid organic growth.
IOTA (SN9) Releases Network Upgrade to Train at Home (TAH)
IOTA (SN9) has rolled out a new network release focused on improving Train at Home (TAH), introducing upgrades that enhance fairness, availability, and research flexibility within the Bittensor ecosystem. The update addresses several user concerns while reinforcing the robustness of distributed training infrastructure.
One of the most significant changes involves the scoring system. SN9 has refined contributor scoring across network layers to improve fairness and reduce asymmetries in how model work is evaluated. Backward activations are now incorporated into the contributor scoring ratio, ensuring that computational contributions are more accurately reflected. This adjustment improves balance across participants and strengthens incentive alignment.
The update also increases availability for Train at Home contributors. Previously, unreliable contributors were removed from active runs every six hours. The new release shortens this interval to one hour. As a result, new participation slots open more frequently, making TAH more accessible to incoming contributors. In addition, error handling for weight downloading has been improved, reducing friction during onboarding and synchronization.
On the research side, IOTA has implemented runtime control switches that allow weighted DiLoCo and gradient clipping to be toggled on demand. These changes provide greater experimental flexibility while maintaining system stability. By introducing configurable runtime controls, SN9 supports more advanced distributed training experimentation without compromising operational integrity.
Overall, the upgrade strengthens Train at Home by making it more fair, more available, and more technically resilient. Within the broader Bittensor framework, these improvements reinforce SN9’s role in advancing decentralized model training infrastructure while improving the contributor experience.
More info about Train At Home here.
New Subnet Launches
GroundLayer (SN20)
GroundLayer, built by Taoshi and Team Rizzo AI, is launching on Subnet 20 to create a capital layer for the Bittensor economy. The platform introduces an OTC marketplace for structured TAO deals with predefined discounts, lock-ups, and smart contract enforcement, all designed to avoid immediate impact on spot markets.
TAO Market Update
Price: $167.31-$184.01
Weekly: +12.26%
Ranking: #36
Market Cap: $2.03B
24h Volume: $128.93M
Top Gainer Subnet: Ridges (SN82)
Ridges is a decentralized AI software engineering platform designed to automate real-world programming tasks using autonomous coding agents. Operating as Subnet 62 within the Bittensor ecosystem, Ridges enables AI agents to write code, fix bugs, generate tests, and solve practical development problems without relying on centralized engineering teams or proprietary platforms.
Weekly change: +27.63%
Price: $8.593277
Market Cap: $1.91M
Volume (24h): $95.19K
