Reports

Hyperliquid is building from the consensus layer up. Jupiter is building from the user layer down. Same playbook, different bets, very different multiples. Microsoft owned the OS. Google owned distribution. Visa owned settlement. The crypto-trading version pays a 2-3.5x multiple premium to the infrastructure side.

Lido yields 3%. Ondo yields 4-5%. Ethena yields double digits. Comparing those numbers is comparing four different risks. The APY is not a return; it is the bill the market sends you for warehousing what is underneath. Includes the credit-spread ladder, the composability premium, and the stETH June 2022 case study.

UNI's fee switch activated on December 28, 2025 via UNIfication. Cumulative holders revenue since: $14.9M. The "AERO clean vs UNI broken" framing died six months ago. Both are now value-accrual protocols; the mechanism choice (direct ve(3,3) pass-through vs universal-eligibility buyback-and-burn) has real per-token consequences.

Most investors model protocol revenue. Very few model who will own that revenue three years from now. SUSHI's annual revenue grew 21x from 2020 to 2021, and the token still collapsed 95%. TI extends its token-vesting-supply framework with Ownership Decay, the velocity dimension that the static MC/FDV ratio misses.

Six DAOs. Six totally different things you're actually buying. AAVE, SKY, UNI, AERO, MORPHO, and LDO confer materially different rights, route revenue to materially different recipients, and behave materially differently in a crisis. TI's six-type DAO taxonomy applied to the cleanest cross-asset stress test in DeFi's history. "DAO governance has become the crypto equivalent of calling every government a democracy."

Four DeFi tokens, four active revenue-funded buybacks, four wildly different outcomes. JUP 18.83% yield / −80% return. HYPE 6.78% yield / +533% return. TI's NBY framework extended with four inputs explains the dispersion. The buyback is the mechanism. It is not the moat.

Three 2026 lending cascades, three architectures, three different places loss showed up while every liquidation engine ran exactly as designed. Aave $230M Scenario-2 bad debt on rsETH; Kamino $0 bad debt on $41.7M liquidated; Morpho $3.8M from a $200K exploit. DeFi solved liquidation tech. Not concentration.

The lazy bear case on Pendle says vesting unlocks will overwhelm the sPENDLE buyback. It's wrong. Insiders are fully vested, NBY is positive at every scenario (+1.9% to +5.5%), and the token still isn't a buy. Revenue-to-NBY sensitivity heatmap, three explicit falsifiers, and the asymmetric trade is PT-USDG, not spot PENDLE.

Aave turned its fee switch on. Morpho kept its off. Euler proposed eliminating protocol fees entirely. For a six-week window, the protocols capturing the least direct value grew the fastest. Applies TI's Five Questions framework + NBY + fee-durability tiers to three opposite lending positions. Growth-vs-yield quadrant chart, architecture diagram, falsification criteria.

P/F multiples treat all revenue the same. Token holders rarely receive what shareholders do. Five protocols (Aave, Morpho, Aerodrome, Sky, Kamino), five value-routing archetypes, three layers of enforceability, the Net Buyback Yield dilution-adjusted lens, fee-quality durability tiers, and the synthesis: five concrete questions any investor should answer before sizing a position.

Token at $659, +13.5pp ahead of ETH over 30 days, with pending US spot ETF catalyst. Cleanest supply mechanics of any large-cap L1 (100% unlocked, ~6M BNB/yr burn). But fee engine is contracting -25.9% vs SOL +9.2%; multiple is 363x P/F vs SOL 24x. Four scenarios, probability-weighted target ~$797 (+21%). Hold existing, don't fresh-allocate.

Token at $1.26 vs $7.86 December 2025 TGE high (-84%), +61% off March 2026 trough. Calendar-defined trade: positive NBY for 7 months, then hard inversion December 30, 2026 cliff (~456K LIT/day at 4.45x investor gain). Four scenarios, probability-weighted target ~$1.72, hard exit by Dec 1, 2026.

Token at $1.91 vs $7.50 April 2024 high, +89% off April 2026 trough. Clean cap table (100% unlocked since Sept 2024), mechanically-live sPENDLE buybacks (80% of V2 fees), but weak fee engine. Five-force teardown, four valuation frameworks, three scenarios, probability-weighted target ~$2.64, hard invalidation below $1.01 ATL.

Headline buyback yield is misleading whenever a token is still in active vesting. Net Buyback Yield = (Buybacks - Sell-Through × Unlocks) / Mcap. Applied to HYPE and PUMP: same 93%-of-fees buyback intensity, opposite outcomes. HYPE faces $4.18B of unlocks vs $337M of buybacks through year-end (NBY -4.9% to -26.7% of mcap). PUMP faces $219M of unlocks vs $210M of buybacks (NBY +15.7% to +24.1%). The unlock side of the equation is the bigger lever.

McKinsey/JPM-style memo on Ethena's token, currently at $0.107 versus a $1.91 March 2024 high. Five-force teardown: revenue cliff ($10.2M to $655K quarterly), fee switch gated by macro, ~172M ENA/month dilution, defensive portfolio rotation. Four scenarios, probability-weighted target ~$0.142, hard invalidation below $0.077 ATL.

Two stablecoin designs ran a market-scale experiment in front of the entire industry. Sky's USDS doubled down on overcollateralized + RWA-backed; Ethena's USDe innovated with synthetic delta-neutral. By dollar supply on 2026-05-22: Sky ecosystem $12.87B, Ethena $5.57B. The 2.3x gap did not exist 12 months ago. Why boring won and what it means for the next synthetic-dollar attempt.

Stream Finance's November 2025 collapse took a contained $93M loss inside one fund and turned it into a multi-protocol cascade across at least 15 Morpho vaults plus Euler positions. Three structural choices amplified it: hardcoded oracle (no pause threshold), structural leverage at the synthetic-stablecoin level (3.25x: $520M xUSD float on $160M deposits), and 100% utilization ceilings.

The BTC/Gold ratio has recovered to 16.85 from a February trough of 12.31 (+37%), still −42% below the October 2025 cycle peak near 28.79. The recovery has been driven almost entirely by gold giving back parabolic gains rather than by Bitcoin ripping. Warsh confirmed as Fed Chair, Operation Epic Fury pushing April CPI to 3.8%, rate-cut path dead.

Aave's WETH reserve was 98.5% LST-funded going into the April rsETH event. The lesson is broader than one protocol: pooled-liquidity lending mispriced concentrated tail risk because the rate framework cross-subsidizes by collateral class. Three protocols, three responses, one structural answer in the V4 risk-premium framework.

Tokenized US equities now have two live routes. DTCC tokenized entitlements treat the chain as a settlement-rail upgrade with no disintermediation. Securitize-Computershare and Bullish-Equiniti tokenize at the registry, where the token is the share. The two largest US transfer agents have publicly aligned with the registry vision.

BlackRock built BUIDL for institutions. Ethena, Ondo, Frax, and Spark adopted it as the base layer, and ecosystem stables (MegaETH USDm) now build on top of those. A three-layer tokenized-Treasury supply chain: BlackRock at the bottom, DeFi protocols in the middle, ecosystem stables at the top. Spark's Tokenization Grand Prix mandated $500M of $1B to BUIDL.

Three of the largest crypto loss events in the last 18 months (ByBit / Gnosis Safe ~$1.5B, Resolv / Morpho ~$3.8M cascade, Kelp / LayerZero ~$290M) were not contract bugs. They were attacks on the surrounding infrastructure. AI capability (Anthropic Mythos solving a 32-step network attack end-to-end) compresses the cost of attacking the soft surface further.

The three biggest "adoption" stories of 2026 share a structure most coverage misses. Strategy holds 66% of public-company BTC. weETH backs ~48% of Aave's e-mode collateral. TRON carries 44% of USDT supply. Same shape, three different parts of the stack. Disaggregating each headline aggregate and committing TI to writing the distribution-shape number every time the adoption number gets quoted.

Galaxy Research's April 22 V3 snapshot reframes the rsETH event as the visible piece of a larger structural exposure still on the book. E-mode is 58.84% of Aave V3's debt, runs at ~10x leverage, and weETH backs ~48% of e-mode collateral. A 10% weETH depeg would move $2.47B of debt against $2.42B of post-shock collateral.

Three proposed factor additions to TI's SLS computation: chain concentration (TRON at 44% of USDT supply), reserve composition risk (S&P-flagged 24% non-cash share), and external rating drift. Calibration material for the regime classifier. Live SLS reads 50 today; with these three factors integrated, same-date composite likely reads 35 to 42.

Tether holds $122B in direct US Treasuries and $141B total Treasury exposure as of year-end 2025. That direct holding alone exceeds the sovereign Treasury position of Germany, the UAE, Spain, or Australia. The right frame for Tether in 2026 is macro-infrastructure: a privately-held money-market-fund-shaped entity that has become a meaningful marginal buyer of short-duration US sovereign debt.

BTC at $78,543, TI signal buy in an accumulation regime. The TI Mean Reversion Index reads Q37 neutral across four cost-basis anchors. The structural absorption of supply by spot ETFs and corporate treasuries continues even as price chops sideways. That is the report's central question.

Two atomic numbers per asset (Trust Assumptions, Override Points), 21 assets scored in a day, zero LLM-inferred counts. Methodology, four-quadrant view, and the five sharpest findings. The framework that drives TI's centralization-map page.