Launched Lending Architectures concept page (monolithic / modular / CeDeFi)
New concept page segmenting onchain lending into three architectures and matching each to the risk profile it's best suited for. Monolithic pools (Aave, Spark, Compound) for continuous price risk. Modular (Morpho, Euler, Kamino) for hybrid/jump risk. CeDeFi (Maple, Pareto, Wildcat) for credit risk that can't live in permissionless systems. April 2026 Kelp/Aave incident treated as the cleanest worked example of architectural trade-offs.
DeFi Risk Methodology integrates Layer 1/2/3 taxonomy + six failure modes
The methodology page now uses Anastasiia's Layer 1 (mechanical) / Layer 2 (governance) / Layer 3 (code integrity) taxonomy, with the dominance condition (Layer 3 must be cleared before Layer 1 optimization is meaningful). Six failure modes documented where TradFi credit analogies break in DeFi. Resolv reference case added as a worked example of oracle latency.
Morpho curator-layer analysis: TVL concentration, network centrality, tail correlations
New curator-layer section on the Morpho research page quantifying where the protocol's risk actually sits. ~$7.27B across 8 major curators (Gauntlet 27.6%, Steakhouse 17.8%, MEV Capital 12.6%, K3 Capital 6.6%, mid-tier ~35.4%). Network centrality finding: B Protocol and Block Analitica are the ecosystem hubs (eigenvector centrality ~0.46 and ~0.43) despite being mid-sized. Lower-tail drawdown correlations stay elevated where collateral overlaps.
Launched Vault Credit Risk concept page (five mechanical channels)
New concept page covering the five mechanical loss channels inside any lending vault: stress-adjusted coverage, recovery endogeneity, liquidity stress, oracle integrity, execution viability. Framework attributes to Anastasiia (@mathy_research) from the Vault Summit paper. Includes the additive/multiplicative Vault Credit Score and honest Layer 1/2/3 scope limits.
SEO discoverability sweep: 5 new concept pages surfaced to Google and AI crawlers
Learn Hub, sitemap.xml, feed.xml, and llms.txt updated to list the five new concept pages (Vault Credit Risk, Lending Architectures, ERM Valuation, RWA Composability, Solver Landscape). FAQPage JSON-LD added to Vault Credit Risk and ERM. Internal cross-links between concept pages tightened.
Maple tagged as CeDeFi Lending architecture
Added a visible architecture tag to Maple's header linking to the CeDeFi section of the Lending Architectures concept page.
Morpho tagged as Modular Lending architecture
Added a visible architecture tag to Morpho's header linking to the Modular section of the Lending Architectures concept page.
Aave tagged as Monolithic Lending architecture
Added a visible architecture tag to Aave's research-meta header linking to the Monolithic section of the Lending Architectures concept page. Makes the classification explicit rather than implicit.
DeFi Risk Methodology: six failure modes now machine-readable
The six failure modes (oracle-to-execution divergence, recovery endogeneity, bank runs, parameter rigidity, oracle latency, congestion-dependent liquidation) now have stable anchor IDs and per-protocol exposure tags in data/defi-risk-scores.json. 7 of 14 protocols are tagged (aave [2,3,4], morpho [1,2,5], maker [5,6], maple [2,3], ethena [3,5], ondo [5], hyperliquid [2,6]). The new js/defi-risk-breakdown.js renders the tags as pills on research pages.
Site-wide jargon sweep: replaced 'V1–V5' labels with plain-English names
User-facing 'V1–V5' terminology across the site was replaced with 'five mechanical risk channels' / 'Vault Credit Risk framework.' Metric-card tags on the concept page went from 'V1, V2...' to 'Channel 1, Channel 2...'. Oracle sub-channels 'V4a / V4b' renamed to 'Oracle latency' and 'Oracle manipulation.' Solver Landscape card rewritten to lead with reader benefit. Anastasiia's original V1–V5 nomenclature is still credited in the attribution.
Taoflow (November 2025) documented on Bittensor page
Page previously described dTAO without the Taoflow upgrade. Emissions are now driven by net TAO flows (staking minus unstaking) with a ~30-day EMA half-life, not alpha price. Subnets with positive net flows share emissions proportionally; negative-flow subnets receive zero. This is the mechanism that makes the 'dTAO flywheel' real rather than narrative.
Rayon Labs concentration: ~24% of Bittensor emissions sit with one dev group
Rayon Labs operates Chutes (SN64, 14.4% of emissions), Nineteen (SN19), and Gradients (SN56), which together capture ~23.7% of daily TAO emissions. A single development group controlling nearly a quarter of the incentive distribution is a structural decentralization problem that gets less airtime than the Covenant drama but is arguably more load-bearing — a Covenant-style exit from Rayon would be materially larger.
BIT-0011 governance proposal: locked-stake conviction for subnet ownership
Const (Jacob Steeves) announced BIT-0011 on April 16 as a direct response to the Covenant AI exit. Participants lock ALPHA tokens for a chosen duration; locked amount × remaining duration produces a decaying conviction score. Highest-conviction staker controls the subnet and captures the 18% owner emission share. Covenant-style exits become economically expensive. But: the triumvirate multisig at the protocol layer is untouched, capital intensity rises, and whale-attack surface on small subnets becomes a live concern. A targeted fix for one failure mode, not a governance overhaul.
Prime Intellect + Nous Research identified as direct decentralized-training competitors to Bittensor
Bittensor's closest competitors on the decentralized-training claim aren't Render or Akash — they're Prime Intellect (INTELLECT-2: 32B parameter model, globally distributed async RL, Apache 2.0, arxiv 2505.07291) and Nous Research (Psyche Network, DisTrO optimizer, $50M Paradigm Series A at $1B valuation). Covenant-72B has more parameters than INTELLECT-2 but INTELLECT-2's async RL is arguably the harder coordination problem. Bittensor's moat in this niche is narrower than the 'first' framing suggests.
Targon (SN4) is Bittensor's cleanest revenue story — Dippy AI migration is real
Manifold Labs' Targon subnet closed a $10.5M Series A in August 2025. Dippy AI, a viral character app with 8.6M users, migrated its entire backend to Targon in a six-figure deal — the first mainstream consumer app to move onto Bittensor-native infrastructure. Confidential compute via Intel TDX / AMD SEV / NVIDIA PPCIe TEEs. Single-anchor concentration is the main caveat.
Prompt-injection added as emerging attack class in DeFi Risk Methodology
New bullet in 'What this framework does not capture': agentic-execution risk — prompt injection via poisoned oracles, ENS records, or contract metadata can hijack agent behaviour and drain wallets with no phishing or malware step. Distinct from the existing Oracle dimension because the attacker surface is the agent's reasoning layer rather than the protocol's pricing layer.
New Agentic Finance Stack section on Ethereum research page
Added coverage of the four Ethereum-native standards that together form the first full stack for machine-to-machine payments, identity, commerce, and execution: x402 (Coinbase, ~119M Base + 35M Solana txns, ~$600M annualized), ERC-8004 (trust, live Jan 29 2026), ERC-8183 (commerce, proposal), ERC-8211 (smart batching, April 6 2026). Competitive landscape (Tempo mainnet, Visa CLI, Stripe ACP) included.
RWA Composability headline numbers refreshed to March 2026
The page's headline figures were updated from '$27B tokenized / $2.7B in DeFi' to '$28.6B tokenized / $3.6B in DeFi' (DefiLlama, March 2026). Year-over-year growth context added: active RWA market cap grew from ~$4.1B in early 2025 to $25.2B in March 2026. Credit concentration restated with cleaner numbers: ~$4.6B of AUM / ~$2.8B of DeFi TVL (~77% of all RWA-in-DeFi usage).
Live RWA private-credit markets on Morpho: Pareto FalconX at 92% utilization
Two live primary examples of the curator-layer-financing-RWA thesis appeared on Morpho: Pareto AA_FalconXUSDC/USDC at ~$48M market size and 91.96% utilization (at 77% LLTV), and Fasanara mF-ONE/USDC at ~$20M and 91.5% LLTV. Gauntlet's Aera vault automates the levered RWA carry into a managed structure with pre-defined risk controls — effectively an onchain repo for private credit.
Maple's syrupUSDC is now the largest tokenized private-credit product onchain
syrupUSDC sits at ~$2.10B active market cap and is integrated across six major DeFi venues: Morpho, Pendle, Aave, Drift, Kamino, and Jupiter. This takes Maple from 'CeDeFi lending venue' to 'composable credit primitive across the stack.' Overall private-credit-in-DeFi TVL reached ~$2.8B in April 2026 and Maple's products drive the majority.
Commodities account for >50% of Hyperliquid RWA perp volume; silver leads gold
Across Hyperliquid and Ostium, commodities make up more than half of all RWA perp volume. On Hyperliquid specifically, silver perps run 2–3× gold volume and account for ~40% of RWA perp activity on peak days, with open interest around $157M and daily volume around $351M. Daily precious-metal perp volume on Hyperliquid has topped $1.3B. Ripple Prime's March 2026 HIP-3 integration brought gold/silver/oil onchain for institutional prime-brokerage clients.
Aave XAUT V3 vault supplied balances grew ~10x over five months
From September 2025 to February 2026, XAUT balances supplied into Aave V3 expanded close to tenfold. Tokenized gold now represents ~95% of the $5.87B onchain commodity market cap. The smart-contract interaction share for XAUT crossed 75% of transaction count and 36% of volume by 2025 — gold holders are deploying into credit markets instead of passively storing. Clean validation of the monolithic-pool thesis.
Solana scored in L1 Scorecard at 64.3 / 51.3
SOL wins Technology & Performance outright (85 vs ETH's 70) and wins stake-weighted Nakamoto coefficient within Security & Decentralization (60 vs ETH's 40). Material gaps on Economic Sustainability (46, fees cover about 9% of issuance) and validator concentration.
Ethereum leads the inaugural L1 Scorecard at 81.6 / 73.1
ETH wins six of eight dimensions in the launch scorecard, with headline strengths on Ecosystem Depth (95) and Investment Accessibility (96). Weakest dimension is Technology & Performance (70) — L1 throughput is still low even when counting L2s. Risk-Adjusted gap to the next-best is ~22 points.
Launched the L1 Investment Scorecard: standardized 8-dimension fundamentals framework
A new TokenIntel-original scorecard for Layer 1 assets as investments. Eight equally-weighted dimensions (Network Fundamentals, Economic Sustainability, Monetary Policy, Technology, Ecosystem Depth, Security, Governance, Investment Accessibility) across ~30 sub-criteria scored against absolute benchmarks. Two parallel composite scores per asset: Fundamentals and Risk-Adjusted. Ships with ETH and SOL fully scored.
L1 Investment Scorecard extended to 4 assets: ETH, SOL, BNB, HYPE
BNB and Hyperliquid scored against the same 8-dimension, ~30-sub-criterion framework. BNB lands at Fundamentals 63.3 / Risk-Adjusted 48.8 — strong deflationary economics and throughput, heavy centralization cost. HYPE at Fundamentals 55.6 / Risk-Adjusted 38.6 — specialist chain with best-in-cohort Monetary Policy and Technology, thin Ecosystem Depth by design. BTC and XRP left as placeholders with explicit notes on why the framework doesn't fit them.
Scatter chart overlap fixes and HYPE symbol clipping resolved
Label offsets on the L1 Scorecard scatter chart were reworked so SOL, BNB, and HYPE don't collide at their near-identical positions. HYPE's 4-character symbol now uses a smaller font so it fits inside the dot. Leader lines added to disambiguate clustered dots.
HYPE scorecard re-scored: Network Fundamentals 47→70, Ecosystem Depth 31→44
Editorial re-score of Hyperliquid on two L1 Scorecard dimensions: active addresses expanded to include Hypercore perp traders (not just HyperEVM users), tx count rebased to include Hypercore onchain order operations, developer activity reflects the live builder-frontend ecosystem. Ecosystem Depth bumped for USDC perp-margin stablecoin flow and DAT/ETP institutional-access improvements. New composite: Fundamentals 60.0 / Risk-Adjusted 43.0.