Sky's USDS supply keeps climbing through 2026 while the rest of DeFi is bleeding TVL. The yield is real, the mechanism is documented, the rebrand from MakerDAO is now two years old. The panda has been watching it for months, looking for the leak. There is one. It is just quieter than the marketing.
From Maker to Sky: a quick origin recap
MakerDAO rebranded to Sky Protocol in August 2024. DAI did not disappear, but USDS became the new flagship stablecoin, with a one-for-one wrapper that lets holders flip between them. The point of the rebrand was practical, not poetic: a cleaner governance design (SKY replaces MKR at a 24,000-to-1 ratio), a new savings product (the Sky Savings Rate, or SSR), and a structure built around endgame "SubDAOs" that nobody has fully delivered yet.
Two years in, the SubDAO part still reads more like roadmap than reality. The stablecoin part, however, shipped. USDS is the largest non-USDT, non-USDC, non-FDUSD stablecoin by supply in mid-2026, and it pays a savings rate that most stablecoin issuers cannot match without inventing risk.
How does the Sky Savings Rate actually work?
This is the part most coverage skips. The SSR is funded by Sky's collateral portfolio. According to the Sky documentation, the protocol's reserves are a mix of three buckets: tokenized US Treasury bills (mostly via Monetalis, BlockTower and Spark Tokenization), crypto-backed Vaults (ETH, stETH, wstETH posting collateral), and Spark Protocol lending markets that recycle USDS into yield-bearing positions.
The mechanism is mechanically simple. Borrowers pay a Stability Fee on Vaults. T-bill allocators earn off-chain yield and remit it on-chain. Spark earns a lending spread. All three feed a Surplus Buffer. The SSR is set by governance at a rate the buffer can sustainably pay. As of June 2026, DefiLlama tracks Sky at multi-billion-dollar TVL, with the savings module concentrating the bulk of that figure.
So the yield is not free money printed by inflation. It is real cashflow from real borrowers and real T-bills, paid out under a rate ceiling that governance can adjust. That is closer to a money market fund than to a Ponzi.
Where USDS sits on the DefiLlama chart
The wider context matters. According to DefiLlama, total DeFi TVL stood at $72.03B on June 5, 2026, with Ethereum holding $37.62B of that. Solana is at $4.82B and BSC at $5.15B, bleeding 5.74% week over week. Sky is concentrated on Ethereum and Spark, which keeps its risk profile readable: one chain to watch, one allocator policy to track.
For perspective on rivals, Ethena's USDe leans on perp funding rates. Tether and Circle lean on bank deposits and corporate paper. Sky leans on a documented mix that is roughly half T-bill, half crypto collateral. Each model breaks differently under stress. They do not all break at the same time. That is the point.
For the broader picture on how issuers fund their balance sheets, see the stablecoin issuer treasury stack. For how lending TVL is migrating across protocols, the Morpho and Aave curator economy covers the same plumbing from a different angle.
Risks
Sky is not risk-free.
Smart-contract risk: the Maker and Sky codebase has years of audit history, but the Spark fork, the SSR module and the new Allocator system are recent additions. More surface area, more bugs to find.
Oracle risk: Vault liquidations depend on Chainlink feeds, MakerDAO-native medians and the new Allocator oracles. A wrong price during a fast wick is the classic DeFi failure mode, and Maker has been bitten before (Black Thursday, March 2020).
Governance risk: SKY voters set the SSR, the Stability Fees, and the RWA allocator policy. A bad governance decision can drain the Surplus Buffer in a quarter. The MKR-to-SKY migration also fragmented voter turnout for several quarters in 2025.
RWA counterparty risk: a chunk of reserves sits with off-chain allocators who hold T-bills with traditional custodians. If one of those custodians has a problem, USDS has the same problem. Trustless onchain this is not.
Peg risk: USDS targets one dollar. The Peg Stability Module burns and mints to defend it, but during a fast outflow event a peg can wobble before policy kicks in.
Nothing here is catastrophic in normal conditions. All of it matters in a tail event.
What to watch next
Three signals over the next quarter. First, the SSR level itself: governance has cut it before when the buffer thinned, so a downward move would say the system is stressed. Second, the USDS supply trajectory: it has grown roughly in line with rate competitiveness, so a flat curve means the yield premium is gone. Third, the share of reserves sitting in tokenized T-bills versus crypto Vaults: drift toward RWAs makes Sky more bank-like, which has political costs in 2026's regulatory climate.
For BSC and Dadacoin holders the read is sideways. Sky's footprint on BSC is negligible, and that is fine. The thesis here is not "buy USDS." It is that a stablecoin paying mechanically funded yield, with transparent accounting, becomes a benchmark the rest of the market quietly competes against. For the broader map of yield, lending and stables, the DeFi cluster pillar collects the rest of the angles.
The numbers add up. The mechanism is documented. The panda does not love every piece of it. But he respects an engine that runs on cashflow instead of vibes.
