Issue 2 — Last week's contraction was repositioning, not retreat: net flow recovered to -0.2%, and Felix's HyperEVM caps filled in seven days

Five bullets, the only thing some readers will read. This is the catch-up weekly for the week ending Saturday 2026-05-30 — we publish issue 2 with a single organising frame: every one of the five watchlist items from issue 1 gets an explicit answer below. Numbers are sampled at the close of the window.

The Morpho Blue universe now spans 329 active vaults holding $4.56B in TVL across 11 chains — ten more vaults than last week, TVL essentially flat ($-9.8M). Aggregate capacity utilization ticked up to 77.7% ($4.56B deployed against $5.87B of stated caps), from 77.0% last week. The universe absorbed new headroom and filled some of it — a healthier signal than last week's "leaning into what's already full."

The cross-chain drift paused — and briefly reversed. Ethereum holds 60.1% of TVL ($2741M) and was net positive this week at +$11.1M (+0.4%) — the first positive Ethereum print we've recorded. Base sits at 32.5% ($1480M) and actually lost $-19.1M (-1.3%) on net. One week does not break a quarter-long Ethereum→Base trend — but the direction flipped, and it's worth watching whether this is a pause or an inflection.

The regulated-stable thesis from issue 1 strengthened. USDC anchors at 50.7% ($2312M). PYUSD climbed to 14.7% ($670M, up from 13.0% last week) and RLUSD to 9.5% ($435M). The PYUSD + RLUSD stack is now 24.2% of the universe — almost four times USDT's 6.6%. The bank-backed payment stables aren't just holding their lead on Morpho; they're extending it.

Curator concentration plateaued. Curator HHI 0.256 — down a hair from last week's 0.259, still just above the 0.25 FTC threshold. Steakhouse Financial (40.1%), Sentora (24.3%), and Gauntlet (18.5%) hold 83% of universe TVL, fractionally below last week's 83%. The composition shifted underneath the headline, though: Sentora overtook the gap to Gauntlet decisively (now a clear #2), and Gauntlet's decline pushed share down to the mid-tier rather than up to the leader.

Net universe TVL was $-9.8M (-0.2%) this week — a near-complete recovery from last week's -3.8% contraction. This is the single most important number in the issue, because it validates the central provisional read from issue 1: the W21 outflow was curated repositioning, not capital fleeing the universe. When a -3.8% week is followed by a flat week with no echo, the contraction was a level-shift from a one-off wind-down, not the start of a trend.

The week's biggest outflow is an order of magnitude smaller than last week's. Gauntlet USDC Prime on Base shed $-42.8M (-11.7%), from $365M to $322M. A -11.7% move on a $322M vault is ordinary book management — not the -73%-class billion-dollar wind-down that defined last week. The Gauntlet repositioning continues, but it has normalised.

Felix HYPE is the headline of the week — and it answers issue 1's open question. Felix HYPE on HyperEVM took +$30.0M (+65.5%), the largest single inflow anywhere in the universe. Last week we flagged Felix's cluster of HyperEVM cap-changes and asked whether the new headroom would fill (supply-constrained) or sit empty (demand-constrained). It filled in days. The whole HyperEVM chain gained +$26.9M (+34.4%) on the back of it. That's a clean supply-constrained read — and it reframes HyperEVM from "speculative new chain" to "borrow demand outrunning supplied capital."

Where the rest of the inflows landed: Sentora PRIME Main (+$26.8M), Steakhouse Prime USDC (+$26.5M), and Steakhouse Ethena USDtb (+$15.1M). The pattern is healthy breadth — inflows spread across Felix (HyperEVM), Sentora (Ethereum), and Steakhouse (Base + Ethereum), not a single concentrated allocation.

Curator-level rollups make the reshuffle plain. Sentora was the week's biggest gainer at +$28.0M (+2.6%), cementing its #2 position. Felix gained +$25.6M (+36.8%) — proportionally the strongest move among the larger curators. Steakhouse Financial held flat (+$7.9M). And Gauntlet was again the lone large loser at $-69.3M (-7.6%) — the second consecutive week of net outflow, but at roughly a third of last week's magnitude.

Universe weighted APY: 4.00%, sustainable 3.94%, sustainable ratio 98.5% — a marginal -0.10 pts over the 7-day window, sustainable ratio flat at 98.5% (98.5% a week ago). The headline is the same as issue 1, and that consistency is itself the signal: this universe earns its yield from borrow demand, not emissions, and that hasn't wavered through a contraction-then-recovery fortnight.

Decomposition by asset. USDC (51% of the universe) prints a weighted 4.39% with sustainable 4.38%, P25–P75 running 3.6%–5.5% — a touch wider than last week, as the AlphaPing book (see spotlight) pulls the top of the USDC cohort up. USDT prints weighted 3.50%, a tighter cohort (P25–P75 3.0%–3.4%). WETH sits at weighted 2.45% — below the LST baseline again, the persistent pattern: WETH supply on Morpho isn't being paid a meaningful spread over directly holding stETH.

Top APY movers this week (TVL ≥ $1M):

Alpha Frax USD Enhanced V2 (AlphaPing, Ethereum): 9.77% → 30.99% (+21.22 pts). Sharp upside — likely a fresh incentive or a risk-on repositioning; will not compound.

SingularV ETH (SingularV, Ethereum): 18.75% → 37.03% (+18.27 pts). Sharp upside — likely a fresh incentive or a risk-on repositioning; will not compound.

MEV Capital wETH (MEV Capital, Ethereum): 19.50% → 1.49% (-18.01 pts). Sharp downside — emission program likely rolled off; do not treat the prior yield as a baseline.

Apyx USDC (Hyperithm, Ethereum): 0.00% → 10.29% (+10.29 pts). Sharp upside — likely a fresh incentive or a risk-on repositioning; will not compound.

Felix USDT0 (Felix, HyperEVM): 5.30% → 14.89% (+9.60 pts). Sharp upside — likely a fresh incentive or a risk-on repositioning; will not compound.

August USDC V2 (August Digital, Monad): 7.15% → 12.60% (+5.45 pts). Sharp upside — likely a fresh incentive or a risk-on repositioning; will not compound.

The cliff issue 1 warned about happened. MEV Capital wETH — the vault we used last week to illustrate the spot-vs-sustainable trap — fell -18.01 pts this week, from 19.5% to 1.5%. An allocator reading the trailing-30d window still sees a strong number today; the spot has collapsed. This is the exact failure mode we flagged, materialising one week later. The separation between spot and sustainable is not academic — it is the difference between a 19% position and a 1.5% one, with no warning in the trailing average.

Vault sustainable APY z-scored against its (loan asset) cohort flags positions paying materially above peers. A vault paying >2σ above cohort is being paid for risk we don't yet see — surface it, decide whether the risk is legible, size accordingly.

This week's standouts:

SingularV ETH (SingularV, WETH): sustainable 37.03% vs WETH cohort median 1.49% — z=6.5, anomaly score 80.

Alpha Frax USD Enhanced V2 (AlphaPing, frxUSD): sustainable 30.99% vs frxUSD cohort median 7.74% — z=1.4, anomaly score 75.

August USDC V2 (August Digital, USDC): sustainable 12.60% vs USDC cohort median 4.20% — z=3.2, anomaly score 64.

Hyperithm cbBTC Apex (Hyperithm, cbBTC): sustainable 3.75% vs cbBTC cohort median 0.02% — z=2.4, anomaly score 55.

Alpha USDC Prime V2 (AlphaPing, USDC): sustainable 11.11% vs USDC cohort median 4.20% — z=2.6, anomaly score 51.

SingularV ETH is the universe's highest anomaly for the second straight week (z=6.5, sustainable 37.03% against an ETH cohort median of 1.49%) — and this week we can partly explain it. SingularV ETH is also the universe's top risk-mover: composite risk 56, dominant factor liquidation=98. The 37.03% isn't free yield — it's a position running close to its liquidation boundary. The premium over peers is the market pricing tail risk that the score now corroborates. Last week we said "investigate before sizing"; the risk readback this week is the reason why.

AlphaPing is the most anomaly-dense curator in the universe — 4 of their vaults sit in this week's top-5 flags (Alpha Frax USD Enhanced V2 z=1.4, Alpha USDC Prime V2 z=2.6, Alpha USDC Asia V2 z=2.4, Alpha USDC Delta V2 z=2.4). That density is exactly why we promised an AlphaPing deep-dive last week, and it's this issue's curator spotlight (§07). Read the two sections together: the anomaly flags are the symptom; the spotlight is the mechanism.

Decision rule, restated: any vault flagged here can be sized, but only once the premium's attribution is mechanically understood. "Paying high" without a mechanical reason is paying for something the market sees that you don't. SingularV's near-liquidation posture is now legible; AlphaPing's frxUSD/USDC cluster is the open question the spotlight tackles.

Third consecutive clean week on credit: 0 RED warnings, 2 YELLOW. Bad debt realized: $0.0M. Bad debt unrealized: $0.0M. Zero credit losses across $4.56B of supplied capital, again. The structural credit performance of the Morpho universe remains the under-discussed positive of the whole story — two contraction-and-recovery weeks with not a dollar of realised bad debt.

The 2 yellow warnings are the same benign kind as last week: unrecognized_collateral_asset — vaults using a collateral asset Morpho's UI hasn't yet labelled canonically (Moonwell Ecosystem USDC Vault; Moonwell Ecosystem USDC). Not a credit signal; a reminder the collateral set is expanding faster than the registries.

No depeg elevations — the third consecutive clean week. All 24 tracked stables sit in the healthy band; no oracle/spot divergence beyond the watch threshold; no issuer-side stress active. With PYUSD and RLUSD now 24.2% of the universe, a clean depeg surface on the regulated-stable stack is the precondition for that share to keep growing.

Pending governance changes: 20 queued. 14 of them are again Felix cap changes across the HyperEVM book (RLP, WHYPE, kHYPE, UBTC, PT-kHYPE-19MAR2026, wstHYPE collaterals). Last week's Felix cluster filled — and they have queued another 14 this week. Felix is running a continuous cap-expansion program on HyperEVM, and given last week's fill-through, each new cap is a forward indicator of where the next inflow lands. The remainder is broad: Gauntlet (3), plus single changes from Clearstar and Anthias Labs.

Risk distribution shifted up this week — read it as composition, not deterioration. The universe risk-score median moved from 19 (issue 1) to 28, with the whole distribution lifting (P25 26, P75 34). With zero realised bad debt and a clean warnings stack, this is almost certainly a composition effect — the ten net-new vaults this week skew toward newer, higher-complexity HyperEVM/Monad books rather than any deterioration in existing positions. We flag it explicitly rather than smooth it over: if next week's median holds or climbs further without a credit event, we'll dig into whether it's the score's sensitivity to new-collateral complexity or a genuine universe-wide risk creep.

We promised this one. Last week's watchlist flagged AlphaPing as the universe's most anomaly-dense curator and committed to a deep-dive. Here it is: AlphaPing runs $76M across 14 vaults — 1.7% of the universe — at a weighted APY of 9.98%. That APY is the story. The universe-weighted figure is 4.00%; AlphaPing pays +6.0 points above the universe average. They are, by a wide margin, the highest-yielding curator of any meaningful size on Morpho.

Where the yield comes from. AlphaPing's book is overwhelmingly USD-denominated (11 vaults, $72M) and that sleeve prints a weighted 10.28% — against a USDC cohort median near 4.4%. The flagship anomalies: Alpha Frax USD Enhanced V2 (frxUSD, sustainable 30.99%, z=1.4); Alpha USDC Prime V2 (USDC, sustainable 11.11%, z=2.6); Alpha USDC Asia V2 (USDC, sustainable 10.71%, z=2.4). Their Alpha Frax USD Enhanced V2 was this week's single biggest APY mover in the entire universe — 9.77% → 30.99% (+21.22 pts). This is a book built on frxUSD and frax-adjacent collateral plus enhanced-carry USDC strategies — a deliberate carry posture, not diversified stable lending.

The crucial qualifier: their reported risk and warning surface is clean. Weighted risk 32 (universe median 28), 0 red / 0 yellow warnings, 0 pending changes. So the deterministic score does not flag this book as dangerous — which is precisely the situation the yield-anomaly screen exists for. A 9.98% weighted APY with a sub-median risk score is, on its face, a free lunch — and free lunches are where the legibility discipline matters most. The score reads the on-chain parameters (LLTV, oracle, collateral); it cannot read the frxUSD redemption mechanics or the enhanced-carry strategy's unwind path. That gap is the underwriting work.

The structural read. AlphaPing is the universe's carry specialist — a fundamentally different animal from the §02 leaders. Steakhouse, Sentora, and Gauntlet run large, low-dispersion books at-or-below universe-weighted yield; AlphaPing runs a small, high-z book at more than double it. For an allocator, the two are not substitutes. AlphaPing is a satellite carry sleeve to be sized as risk budget — emphatically not a core stable allocation, however clean the headline score looks.

Felix HYPE on HyperEVM, now $76M TVL — up +$30.0M (+65.5%) this week, the largest single inflow anywhere in the universe. Last week we spotlighted the universe's biggest outflow and asked "why?"; this week the more interesting question is the biggest inflow, because it answers a question we posed explicitly seven days ago.

The setup, from issue 1. We flagged Felix's cluster of pending HyperEVM cap-changes and asked the diagnostic question: when the new headroom opens, does it fill instantly (the HyperEVM borrow side is supply-constrained) or sit empty (demand-constrained)? The answer determines whether HyperEVM is a real lending market or a chain looking for borrowers.

The answer: it filled in days. Felix HYPE absorbed +$30.0M into the freshly-raised caps almost immediately, taking the vault to $76M. The whole HyperEVM chain gained +$26.9M (+34.4%) — almost entirely on Felix's book. There was supplied capital waiting for the headroom. That is a textbook supply-constrained signal: borrow demand on HyperEVM was real and unmet, and the binding constraint was Felix's cap, not borrower appetite.

What it means structurally. HyperEVM graduates, on this evidence, from "speculative new chain" to "a market where lending demand outruns the supplied book." That's the strongest possible read for a young chain — it means TVL growth there is demand-led, not incentive-led (consistent with the universe's 99% sustainable ratio). And it makes Felix's remaining 14 queued cap-changes (§06) a genuine forward indicator: if this week's cap filled in days, next week's are the best available signal for where the next HyperEVM inflow lands.

The fragility note. A chain whose flagship lending book is one curator (Felix) on a handful of HyperEVM-native collaterals (WHYPE, kHYPE, UBTC, wstHYPE) is concentrated by construction. The supply-constrained read is bullish for growth and bearish for diversification simultaneously: HyperEVM's lending universe is Felix's book right now. An allocator chasing the HyperEVM yield is, in practice, taking a concentrated Felix-curator and HYPE-collateral bet — size it as such.

Five items the editorial team will track through the week ending Saturday 2026-06-06. Issue 1's five all resolved this week — that's the bar we hold the format to.

(1) Gauntlet — repositioning complete, or structural decline? Two consecutive net-outflow weeks (-3.8% contribution then, $-69.3M now), share down from 20.3% to 18.5%. Week three is decisive: stabilisation says the wind-down is done; a third outflow week says it's a trend in their universe share.

(2) AlphaPing's carry — does Alpha Frax USD Enhanced V2 hold or cliff? Their flagship just spiked +21.22 pts to 30.99%. The MEV Capital wETH precedent (§04) is exactly how these unwind. If it holds, the carry is structural; if it cliffs, it was a transient and the §07 underwriting caution was warranted.

(3) SingularV ETH liquidation watch. Now both the top yield anomaly (z=6.5) and the top risk-mover (risk 56, liquidation factor 98). A vault paying 37% while sitting near its liquidation boundary is a binary — it either continues to print or it liquidates. We'll report which.

(4) HyperEVM fill-through. Felix has 14 more cap-changes queued after this week's filled in days. If the next batch fills as fast, HyperEVM's supply-constrained status is confirmed structural, not a one-off — and the chain's TVL trajectory becomes a function of how fast Felix raises caps.

(5) The cross-chain drift — pause or inflection? Ethereum turned net positive (+0.4%) and Base went negative (-1.3%) this week, breaking the quarter-long Ethereum→Base trend for the first time. One week is noise; two weeks would be a signal. We're watching whether the drift resumes or has genuinely inflected.

Morpho vs Aave V3 vs Compound V3 on USDC, Ethereum, 7d trailing supply APY (gross): Morpho 4.39% · Aave V3 ~4.2% · Compound V3 ~3.9%. (Aave and Compound sampled from their UIs at issue time; Morpho figure is universe-weighted USDC.) The Morpho premium over Aave on USDC is ~0% this week — back toward the trailing-month median after last week's compression, as Morpho's USDC cohort firmed.

On USDT, Ethereum: Morpho 3.50% · Aave V3 ~4.6% · Compound V3 N/A. Notably, Morpho's USDT cohort sits below Aave this week — a reversal of the usual pattern. USDT borrow demand on Morpho softened relative to the universal pool; the USDT cohort is tight (P25–P75 3.0%–3.4%) and currently unremarkable.

On Base, Aave V3 USDC supply APY ~4.4%; Morpho's Base USDC vaults weighted ~4.4%. The premium has been remarkably stable on Base over the quarter — though note Base saw net outflow this week (§03), the first time we've recorded the drift pausing. The yield premium and the flow direction diverged this week; worth watching whether flow follows yield back.

WETH remains the cohort where Morpho is not winning. Universe-weighted WETH APY 2.45% sits below Spark's stETH-leveraged route (~2.7%) and barely above Aave's WETH supply. The conclusion from issue 1 stands: WETH allocators at scale should not assume Morpho is the default route — the case is per-vault, not per-cohort.

3-month US T-bill yield: 4.18%. The institutional benchmark every yield-allocator implicitly carries. A Morpho USDC vault at 4.4% (current universe-weighted USDC) pays a 0% spread over T-bills; the cohort top quartile at 5.5% pays 1%. Any sustainable yield ≤4.5% on a stable vault is, by simple comparison, a worse risk-adjusted trade than the bill — and the universe-weighted sustainable USDC is 4.4%, so the median vault clears that bar only modestly. Selection is the whole game.

Stablecoin context: USDC market cap ~$54B, USDT ~$155B, PYUSD ~$1.3B, RLUSD ~$2.0B (CoinGecko, issue time). PYUSD and RLUSD's combined ~$3.3B is a rounding error against USDT — yet they are 24.2% of the Morpho universe against USDT's 6.6%, and that gap widened this week. The regulated-issuer stables continue to punch far above their broader-market weight in curated lending — the clearest structural signal of where institutional supply prefers to sit, and it strengthened over the fortnight.

LST APY benchmark: stETH 2.93%, rETH 2.80%, cbETH 2.78% — flat over 90 days. Morpho's WETH cohort weighted sustainable of 2.43% sits below the direct LST baseline — confirming again that WETH on Morpho isn't earning a meaningful spread over simply holding stETH.

Data sources. Per-vault state: Morpho Blue GraphQL indexer. Cross-protocol rates: Aave V3 UiPoolDataProvider, Compound V3 subgraph (sampled at issue time). Stablecoin market caps: CoinGecko. T-bill yield: US Treasury Direct. Every cited number traces to one of these sources and the methodology link below.

Methodology. Risk score, complexity score, depeg composite, sustainable-APY computation, and yield-anomaly z-score — all documented at sharpe.finance/methodology. No black-box scoring. Re-running the methodology against the same sampled state produces the same composite — the payload hash below is the integrity check.

Sampling. Figures sampled at the close of the week ending Saturday 2026-05-30 UTC; window deltas computed against the matching time 7 days prior. Vaults below $1M TVL are filtered out of cohort statistics (movers, anomalies) to avoid dust noise; they remain in the universe-level footprint. Vaults with reported APY ≥50% are filtered out of weighted aggregates as pathological. This is the W22 catch-up issue (published after the W22 Saturday slot); the window and cadence return to the standard Saturday close from W23.

Glossary. Sustainable APY: spot APY net of the reward-emission slice — the rate that survives an immediate emission rolloff. Cohort z-score: vault sustainable APY normalised to its (loan asset) peer distribution. Curator HHI: Herfindahl-Hirschman index of curator-level TVL shares. LLTV: per-market liquidation LTV, immutable post-deployment. Supply-constrained: a market where raising the supply cap is immediately met with deposits — borrow demand exceeds supplied capital.

Payload hash. 8b5fda93394d65a94e4255c3… (sha-256 over the frozen data JSON).