22 research updates published. Positive: 14 · Negative: 1 · Neutral: 7.
Previously the worker fetched the previous snapshot from the static /data/leverage-audit-latest.json, which only updated when a human ran the CLI locally and committed. This commit adds Supabase persistence so the diff auto-updates each run.
One-shot fix sweep driven by the content-leverage-audit output. Takes the audit from 81 total violations to 0 and from 5 cross-protocol awareness gaps to 0 in a single commit.
Two bugs in renderScatter() the expert review caught:
Ships a weekly (Wed 14:00 UTC) audit that verifies TokenIntel content is cross-leveraged and SEO-consistent. Two layers:
Closes the two obvious gaps from the Phase 1/2 launch: the module was only on 5 research pages, and events still had to be added to the JSON by hand. Both solved.
Closes the two obvious gaps from the Phase 1/2 launch: the module was only on 5 research pages, and events still had to be added to the JSON by hand. Both solved.
Two small pieces that close Item #3 of the research-changelog plan and put a hard reminder on the calendar for the Phase 3 decision.
Two small pieces that close Item #3 of the research-changelog plan and put a hard reminder on the calendar for the Phase 3 decision.
Expert review feedback called out three high-leverage gaps on the L1 Investment Scorecard page: no "so what" layer per asset, bar chart doesn't show who wins/loses each dimension, weakness info buried in prose. This commit addresses those six items, bounded in scope:
Expert review feedback called out three high-leverage gaps on the L1 Investment Scorecard page: no "so what" layer per asset, bar chart doesn't show who wins/loses each dimension, weakness info buried in prose. This commit addresses those six items, bounded in scope:
Phase 1 + 2 of the trust-by-transparency surface the owner asked for. Every research update TokenIntel ships now gets logged, classified by underlying-event direction, and aggregated into per-project momentum. Two-week history back-tagged so the scorecard has real data on day one.
Table was rendering with jittery values (+0 / -0 / mixed positional treatment of positive vs neutral columns) and center-floating numbers because headers lacked explicit right alignment.
The yield-spread thesis is visible in real time: USD Curator ARR converging with Treasuries at 3.79% vs 3.80% is exactly the configuration in which the composability-vs-holding trade-off stops favoring structured loops. The callout explains what happens when spreads widen again.
Avenir (Li Lin's family office), Asia's largest BTC ETF holder at >$900M in BlackRock IBIT as of December 2025, was acquired by Xinhuo Group. Xinhuo plans to launch Hong Kong's first compliant Bitcoin-backed wealth product, absorbing Avenir's investment team and long-running Bitcoin strategy capability.
Fairfax County Employees' Retirement System — a roughly $6B Virginia public pension fund — increased its digital-asset exposure to 4.5% of the total fund. The allocation blends blockchain-infrastructure equity, direct crypto exposure, and short-duration yield strategies.
After negative January and February reads, spot BTC ETF flows swung to +17.6K BTC by month-end and corporate treasury (DAT) flows to +30.9K BTC, with mid-month peaks of +30.6K and +46.8K. BTC capital flow overall still closed March at −$7.0B, but that's a meaningful improvement from −$9.6B mid-February.
Aggregate on-chain curator TVL has recovered toward ~$5B from the ~$10B late-2025 peak drawdown following February's leverage unwind. The USD Curator Benchmark's 3-month annualised return (3.79%) has converged with US Treasury yields (3.80%) — the first period of near-parity since the yield compression began. Morpho is the single largest venue inside this category.
Glassnode's March 2026 Strategy Watch survey shows Hyperliquid's share of fund-manager ecosystem focus rose sharply, now approaching Solana for third place behind Ethereum and Bitcoin. Ethereum's share fell from roughly 50% to 40% over the same period. Glassnode framed the shift as 'increasingly structural rather than speculative.'
At current compressed spreads, even the best-designed fee-switch activation produces only ~$26M annualised buyback against ENA's ~$74M daily trading volume and $300M+ of 2026 emissions. OAK Research's own recommendation: wait. The real activation window probably sits in late 2026 or beyond — meaningfully later than the 'imminent catalyst' narrative typically prices in.
Bybit incident ($30M resolved through Copper Clearloop without a depeg), Oct 10/10 orderly contraction (supply shrunk mechanically as yields compressed, no panic redemption), and the April 2026 Kelp/LayerZero response (bridges paused within hours, proof of reserves published on-demand, zero rsETH exposure confirmed at 101%+ overcollateralised). Pattern: credit-trained management doing what credit-trained management does in stress.
~$2.9B of USDe/sUSDe — roughly half of total supply — is deposited in Aave as leveraged-loop collateral, with peak Ethena-related exposure on Aave reaching $6.6B. The composability flywheel that accelerates USDe demand also means a shock that breaks the loop economics (sUSDe APY dropping below USDC borrow cost, or an adjacent-asset failure) triggers mechanical unwinds.
Ethena announced a structural shift away from pure basis-trade backing: institutional USDC lending via Anchorage Digital (OCC-chartered), Maple Institutional, and Coinbase Asset Management, plus expanded RWA exposure, HyENA's equity/commodity perp basis on Hyperliquid, and prime lending to trading firms. Counter-cyclical by design — the new credit book should yield when perp funding compresses. If executed, sUSDe APY could step up from ~3.75% toward 8-10%.