Sky: DeFi's Biggest Yield Fund

Every Monday, we zoom in on a specific project that offers unique and attractive opportunities to earn yield. We tell you exactly how we’re using the project to put money to work in the Machines & Money Portfolio, and show how you can do the same.

You can find our most recent portfolio report here.

Want daily DeFi insights? Join our Telegram group!

We believe that no newsletter should simply be talking at people. We want to foster productive discussions and have an ongoing relationship with each subscriber. The Telegram exists for exactly that reason.

If you missed the original article, here's the 30-second version.

Sky Agents are independent capital allocators that use Sky’s USDS stablecoin to generate yield across various specialized strategies. Additionally, a portion of what they generate flows back to Sky as protocol revenue.

That revenue is one of the main inputs governance uses to calibrate the Sky Savings Rate (SSR) – the yield behind sUSDS, which is currently sitting at ~3.65% APY with ~$6.2B supplied.

The model is simple but powerful; instead of Sky directly underwriting every yield source, it outsources allocation to specialists. Sky stays focused on risk management and governance. Agents handle execution.

The more Agents in the network, and the more diverse their strategies, the more resilient the foundation underneath sUSDS becomes.

That's what makes Obex Cohort 1 a big deal.

For context, Obex operates with a $2.5B mandate from Sky to identify and onboard institutional-grade allocators. Cohort 1 uses roughly 40% of that mandate.

Here's who's joining and what each brings to the table:

One of the most established onchain capital markets platforms, Maple has originated over $12B in loans since 2021. As a Sky Agent, Maple deploys USDS into institutional lending strategies – connecting Sky's liquidity to professionally managed credit pools. Reports suggest Maple has already received around $600M of the cohort's capital through its credit markets.

Securitize is the leading tokenization platform, managing over $4B in tokenized assets, backed by BlackRock and Morgan Stanley, and registered with the SEC as a broker-dealer and digital transfer agent. They issue tokenized products from BlackRock, Apollo, and KKR, among others.

Centrifuge has brought $1.3B+ in tokenized assets onchain, specializing in institutional-grade credit, treasury, and fund products. As an Agent, they channel USDS into tokenized credit strategies.

Better is a fintech mortgage and home equity firm with $110B+ in cumulative loan volume. Sky's partnership brings $500M of USDS into Better's platform, opening up mortgage yield, HELOCs, and loan sale proceeds as new income streams for USDS holders.

A relatively new protocol focused on backing AI compute infrastructure, USDai directs yield toward USDS via GPU/data center financing returns.

Daylight allocates capital toward solar and energy projects – the kind of real-world assets that traditionally sit far outside DeFi.

River deploys USDS into structured credit products, expanding the credit risk profile available to Sky.

TVL is focused on crypto capital markets and chain-traded products – bringing crypto-native yield strategies into the Agent Network.

The pre-Obex Sky yield base relied heavily on a few sources: ETH-collateralized loans, US Treasury allocations, and Spark vault revenue.

Cohort 1 widens that significantly:

This is the diversification flywheel: each Agent contributes a different revenue stream, and the SSR sits on top of a wider, less correlated yield base.

If Treasury rates fall, mortgage yields hold up. If onchain lending slows, RWA credit keeps generating. If one sector underperforms, the others absorb it.

Sky isn't betting on a single yield source. It's building a portfolio.

Now to the second piece – the proposed structural changes to how Sky handles its reserves and pays out staking rewards.

There are two upgrades on the table:

The Backstop is Sky's solvency buffer – the reserve that absorbs bad debt before it ever reaches USDS holders. It's the first line of defense for every dollar in the system.

The proposal raises the Target Aggregate Backstop Capital from 125M to 150M USDS – a 20% increase.

The current Backstop is already sufficient by Sky's own risk model. This just makes it stronger. With $1B+ flowing into newer, more diverse strategies via Obex Cohort 1, having a deeper buffer makes sense – more yield sources means more potential failure points, even if individual exposures are well-managed.

This is the more interesting one for SKY holders.

Right now, SKY staking rewards are funded through a flat allocation – mostly using treasury reserves to fund SKY buybacks that get redistributed to stakers. The proposal replaces this with a three-stage transition model:

Stage 0 (Current) Reflects the changes made on March 14, 2026, when Sky temporarily cut SKY buybacks from 300,000 USDS/day down to 37,600 USDS/day for an estimated three months. This was the first move to redirect more capital toward the Backstop.

Stage 1 (Next Step) Staking reward funding drops from 75% to 50% of the prior month's protocol surplus. SKY buybacks and Surplus Buffer allocation stay the same. Activated by executive vote.

Stage 2 (The Sustainable Model) This is the long-term endgame:

Stage 2 activates only when SKY treasury reserves hit a predefined threshold.

For SKY stakers: rewards become fully revenue-funded rather than partially treasury-funded. The current 10.8% APY (mentioned in the original article) will likely compress in the near term as buybacks remain reduced. But it becomes far more durable long term – tied directly to protocol revenue, which has been growing (Sky did $53M in gross revenue in Q1 2026 alone).

The reintroduction of USDS staking rewards is the part worth watching. SKY stakers will eventually be able to choose between getting paid in SKY (for upside exposure) or USDS (for stable yield). That's a meaningful flexibility upgrade.

For sUSDS holders: the SSR isn't directly changed by these proposals. But a stronger Backstop + a broader Agent revenue base = a more sustainable foundation for the rate going forward. The 3.75% APY today is sitting on a much wider yield base than it was six months ago.

Both proposals point in the same direction: prioritizing long-term protocol stability over short-term yield.

The trade-off is that SKY staking yields will be lower in the short term. But the underlying engine generating those yields is becoming more durable – and the diversification of the Agent Network gives Sky a real path to scaling USDS without scaling risk proportionally.

For reference, USDS supply grew 86% in 2025 to $9.86B, and projections for total USDS supply in 2026 are around $20.6B. The infrastructure changes happening right now are designed to support that kind of growth without breaking the risk model underneath it.

If you hold SKY or sUSDS, you can participate in governance at vote.sky.money.