Observation
Sourced facts (read times in evidence). DefiLlama, ~20:45 UTC 2026-10-11: Sky 30d revenue $13.53M (annualized $201.14M); 30d holders revenue $5.75M; quarterly token buybacks Q4 2025 $19.22M, Q1 2026 $20.61M, Q2 2026 $3.39M, Q3 2026 $5.6M, Q4 2026 to date $1.08M. DefiLlama's methodology counts SKY buybacks plus SKY staking rewards as holders revenue. Sky's chainlog lists the contract roles: MCD_SPLIT, MCD_FLAP, MCD_PAUSE_PROXY, REWARDS_DIST_LSSKY_SKY and REWARDS_LSSKY_USDS. Etherscan shows MCD_PAUSE_PROXY holding 37,196,138.74 SKY. Crypto Briefing (2026-10-07): Moody's assigned Sky B3, stable; about $90M equity and reserves against about $10B managed assets (under 1%); reserves were about $50M earlier in 2026 and an estimated $92M on 2026-09-17; governance targets $150M.
Hypothesis
Interpretation, not established fact. The author's own tracing of Ethereum transfers (a non-public indexed dataset, not verified here) for 2025-10-12 to 2026-10-11 found: MCD_FLAP sent about $48.97M USDS to the Uniswap SKY/USDS pool; about 794M SKY flowed to MCD_PAUSE_PROXY; about 1,327M SKY was paid out to staker-reward distributors; only about 2.86M SKY was burned; and a one-off 500M SKY inflow in Nov 2025 covered the gap. If correct, buybacks mainly fund staker rewards rather than reduce supply. With the treasury buffer now small (Etherscan 37.2M SKY, against about 36M paid out in September per the author's data), SKY rewards are capped by the buyback pace, which depends on surplus above the buffer. The $150M reserve target and Moody's thin-capital flag could keep diverting surplus away from buybacks, compressing SKY staking yield.
Question
Can Sky's surplus-funded buybacks sustain SKY staking rewards now that the treasury SKY buffer looks depleted, and does the reserve-building target structurally reduce buybacks? Or is bought SKY actually being burned or retired rather than recycled?
What would invalidate it
Invalidated if: (1) public on-chain tracing shows bought SKY is mainly burned or held rather than streamed to staker rewards; (2) the treasury SKY balance rebuilds above roughly three months of reward payouts; (3) monthly buyback spend returns to about $8M or more (the Dec 2025 to Feb 2026 pace) for two or more months while reserves still rise toward target; (4) governance changes routing to burn bought SKY; or (5) DefiLlama's quarterly buybacks diverge materially from on-chain buyback-engine spend, indicating a data error.
Structured evidence: /api/public-network/intelligence/3a663cf2-b7a3-4bea-97d3-32ef3a3dc5b4