Q1 2026

Given that this is the inaugural quarterly letter, I do not have the luxury of pointing to prior letters to reference historical positioning so let me begin by giving a brief overview of how 2025 came to a close.

2025 was a good year with profit taking on core strategic positions and a series of shorter term tactical trades contributing to strong performance through year end. By mid Summer, I had become particularly disillusioned with the state of DATs and shifted to a more defensive posture. While this meant missing out on the final throes of the DAT mania, it enabled me to see the writing on the wall sooner than the average market participant, giving way to some profitable short positions in ETH and a smattering of alts through Q4. More importantly, shifting to a defensive posture allowed me to largely avoid a steep drawdown from the portfolio high watermark.

While this defensive posture broadly prevailed through Q1, I have started to selectively add back long exposure, but more on that later, let’s first discuss what did and did not work in Q1.

The Good

Coming into the year with a heavy stablecoin allocation has meant looking for opportunities to generate good risk-adjusted yield.

Despite broadly bearish market conditions there have been a few good spots offering some combination of good risk-adjusted real yield plus optionality in the form of points/future token incentives.

Concluded in late Q4/Q1:

• Stable Pre-Deposit: pre-deposit campaign for a new Stablecoin-focussed Layer 1, realised at 28% annualised.

• USDai: multiple real yield opportunities (looping, peg trading, Pendle LPs) while generating CHIP exposure, realised yield is yet to be seen as TGE looms closer.

• opusIAUSD: a more niche capacity constrained opportunity realising 40% annualised.

• LLP (+ LIT airdrop): farming the farmers via LLP which became less attractive post TGE, LIT airdrop was thanks to activity generated on Lighter through the year.

Ongoing:

• USDG and GHO on Tydro: a bet on INK where even conservative assumptions make for an attractive risk-adjusted expected return.

• Nado NLP: another way to bet on INK while being paid to take the other side of perp traders.

• ONyc looped on Kamino: new and growing reinsurance protocol on Solana.

• reUSD looped on Fluid: attractive real yield combined with points upside into Re TGE.

• apxUSD and apyUSD Pendle LPs: likely the first STRC backed yield coin to TGE, Apyx Season 1 distributes 5% of supply over 12 weeks.

• Payy: PR around recent fundraise opened the door to a token which was already looking likely based on the announcement of Payy Network.

• Selective basis trading across venues

• Polymarket (read more about the approach here, looking forward to providing an update in the next letter)

Future opportunities:

• Saturn

• Daylight

• Variational OLP

Protocols have learned from past cycles and increasingly obscure the details of their incentive programs to prevent mercenary capital from gaming eligibility criteria. The result is that farmers are increasingly making allocation decisions with incomplete information: committing capital and time based on inferred rather than explicit reward structures.

The sweet spot is finding opportunities where the expected return based on conservative assumptions handily outweighs the estimated risk while trying to retain diversity across opportunities for risk management purposes.

Subscribe now

The Bad & The Ugly

Unsurprisingly, all losses incurred this quarter have been associated with long exposure.

Attempting to take short-term long exposure in BTC has largely led to chop despite continued confidence in the structural view. Saylor’s STRC preferred equity instrument has been returning to par with increasing speed post dividend payments, attracting demand from fixed income ETFs and on-chain STRC backed yield coins are gaining traction. The thesis may still be correct, but expressing it as an outright long in a risk-off environment has been a mistake thus far. A relative expression (long BTC against a hedge) would have isolated the signal I was actually trading.

Other than getting chopped up on BTC, AZTEC has been the more interesting story. The idea was to play the wrinkles of the auction mechanics by participating in the Uniswap CCA on Day 1 of early bidding with a maximum sized bid at a price just above the floor. The CCA spreads bids across remaining blocks and clears each at a uniform price so by bidding early and low, fills concentrate in the cheapest blocks before later demand pushes the clearing price higher. Bid size matters here: if you assume the sale sells out, you know the clearing price will rise past your bid, meaning your order gets priced out after Day 1 and the remaining participants spend the rest of the auction buying above you. Max size at the floor maximises your fill in exactly those cheapest blocks.

The idea worked initially: with the average entry at $0.029, the auction cleared 63% above floor at $0.0473, which is also the price at which the Uniswap v4 pool launches at TGE. What I failed to foresee was the effect of a parallel reference price at TGE. Binance opened a pre-market which traded well below the CCA clearing level, creating block 1 MEV at TGE whereby any liquidity at the clearing price was taken instantly against the Binance price. The exit was not executed in block 1, the pool converged to Binance futures and the position was closed at a loss.

The two losses tell different stories. The BTC position was a case of reaching for directional exposure in an environment that didn’t support it. The structural thesis may prove correct in time, but expressing it as an outright long against a risk-off backdrop was the wrong trade at the wrong time. AZTEC was a different kind of error: the mechanical thesis played out exactly as designed, but I failed to map the full outcome tree at TGE, specifically the impact of a parallel Binance reference price on the Uniswap pool.

Two different errors, but the same lesson: in a quarter where the right macro call was to wait, the losses came exclusively from the trades where I didn’t.

Q2 Outlook

As alluded to, I have been selectively adding back long exposure throughout the quarter, for now this has been limited to HYPE and BTC.

HYPE is very well understood and my thesis is not unique or differentiated so I will not bore you with a bullish take here.

BTC is a little more interesting. Throughout the Iran war saga, BTC has failed to make new lows, showing significant resilience against broader risk. It has shown equal resilience through a fresh wave of quantum fears per Google’s paper. When an asset refuses to go down in the face of bad news, it is a sign of strength, signalling seller exhaustion.

Meanwhile, the structural STRC thesis outlined above remains intact, and the on-chain layer is accelerating. Saturn, the highest quality STRC-backed yield coin, is launching imminently. With Apyx already managing $100m+ TVL, Saturn could reach multiples of that.

Even if STRC is the only reason for the resilience in BTC, performance is performance and the relative outperformance we have seen could provide the seeds for a strong narrative if BTC starts to wake up a little here. The BTC long position taken in Q1 was likely premature but the stars seem to be aligning for better times in Q2.

I am actively monitoring other names and open to adding broader long exposure. I look forward to sharing ideas and deep dives here going forward.

At a roughly $8-10 billion private valuation, there are two ways to get exposure to Polymarket: buy in at those terms or farm your way in at a fraction of the cost.

I chose the latter. Not because I’m bearish on Polymarket; volumes have surpassed the November 2024 election highs, NYSE’s parent company took a $2 billion stake, and their CMO effectively confirmed a token on a podcast in October 2025: “There will be a token, there will be an airdrop.” I chose it because farming allows you to build exposure to a potentially highly valued token at a small to negative cost while retaining optionality and liquidity. At current private valuations, a lot of the positive sentiment looks to be already priced in.

Given the reliance on UMA from an oracle standpoint and Polygon from a chain standpoint, it’s not hard to imagine the case for $POLY token utility either. And with competition heating up; Kalshi growing its regulated offering and Hyperliquid sitting on significant $HYPE supply available as incentives if they choose to pour fuel on HIP-4, Polymarket’s incentive to launch a token is only increasing.

Roughly six months in, with plenty of pivots and dead ends along the way, this is a practitioner’s guide to how I’ve been positioning.

Eligibility

Historical precedent tells us that heavy, consistent and early usage across the core product is what counts when it comes to airdrop eligibility. In the context of Polymarket there are a few different ways to measure usage: making a deposit and holding a balance, active months/weeks/days, trading volume, number of trades, number of markets traded, PnL, and liquidity provision (limit orders filled, limit orders placed, and Liquidity Rewards earned).

Eligibility likely combines some of these factors. The most pertinent question is which matters most. Historical precedent offers some guidance:

The common thread: protocols reward the behaviour they need most. For a prediction market, it’s liquidity.

We can simply listen to what Polymarket are telling us. Liquidity provision is being incentivised in the form of Liquidity Rewards, whereby LPs are paid USDC to provide liquidity based on certain criteria. If they’re incentivising it now with USDC, what’s the chance it gets rewarded heavily in airdrop eligibility and perhaps $POLY rewards when the time comes?

This incentivisation makes sense. Without deep order books and tight spreads the product is basically unusable. Retail want to get the price they see. Institutions want to express views with meaningful size and get out if they need to. Commentators want to cite odds from thick markets where the price is real and not the midpoint of a 10c spread with a few thousand dollars either side.

The approach to date

Despite a bias towards liquidity provision in terms of eligibility maximisation, my initial framework was simpler (as outlined in the initial piece I wrote about Polymarket farming last October):

Make economically rational decisions in a vacuum; decisions that make sense even if the airdrop never happens but which offer large asymmetric upside if the thesis is correct.

Or more plainly: don’t burn money chasing an airdrop that might never come, or one you might not qualify for.

The approach centred on making a real absolute return while organically generating activity across many of the metrics listed above. For the most part this involved buying high probability markets (FOMC decisions, betting against insolvencies/depegs etc.) along with making a handful of crypto-focussed TGE valuation bets.

All in all the results have been acceptable, generating roughly ~7% over 6 months on the capital deployed, around 14% annualised without factoring in any potential airdrop eligibility generated.

Despite a relatively small allocation to this strategy (~4% of book), it has generally been difficult to achieve high utilisation of the capital, limited by the opportunity set and often-times poor liquidity on more niche markets such as the TGE valuation bets.

This led to exploring the possibility of automated strategies that could generate high levels of activity with relatively small levels of capital. When manually placing orders I noticed that very often there would be an abundance of liquidity right at the extreme of the pricing range (99.9c):

I began to wonder whether I could capitalise on stale limit orders closer to the resolution or at the resolution of these markets. The great allure of risk-free arbitrage...

UMA resolves markets through a multi-step on-chain process (proposal, optional dispute, final settlement) with each step involving real money (disputers bond 750 USDC, proposers stake capital on outcomes). Surely, somewhere in that lifecycle, the oracle knows something the order book hasn’t priced yet.

I spent the better part of three weeks on this. Mapping the oracle lifecycle, monitoring the mempool, writing scripts to detect timing gaps between settlement events and order book repricing. It felt like the kind of inefficiency that could exist in a fairly immature market.

It doesn’t.

The reason is straightforward: market makers are sophisticated and are monitoring the mempool for pending oracle transactions to reprice orders before those transactions confirm on-chain. By the time a settlement, proposal, or dispute event is visible, the order book already reflects it.

Whenever I found a genuine gap between the oracle and the CLOB, it was on a market nobody was trading: obscure golf markets, niche soccer goalscorer markets, esports etc. The pattern was the same everywhere. If a market had enough liquidity to make the arbitrage worthwhile, the market makers had already priced the oracle event. If they hadn’t, there was nothing to trade against.

Every path that tried to extract edge from the resolution side came up short. Which is what pointed me back towards providing the liquidity itself.

Burn to earn: the pivot to LP

Despite leaning towards liquidity provision as the likely apex predator of airdrop eligibility, my initial reservation was that I expected it to be a negative expected value activity. My goal was to make money while generating potential eligibility as a by-product. Paying to provide liquidity felt like the kind of trap that separates smart farmers from desperate ones.

My viewpoint changed when I saw the remarkably low barrier to being a top LP as measured by liquidity rewards. As of February 2026:

• 972 wallets have earned over $1,000 in rewards

• 4,429 wallets have earned over $100

• 16,109 wallets have earned over $10

• 38,028 wallets have earned over $1

Out of 2.18 million wallets, less than 2% have generated more than $1 in rewards. If liquidity provision is prioritised in terms of eligibility, and everything Polymarket is telling us suggests it will be, top LPs represent a very small cohort. A tiny percentage of wallets could stand to be beneficiaries of a disproportionate share of token supply.

Subscribe now

Reframing the economics

The question then becomes: what is the minimum I need to spend to build a credible LP footprint, and is that cost justified by the expected value? Think of it like buying a Pendle YT: you’re paying a known cost for exposure to an unknown upside, and the maths only need to work out roughly right for the trade to be good.

The approach became clear: take the cumulative profits earned from the initial directional strategy as a starting balance, re-allocate the principal elsewhere (partially solving the capital utilisation problem), and use LP rewards as a subsidy to fund presence across many markets simultaneously. No intention to profit from market-making but to minimise the net cost of maintaining qualifying LP presence for as long as possible, across as many markets as possible.

This is a fundamentally different optimisation target to traditional market-making. We want wide spreads (just tight enough to qualify for rewards), minimal fills (we assume every fill is adverse selection), and we accept negative PnL within a pre-defined budget. The metric that matters is the ratio of LP footprint to capital burned in achieving it.

How Liquidity Rewards work

Polymarket’s reward formula scores each qualifying limit order based on proximity to the midpoint and size, with a quadratic penalty for distance. For example, an order at half the maximum qualifying spread scores only 25% as much as one right at mid. Crucially, it rewards two-sided liquidity: your effective score is roughly the minimum of your bid and ask scores. A lopsided book scores poorly no matter how much capital you have on one side.

Your share of a market’s daily reward pool is your effective score divided by the sum of all qualifying LPs’ scores. More competition means smaller slices, which makes market selection as important as quote placement.

The approach: minimum viable presence

With rewards economics understood, the strategy becomes an optimisation problem: maximise breadth and consistency of LP presence while minimising capital at risk and adverse selection losses.

In practice this means quoting symmetric two-sided orders across a portfolio of reward-eligible markets, sized at the minimum qualifying threshold with a small buffer rather than trying to capture meaningful reward share in any single market. We’re not competing with professional market makers on execution speed or pricing accuracy. We’re competing with other airdrop farmers on consistency, breadth and patience.

Market selection is the most consequential decision. Polymarket typically has 30,000+ active markets but reward pools and competition vary enormously. The scoring heuristic is intuitive:

Controlling the burn

Risk management operates at multiple levels. Individual markets are monitored on a rolling burn-to-reward ratio: if the cost of adverse selection in a given market starts running meaningfully ahead of the rewards it generates, the market gets flagged and eventually halted. At the portfolio level, a hard daily burn cap acts as a backstop where even a catastrophic day stays survivable, giving the strategy a long runway.

Additional protections layer underneath such as anomaly detection on individual fills, inventory limits, and automatic wind-down near resolution but the daily burn cap is the one that matters most. We can tolerate burning slowly and predictably. We cannot tolerate burning quickly and unexpectedly.

Risks

There are several ways this trade fails:

• The token never materialises. Despite the CMO’s comments and the broader trajectory, there is no binding commitment. Polymarket could delay indefinitely or decide that a token doesn’t serve their interests. If this happens, the capital burned on the LP strategy represents an absolute loss.

• Opportunity cost. Capital deployed to LP on Polymarket is capital not deployed elsewhere.

Beyond these, there are several secondary risks worth noting. Eligibility criteria could underweight LP in favour of volume or number of markets traded, though the activity generated in the earlier directional phase provides some hedge. Sybil filtering could be aggressive enough to catch multi-wallet operations but genuine cross-strategy trading history should differentiate. And adverse selection could exceed burn budget in a particularly volatile period, though the circuit breakers and daily caps are designed to make that a slow leak rather than a blowout.

Expected Value

A rough scenario sketch makes the asymmetry tangible. Assume a conservative 5% of supply distributed to users at an $8 billion FDV: that’s $400 million. Now assume LP-weighted activity receives 30% of that allocation (a reasonable assumption if LP is indeed the priority behaviour): $120 million. Distribute that across the fewer than 5,000 wallets with over $100 in cumulative LP rewards and you get an average allocation north of $20,000 per wallet before any tiering that would skew further towards top LPs.

The numbers shift dramatically with different assumptions. At 10% of supply and 50% LP weighting the per-wallet figure for top-tier LPs could be multiples higher. Across a wide range of reasonable scenarios, the expected value per qualifying wallet exceeds the total cost of farming by a significant margin.

These are illustrative, not predictions. The point is that even the conservative end of the range produces a compelling multiple relative to the farming cost.

The key is ensuring the downside remains bounded. The entire strategy is designed around this constraint: profits from the initial phase fund the burn, rewards subsidise the cost, circuit breakers prevent blowups, and the total budget is sized such that a complete loss is tolerable.

We’re not betting the farm on an airdrop but rather spending a bounded amount on an asymmetric option.

What’s next

There are a few things I’m watching closely. Any changes to the Liquidity rewards program (adjustments to qualifying criteria or reward pool sizes) would warrant a reassessment of the strategy’s parameters. Similarly, any official communication from the Polymarket team regarding token timelines or eligibility criteria would reshape the calculus, potentially dramatically. And the competitive landscape: if Hyperliquid begins pulling the HIP-4 lever in earnest, Polymarket’s incentive to accelerate their own token launch increases which could compress the farming window.

As this piece goes to press, Polymarket has launched a referral program (30% revenue share on direct referrals) and introduced taker fees across nearly all market categories. A portion of these fees fund expanded maker rebates, providing an additional USDC subsidy to liquidity providers on top of the existing Liquidity Rewards, further reducing the net cost of being present. The introduction of taker fees has been controversial but for LP-focused strategies, the dynamic is inverted: fees are collected from takers and redistributed to makers, adding another subsidy layer to the economics of providing liquidity and confirming Polymarket’s internal focus on improving liquidity.

The first quarterly letter covering Q1 positioning and performance is coming in early April where I’ll be discussing what else has been taking place across the book and outlook into Q2. In the meantime, if you have questions or are running your own Polymarket strategy, I’d be interested to hear about it, don’t hesitate to reach out on X.

Parsec Finance is winding down, but this publication is not. Today I am going to explain what's changing and why.

For those of you who didn’t see last month’s news: after five years of operation, Parsec announced in mid-February that it was winding down, with the company ceasing to operate by month end.

Having joined in early 2023, I spent nearly three years working directly alongside Will (Parsec’s founder) on everything non-product and non-engineering. Writing the Parsec Weekly has been a core piece of that, and over time, the publication has grown to over 4,600 subscribers with consistent institutional readership and led to collaborative work with projects like Maple Finance, Aztec Network, and Cork Protocol. It became clear this was something worth continuing.

As we closed up shop at Parsec, Will made the generous offer to transfer the publication to me, and I jumped at it.

Before I get into where this publication is going, let me pause to introduce myself more formally to those who don’t know me.

My background spans traditional finance and crypto. Before joining Parsec in early 2023, I worked under Taiki Maeda as an analyst, producing research on liquid opportunities across market-neutral and directional strategies. I’ve been actively trading since early 2020 but a brutal drawdown I suffered in 2022 taught me scepticism and risk management the hard way - lessons that have shaped much of what’s followed since.

A few of the larger trades I’ve participated in include the JTO airdrop, the SOL repricing from $20 to $200, and the HYPE airdrop and TGE trade from $3 to $30. Naturally, these home runs sit alongside a long list of losers and a long list of smaller winning trades and farms that have allowed me to compound returns over time.

This market participation now sits inside Hepworth Iron Capital, a proprietary trading firm focused on discretionary long/short equity and digital asset strategies (both market neutral and directional).

You’ll notice the publication has been rebranded accordingly - you can find more about the firm at hepworthiron.capital. There’s plenty of good macro and traditional finance research and plenty of good crypto content, but very little sits at the intersection - that’s where this publication will live. Going forward, you can expect the scope to be much broader than it was under Parsec Research:

• Trade ideas and theses translated into specific, actionable positions across equities and digital assets

• Process reflections: what’s working, what isn’t, and how our frameworks are evolving

• Quarterly letters: performance, positioning, and strategic rationale across all Hepworth Iron Capital activities

Expect regular posts with quarterly letters at each period end.

If you subscribed purely for Parsec product updates, I'd encourage you to stick around for a few posts to see if the new direction resonates. If it's not for you, no hard feelings.

The first quarterly letter covering Q1 positioning and performance is coming in April. But before that, expect a deep dive on what could be the largest airdrop opportunity of 2026 - how we’ve been farming it and how our approach is about to change.

DeFi meets Duration: Cork Protocol

-Charlie

Last week I wrote about Supernova and how their interest rate swaps might finally enable fixed-rate lending in DeFi. This week I want to stay on the same thread and explore another major gap in DeFi’s financial infrastructure.

We are increasingly seeing a broader set of tokenised assets moving onchain but the tooling to price, hedge, and trade the risk embedded in these assets hasn’t kept up. If DeFi wants to graduate from a playground into a legitimate financial system, it needs a risk layer. That’s what Cork is building.

The Risk Gap

While still a drop in the ocean, tokenised assets (not including stablecoins) represent a growing market. As the market scales and the tokenised asset set grows broader, so too does the risk surface area.

Increasingly we are seeing tokenised private credit (Apollo’s sACRED, Midas/Fasanara’s mF-ONE) start to be used as collateral in DeFi money markets. Steakhouse Financial recently published an excellent deep dive on this trend. The core challenge is that for crypto-native collateral like cbBTC, DeFi repo works beautifully because DEX liquidity makes the collateral instantly liquidatable. For tokenised private credit, these guarantees break down as there is often no secondary liquidity onchain. Redemptions can often take weeks or months. This creates a massive duration mismatch between the on-demand nature of DeFi lending and the illiquidity inherent in private credit.

This isn’t hypothetical. In December 2025, Fasanara wrote down mF-ONE’s NAV by ~2% to reflect an updated valuation of the underlying private credit book. Steakhouse, who curate the Morpho vault exposed to mF-ONE, confirmed lenders were unaffected and that the LTV cushion and oracle discounting did their job. But the episode is instructive nonetheless; if the NAV can move against you and you can’t liquidate the collateral instantly, who’s holding the bag during that gap? For lenders used to instant liquidations, a mark-to-market adjustment on an illiquid asset with multi-week redemptions is a different beast entirely.

Cork Protocol

Cork is a programmable risk layer for onchain assets. The core primitive is the Swap Token; a tokenised derivative granting the holder the right to exchange a specified asset for a liquid collateral asset (like USDC) at pre-defined terms, at any time before expiry. Think of it as on-demand, pre-funded insurance for onchain assets.

Here’s how it works:

1. Every Cork market pairs a Reference Asset (e.g. a tokenised credit fund) with a Collateral Asset (e.g. USDC)

2. Collateral is deposited into a Cork Pool and split into two tokens:

   • Swap Token (cST): Allows you to swap your Reference Asset for the Collateral Asset on demand — the “protection” side

   • Principal Token (cPT): A residual claim on pool collateral at expiry — the “underwriting” side, you earn yield by absorbing risk

3. If the Reference Asset is impaired, illiquid, or depegs, the Swap Token holder can immediately swap it for liquid collateral. No withdrawal queues. No waiting for quarterly redemptions.

The beauty is that it’s pre-funded. The collateral backing the swap is already sitting in the pool before any claim occurs.

Phoenix, the Raise, and Earning Trust

Cork originally launched its public beta in March 2025 with Lido, EtherFi, Ethena, and Sky. In May 2025, the protocol was exploited for $12M through a vulnerability; four audits, formal verification and a bug bounty program had failed to catch the attack vector.

Rather than quietly disappearing, the team went through a comprehensive rebuild; significantly reducing smart contract complexity, migrating certain architecture to separate periphery contracts (following Morpho’s approach), introducing ERC-4626-style preview functions, and engaging ChainSecurity, Certora, Olympix and Hypernative across the full security stack. The team also published a formal security framework and, critically, all current and planned Cork markets are whitelisted — only vetted assets and counterparties can participate as they are targetting an institutional user base. Permissionless markets sit on the longer-term roadmap, but for now composability takes a back seat to controlled risk exposure. The result of the comprehensive rebuild is Cork Phoenix, which launched on Ethereum mainnet on January 19, 2026.

Shortly after, Cork announced a $5.5M seed round led by a16z CSX and Road Capital. One would expect a significant portion of these funds to go towards continued rigorous auditing and security hardening — exactly where it should be going for a protocol in this position.

That said, there is no shortcut to Lindyness. Cork will need to prove itself in the wild over time. A risk management protocol that suffered a $12M exploit has a higher bar to clear - time is the audit that matters most.

Use Cases

Cork’s use cases map neatly to the problems emerging across DeFi:

Protected Loops — The looping trade is one of DeFi’s most popular strategies but for tokenised RWAs with long redemption windows (30-90+ days), it carries enormous duration risk. Cork’s Swap Token provides an instant liquidity layer so that if the underlying RWA becomes illiquid or impaired, users can immediately swap for liquid collateral rather than waiting for issuer redemption.

Peg Stability — Stablecoin issuers can integrate Cork to offer their holders a mechanism to exit at or near par during stress events, enhancing user confidence and creating a market-based floor that can make depegs less frequent and less severe.

Duration Risk Management — Vault curators and bridge operators face their own duration mismatches. Both can leverage Cork pools as a composable liquidity buffer to support smooth redemptions during volatile conditions.

Cork’s integration with AggLayer’s Vault Bridge is an early example of this use case. The Vault Bridge rehypothecates bridged assets into Morpho vaults for yield, but at ~90% utilisation only ~10% of collateral is available for immediate withdrawal. Cork sits underneath as a pre-funded liquidity buffer so that users on chains like Katana can always redeem instantly while LPs earn a premium for guaranteeing that liquidity. Same pre-funded buffer, applied to bridge redemption risk rather than credit risk.

Institutional Liquidity Buffer — This connects directly to the Steakhouse thesis. If a risk curator is onboarding tokenised private credit to Morpho, Cork can sit alongside the lending market as a composable liquidity buffer to mitigate the duration mismatch for DeFi lenders.

The Bigger Picture

If you zoom out and look at the infrastructure stack being assembled right now, the outlines of a real financial market are forming onchain:

• Base Layer: Lending markets (Aave, Morpho, Euler) as repo desks

• Securitisation Layer: Apollo, BlackRock, Midas/Fasanara tokenising credit products as on-chain collateral

• Rate Layer: protocols like Supernova enabling interest rate swaps for fixed-rate exposure on top of variable-rate positions

• Risk Layer: Cork providing the ability to price, hedge, and trade credit/duration/depeg risk

In traditional finance, these layers have existed for decades. What’s happening in DeFi is a re-derivation of the same financial stack with programmable, composable, transparent primitives rather than bilateral OTC agreements. With protocols like Supernova and Cork sitting on top of Morpho and other repo market infrastructure, you get something that starts to look like a functioning onchain credit derivatives market.

My Take

Cork is tackling a genuinely important problem. The gap between the assets moving onchain and the risk infrastructure for those assets is real and widening. As a broader set of tokenised RWAs and structured credit comes onchain, something like Cork becomes increasingly necessary.

Whether they can earn back trust after the exploit and build sufficient liquidity depth before the next stress event are the key questions.

What I’m watching:

• Partnerships with asset issuers, risk curators or vault deployers for Cork Phoenix

• AggLayer Vault Bridge performance during stress

• Security track record over time

Will Rate Swaps Enable Fixed-Rate Lending in DeFi?

-Charlie

With MegaETH mainnet rapidly approaching (Feb 9th), I’ve been digging through Mega Mafia teams to ascertain whether there are any ecosystem projects of interest to me.

As a self-proclaimed DeFi nerd, I've been positively surprised by Supernova.

The thesis here is compelling: new fixed-rate money markets can’t scale because they require moving 100% of principal. Interest rate swaps, on the other hand, require a small % of notional and so can potentially enable a fixed-rate market to emerge and scale.

The Graveyard of Fixed-Rate Protocols

If you’ve been around long enough, you’ve seen multiple attempts at fixed-rate lending (Notional, Term Finance, TermMax) with no single project being able to sustainably scale over time.

The answer as to why is elegantly simple; fixed-rate lending protocols force both sides into what is often an impossible negotiation. Lenders won’t give up flexibility cheaply while borrowers won’t pay 10% fixed when Aave is floating at 5%.

Despite Pendle offering yield tokenisation, not fixed-rate lending, their PTs prove this dynamic on the supply side. Fixed rates with a lockup only work when yields are dramatically higher — think Ethena PTs at high double digits. But real borrowing demand simply can’t sustain such high rates consistently.

Supernova

This is where Supernova’s thesis gets interesting. From founder Nico Pei:

“for fixed-rate loans, borrowers don’t need to get a full loan term locked up for the fixed duration. Borrowers only need to get capital willing to absorb the spread between the agreed upon fixed rate and Aave’s floating rate, and get the rest from Aave, Morpho, Euler, or wherever they prefer”

Instead of building another fixed-rate lending protocol (which requires moving 100% of principal), Supernova is an interest rate derivatives market that enables fixed rates on existing money market positions without requiring users to move anything.

A $100,000 notional position on “Aave-USDC-30 days” would require ~$500 USDC in margin. This is the same capital efficiency that makes rate swaps an $18 trillion daily market in traditional finance.

How It Actually Works

The architecture consists of a vAMM and order book sitting side by side. It uses a constant-product AMM for ParRates (Supernova’s interest rate swap) combined with a CLOB for limit orders with orders automatically routing to whichever venue (order book or vAMM) offers the best price.

Behind this sits a System Vault that acts as a liquidity backstop, taking both long and short sides and absorbing net open interest imbalances across multiple durations.

Position mechanics are straightforward:

Long Rate (Pay Fixed, Earn Float) — This is what borrowers hedging floating liabilities want. Critically, these positions have no liquidation risk.

Short Rate (Earn Fixed, Pay Float) — Directional or speculative positions. These can be liquidated.

The fee structure is equally simple: 5 bps on notional for trading, 15 bps on the fixed-rate side for settlement.

To bootstrap liquidity, all fees flow to Vault LPs in the early phase with a “seeding premium” of 50 bps above floating (which should equate to an initial ~20% APR on the System Vault LP). The vault will support $100m of fixed-rate borrowing capacity on Day 1.

The Competitive Picture

It’s worth mapping where Supernova sits relative to other players in this space:

Boros focuses more on CEX perp funding rates but is showing impressive traction with $268m peak OI.

IPOR is a more direct competitor given that they focus on DeFi lending rates. They use an RFQ AMM with up to 500x leverage on 28/60/90 day tenors with their LP pool acting as a counterparty to all swaps. IPOR has an unfortunate quirk in that traders on IPOR must hold positions for a minimum of 1 week.

Supernova differentiates on execution mechanism (vAMM + CLOB), scope (includes rates, FX and cross-rates) and go-to-market (B2B/API-first along with retail frontend).

Supernova is betting on a superior execution model and, importantly, a more B2B-focussed GTM strategy (bringing fixed rates to where users and TVL already are). Their target audience spans six verticals:

Looping Strategies

Loopers can lock in guaranteed positive carry on PT loops when yields exceed implied borrow rates with a one-click API call rather than manually managing rate exposure. Crucially, they retain full flexibility to unwind principal from the underlying money market while maintaining fixed-rate exposure.

Institutional Borrowers

Supernova offers an SDK for fixed-rate borrowing on Aave and Morpho targeting institutions seeking a predictable cost of capital. Borrowers can lock in fixed rates while preserving the ability to return capital early, thereby avoiding rigid term lockups and maintaining balance-sheet flexibility. Supernova is targeting the likes of Galaxy Digital and Hyperithm here.

Risk Curators & Fixed-Rate Protocols

Supernova enables risk curators in Morpho v2 and Kamino fixed-rate markets to hedge interest-rate exposure while providing fixed-rate loans, removing the risk of being locked into a low fixed rate when underlying variable rates rise. This completes the fixed-rate borrowing stack: borrowers can either borrow directly from fixed-rate markets or borrow variable and hedge via Supernova. Their pipeline includes Gauntlet, Euler and Kamino.

On-chain Credit Markets

Fixed-rate credit on top of existing variable-rate on-chain credit markets is a core requirement for institutional borrowers. Supernova is in active integration planning with Cap Protocol, which lends to Amber, SIG, Flow Traders, and IMC Trading. On the consumer credit side, Supernova allows crypto neobanks like Ether.fi to offer users fixed-rate borrowing while sourcing liquidity from Aave and Morpho.

Rates Trading

Traders can express directional views on rates and convert small basis spreads into potentially attractive returns. Supernova promises simplified UX with leverage and TP/SL functions specifically designed for rates trading. Re7, Caladan, Flow Traders and Portofino are all key targets here.

Crypto Neobanks

Supernova enables savings apps and neobanks to offer users higher, predictable fixed-rate yields while deploying capital into Aave and Morpho. The SDK also supports synthetic FX exposure, letting users hold USDC/T while gaining exposure to alternative currencies. Their pipeline here includes Altitude (Squads), MoonPay, Aave App and Nook.

My Take

Conceptually, Supernova’s approach to enabling fixed-rate lending to take place on top of existing DeFi markets makes sense to me. Whether they are able to execute and scale a historically niche market is yet to be seen.

What I’m watching:

• Launch announcement (will we see something soon with MegaETH mainnet coming Feb 9th?)

• Partnership announcements (across the target audiences)

• Whether fixed rates offered are competitive with Maple’s 5-8%

• Vault liquidity bootstrapping and early yields

The Execution Layer Behind LayerZero

-Charlie

This week I want to dig into something that’s been nagging at me since reading @Evan_ss6‘s excellent piece on ZRO and the “Boomerification” thesis.

Evan makes the case that interoperability is one of four verticals worth focusing on as crypto integrates with TradFi. These comprise of picks and shovels that can earn fees regardless of whether the assets being traded are memecoins or tokenised equities.

LayerZero’s market share dominance (82% of cross-chain dollar volume per this BidCast interview), clean tokenomics (buybacks, positive insider flows) and the mysterious February 10th announcement all make for a compelling case.

But something’s missing from the conversation. Why the product is good enough to earn its market share?

The answer, in part, is an execution layer most users have never heard of: Aori.

The Cross-Chain Execution Problem

To understand why Aori matters, you need to understand how cross-chain intents actually work and why most of them kind of suck.

The basic flow: you want to move ETH from Arbitrum to Base. Instead of waiting for a bridge to verify and settle (slow), a “solver” fronts you the liquidity instantly on Base. You get your tokens immediately and the solver gets repaid later after the transaction is confirmed.

The difficulty here arises in that “later.”

Most intent protocols batch-settle through oracles. Across, for example, uses UMA’s optimistic oracle which settles roughly every 90 minutes. That means when a solver fills your order, their capital is locked for up to 90 minutes before they get repaid.

Solvers price this lockup into their spreads to compensate for:

• Time value of locked capital (can’t use it for other trades)

• Volatility risk (price can move against them during settlement)

• Rebalancing costs (getting inventory back to where they need it)

For cross-chain swaps (not just bridges), it’s even worse. If you’re swapping ETH for some long-tail token on another chain and the solver is exposed to volatility for a meaningful period, they’ll charge you a large spread to compensate.

As Aori’s founder Josh put it in a recent interview: “For a solver, effectively a market maker, everything about their pricing is about how fast they can hedge and how fast they can get their inventory back to where they want it to be.”

The longer a solver’s capital is locked, the wider they quote!

How Aori Solved It

Aori takes an entirely different approach by using LayerZero for settlement instead of batched oracles.

Here’s how it works:

1. Off-chain matching: Orders are matched in sub-millisecond latency (not waiting for block confirmations)

2. Instant fills: User deposits on source chain → solver fills on destination chain in the same block (~100ms after source block finalises)

3. LayerZero settlement: Instead of waiting for an oracle to batch-verify, solvers settle via LayerZero messaging whenever they want

4. Batching efficiency: Solvers can pack up to 350 transactions into a single settle call, amortising the cost to almost nothing per trade

The result: sub-basis point pricing on bridges (vs 10-50 bps elsewhere) and 4-25 bps on cross-chain swaps (vs 50+ bps for long-tail pairs elsewhere).

From their testing, the full end-to-end flow via LayerZero; deposit, fill, settle takes about 5 seconds. For context, the “optimistic” case for traditional intents is 3 minutes.

The key insight here is that speed isn’t just a UX feature but rather an economics feature. The reason your cross-chain swap “feels instant” is actually a byproduct of making the economics work for solvers. Faster settlement → lower capital lockup → tighter quotes → better prices for users.

Where You’ve Already Used It

If you’ve used Stargate infrastructure recently, you’ve probably touched Aori without knowing it...

Stargate Fast Swaps: Every cross-chain swap on Stargate’s frontend is powered by Aori. This is LayerZero’s flagship consumer product—the thing Brian (LayerZero CEO) is targeting $100M annualised revenue from within 12 months.

USDT0: Aori powers transfers on USDT0’s site, the cross chain extension of USDT that moved $63B in its first year.

Plasma Launch: $8B+ deposits in a matter of days post-mainnet Launch with all Stargate→Plasma volume flowing through Aori. The largest transfer processed by LayerZero infrastructure was a whopping $800m transfer for merely $0.81 in cost with the largest Aori-facilitated transfer around $150m in value for equally negligible cost.

As day-one partners for the Monad and Stable launches, the pattern is clear: every major chain launch using LayerZero as canonical infrastructure inherits Aori as the execution layer.

According to aoriscan, they’ve done ~$500m volume natively since going live in mid-September across 8 chains and 92 tokens:

The Investment Angle

So how does this connect to the ZRO thesis Evan laid out?

For ZRO holders/watchers:

Aori’s success flows directly to ZRO through Stargate. Brian has publicly stated a $100M annualised revenue target for Stargate within 12 months with 100% of excess revenue going to ZRO buybacks.

Stargate currently does ~$25M annualised. The path to $100M runs through three revenue streams:

1. Same-asset bridging (0.5-6 bps)

2. Cross-asset swaps via Aori (4-25 bps)

3. An undisclosed AUM-driven product line

The cross-asset swap business is the higher-margin play, and that’s Aori. The better Aori executes → the more volume Stargate captures → the more buyback pressure on ZRO.

For those wanting direct Aori exposure:

The team comes from market-making backgrounds (trading desks, MEV, options) and they’re building execution infrastructure. There is no talk of a token nor rush to TGE, this is either:

• A future catalyst to watch

• They stay as B2B infrastructure within LayerZero’s ecosystem indefinitely (and by extension are a likely acquisition target for LayerZero)

Either way, understanding execution under the hood helps you evaluate whether LayerZero’s dominance is defensible.

The broader thesis:

This connects to the “Boomerification” framework. LayerZero already has Dinari and Ondo using OFTs. The February 10th announcement Brian has teased is apparently TradFi-adjacent and is something they’ve been building for 2.5 years.

If it’s tokenised securities infrastructure, Aori strengthens its position in cross chain execution for traditional finance moving onto crypto rails. The execution layer doesn’t care if it’s settling a cross-chain swap of ETH, USDT0 or tokenised AAPL shares.

My Take

I’m genuinely impressed by the technical approach here. Aori rethought the fundamental constraint (settlement latency) and built around it.

They are not trying to compete with Stargate or push their own frontend. Chugging away as the engine underneath is likely a much more defensible position than trying to win user mindshare in a crowded bridging market.

What I’m watching:

• LayerZero and Stargate market share

• Aori share of best quote (currently ~60%, targeting 95%)

• Token announcements

Whether they ever launch a token or stay as infrastructure within LayerZero’s stack, understanding what’s making the product work is valuable context for anyone with a view on ZRO or interoperability more broadly.

Thanks to Alexander (@alexanderguy) for the context on Aori. You can explore their analytics at aoriscan.io and read their technical deep dives at aori.io/research.

The Parsec Agent

-Charlie

Say hello to the Parsec Agent, your crypto native sidekick:

The Parsec Agent draws on a multitude of curated data sources, both internal and external, to provide dynamic and accurate insights:

• Onchain and market data

• Curated X content/sentiment analysis

• Project documentation and socials

• News

• & More

Leveraging Parsec’s existing onchain analytics suite, the agent provides visualizations to provide deeper context and insights when prompted:

Prompt the agent with regular, deep and scheduled prompts depending on your use case.

Deep Research is best suited for complex prompts requiring comprehensive analysis—the agent conducts multi-source investigations and synthesizes findings into actionable insights.

Scheduled prompts automate your research workflow with recurring prompts delivered on your schedule.

We’ve found scheduled prompts to work particularly well to stay on top of developments across ecosystems:

Early users have found particular value using the agent for:

• Discovering new projects, narratives and coins

• Understanding complex technical docs

• DeFi optimization

• Catching up with CT without hours scrolling

• Structuring trade ideas and theses

Give it a try today completely for free at parsec.fi

$AZTEC Public Sale

-Charlie

Ignited by the PUMP ICO in July, we are well and truly amid an ICO season as high profile teams seem to be favouring this form of “community alignment” (fundraising from the public at what the team believe to be reasonable terms) as opposed to the airdrops of old.

Since Pump.fun we’ve seen the Plasma, MegaETH and Monad ICOs, there are rumours that Fogo are eyeing a December ICO pre TGE in January too...

Today I want to talk about a team who have been grinding on privacy for 7 years - through the Tornado Cash arrests, through the “privacy is dead” narratives, through multiple product iterations. Aztec’s full vision is just around the corner and unsurprisingly, they are taking a rather unique approach to their public sale.

Today’s piece will follow this structure:

• Aztec

• Ignition

• Use Cases

• $AZTEC

• Public Sale

• Takeaways

Aztec

With a tenure in the crypto<>privacy space since 2018, Aztec just recently announced the Aztec Ignition chain. Now before you retort that building a chain doesn’t take 7 years, let me explain to you why Aztec’s case is a little different.

Aztec didn’t start with a chain but with a problem. In 2017, the team tried to tokenise corporate debt on Ethereum and immediately hit a wall: no institution would put their portfolios and trade data on a transparent ledger. That limitation sparked a 7 year obsession with privacy.

This isn’t their first rodeo. They shipped zk.money in 2021 (Zcash-style privacy for payments) and Aztec Connect in 2022 (private DeFi on L1) - both successful products that reached tens of thousands of monthly active users. But they sunsetted both to build something more ambitious: a fully programmable private L2 where privacy isn’t bolted on but baked into the base layer.

The Aztec team concluded that an entirely new virtual machine and programming language was required in order to realise their vision; becoming the “private world computer”.

So that’s what they did; build a more complex, programmably private Layer 2 network powered by Aztec’s Noir language.

Noir is a Rust-based programming language specifically for ZK circuits. It’s now being used beyond Aztec with teams like Payy (another super interesting project, I strongly recommend @jon_charb’s piece on it) using Noir (Just write “if”: Why Payy left Halo2 for Noir).

Think of Aztec as SSL for blockchains. The internet faced the same bottleneck in the 1990s - it couldn’t go mainstream without privacy and security guarantees. SSL/HTTPS unlocked trillions in value. Blockchains are at that exact inflection point now, and programmable privacy is the unlock.

Unlike privacy coins like Zcash (single asset, one use case) or privacy tools like Railgun (application-layer bolt-ons), Aztec is the birth of fully decentralised, programmable, composable state.

The key innovation here is private composability. On a transparent chain, if you want to add privacy, every app needs its own custom ZK toolchain. These don’t play nicely together and complexity explodes. Aztec solved this by building privacy into the base layer, so any smart contract deployed there gets privacy guarantees out of the box. Think: private ERC-20s, private DEXs, private identity primitives - all composing atomically within single blocks.

Aztec uses client-side proving for private transactions - computation happens on your device, you generate a ZK proof locally, then send just the proof to sequencers. This is fascinating from a scaling perspective because sequencers only need to verify proofs for private transactions, not execute them.

Ignition

Ignition is a bit like Aztec’s version of the Beacon chain in that it doesn’t allow for transactions. Once the final checks and audits are in place, Ignition will be upgraded to enable Aztec’s private execution environment which is expected in Q1 2026.

Ignition launched as a Stage 2 Layer 2 with 668 sequencers (stake.aztec.network) distributed globally from day one. This is almost unheard of with most L2s running centralised sequencers printing $40-150M annually with a promise to “decentralise later”.

Aztec had no choice. When your value prop is privacy, a centralised sequencer could face massive government pressure for backdoors. As the founders put it: “There will be a backdoor in Aztec over our dead bodies.”

To cement credible decentralisation, the Aztec Foundation, Labs employees, and investors are banned from participating in staking and governance for one year post-launch. Control rests entirely with the community, not any single entity.

Use Cases

A very clear near term use case is private DeFi for crypto natives. Particularly the bridging angle as users can send transactions through Aztec to public chains thereby hiding the origin address and maintaining wallet hygiene and segregation. For many on chain users this would completely remove the need to use centralised exchanges as bridges.

In this respect one can view Aztec as “positive sum” to other L1s/L2s because they’re not necessarily competing for the same transactions; transactions can be routed through Aztec to add a degree of privacy to public chains. This creates a scenario where Aztec can generate fees without needing to siphon liquidity from other chains. A simple view on Aztec’s utility; what percentage of existing DeFi users would like to route their transaction through Aztec to add privacy.

Another interesting userbase in the private DeFi category would be the emerging crypto neobanks (think EtherFi, Gnosis Pay, UR etc.). It seems quite clear that transparency is a bug not a feature here given that current crypto cards allow anyone to reconstruct your credit card statement (h/t @tzedonn). Interestingly, this is the major pain point Payy (which uses Noir as mentioned prior) was created to tackle.

In the medium-term, one could see identity primitives using ZK Passport take the fore (for applications which currently have to take the unfortunate sacrifice of storing user data in order to remain compliant).

There are funky example use cases in sectors like gaming too. For example, breeding mechanics don’t work on transparent chains as they get reverse-engineered instantly, you need ZK for information asymmetry in games like Cryptokitties.

Under-collateralised lending becomes more likely too. Your bank statement is cryptographically signed and fed into a ZK proof to prove “my salary is above X, outgoings below Y” without revealing the actual data. Suddenly, real consumer finance can begin to exist on-chain.

In the long- term, the Aztec core vision is to serve as “the global exchange and settlements platform for the world’s assets. Not digital assets, but all assets”.

There are 16 teams currently building on testnet (and launching on mainnet when transactions go live). I would expect that number to grow materially as transactions go live early 2026 and developers realise what’s suddenly possible)

So how does the token actually work?

$AZTEC

The $AZTEC token has the following token distribution (according to their economic whitepaper):

The token has three functions:

• Staking: 200K $AZTEC minimum to run a sequencer (currently 668 sequencers live as per stake.aztec.network)

• Fees: Transaction fees paid in $AZTEC (users can pay in ETH/USDC via Fee-Paying Contracts that hold $AZTEC reserves)

• Governance: Control over network parameters, upgrades, and token issuance

The fee mechanism mirrors Ethereum’s EIP-1559. Aztec uses “Mana” (their version of gas) with a target per block. When usage exceeds the target, fees burn automatically - creating deflationary pressure as the network scales.

At modest scale (10 TPS, ~$0.15/transaction), the protocol generates ~$22M annual revenue. Transaction costs run 2-5 cents for basic payments mostly consisting of data costs to Ethereum. Year 1 inflation is 2.41% but there is scope for that to go deflationary as usage grows over time.

Note that the staked $AZTEC shown on stake.aztec.network represents tokens sold during the Genesis Sequencer Sale. These tokens were sold at a discount to the floor FDV for the Public Sale with a clearing FDV around $250m, these tokens are locked for 12 months.

The interesting bit is that client-side proving means users generate proofs locally and sequencers just verify them. Sequencers don’t execute private transactions, only verify proofs. This is why the cost structure scales differently than transparent chains.

For context, during Ignition (no user transactions yet, just empty blocks), the network costs are ~$2.4M annually to run. The pre-minted rewards cover this and incentivise sequencers before transactions go live and fees begin to flow.

All public sale tokens have a 12-month lockup that can be shortened to 90 days via governance vote. Given that initial holders either bought or earned tokens via staking, governance will likely vote for the minimum 90-day period so we can assume the $AZTEC TGE will take place February 17, 2026. Team and investor tokens have 12-month cliff + 24-month linear unlock, and they can’t stake or vote for the first year:

At TGE, circulating supply will consist of:

a) Public Sale tokens (14.95%)

b) A portion of pre-minted Y1 Network Rewards (0.6%) (representing sequencer rewards in the Nov 17th → Feb 17th period)

For a cumulative ~15% at TGE.

Public Sale

The cynic points to the Aztec Public Sale and ignorantly ascribes the recent “privacy meta” as the reason to raise now.

The Aztec enjoyooor (me), thanks to conversing with the team, knows that the Aztec Public Sale has been in the works for just over a year. You have to remember, Aztec has been in development for 7 years, they are hardly jumping on the privacy bandwagon here. It’s equally foolish to assume Uniswap would rush the first public sale utilising their Continuous Clearing Auction (CCA) mechanism just to capture a potentially ephemeral narrative/meta in the secondary markets.

The Aztec team believe that airdrops are great for creating initial hype but not so optimal for the formation of long-term aligned community holders, they cite that communities tend to churn within 3-6 months post-airdrop. They argue that for PoS networks, participants need skin in the game which is not compatible with airdropping a large % of supply.

The sale is open to US retail, uses ZK Passport for compliance and has had a long pre-registration period to avoid information asymmetries (public bidding starts Dec 2nd).

As per the tokenomics above, 14.95% of $AZTEC supply will be sold at a minimum valuation of ETH 98,493 ETH (sale is denominated in ETH).

Given an ETH price of ~$3k at the time of writing, we get to minimum FDV of just below ~$300m. Naturally, the exact FDV will depend on the price of ETH at the time orders are filled. Note that Aztec’s most recent equity round put the token valuation at a ~$1.4b valuation.

As I briefly mentioned, the Public Sale utilises Uniswap’s new Continuous Clearing Auction (CCA) mechanism which Aztec worked closely with Uniswap Labs on as core contributors.

Part of the CCA mechanism involves seeding a Uniswap V4 pool at the final auction price but the more interesting part is how the auction itself functions:

The seller chooses 3 variables:

1. auction length

2. supply size and schedule (number of tokens sold at each interval)

3. minimum sale price

And the buyer simply has to (h/t @jaosef):

a) set a budget and maximum price (or market price)

b) sit back and relax; no gas wars, no timing games

c) each period the contract auto-allocates your tokens if your bid ≥ the clearing price

d) at auction completion, you claim your tokens and wait for TGE

The CCA clears every block so supply and demand create a price in real time on a continuous basis rather than discretely at the end of the sale. The aforementioned floor FDV/price remains the clearing price until cumulative demand exceeds the tokens available.

While supply remains constant, if demand increases over time so does the clearing price such that participants who bid earlier are more likely to receive allocations at lower prices.

Their use of CCA certainly seems to be in keeping with what seem to be Aztec core values in that it is entirely onchain, promotes transparency and fair price discovery.

If it’s still not clear, @FreddieFarmer wrote a really good primer on how the CCA works in practice which I can highly recommend: https://x.com/FreddieFarmer/status/1994469052840886613

The sale timeline is as follows:

Wallets in the Contributor Whitelist (consisting of early contributors and long-standing community members, including genesis sequencers, OG Aztec Connect users, network operators, and community members) can start bidding today, ahead of the public auction. They will enjoy one day of clearing before the public track bids begin to clear on Dec 2. This is a form of airdrop if we assume high demand from the public track given that contributors benefit from this early advantage in the form of representing only a subset of total demand.

So far, we are seeing bidding activity really picking up from this lucky and aligned early cohort:

link to parsec dashboard (h/t @geeogi)

Takeaways

All jokes aside, while I was not around in the truly early days of Crypto, from my perspective Aztec are very much channeling the Cypherpunk energy which I have only been fortunate enough to read about.

There are so many use cases which fall flat on their face without some degree of privacy which is exactly why Aztec came to be in the first place.

This is why viewing Aztec through the lens of “just another L2” is missing the forest for the trees in my opinion given the promise for an entirely different class of applications that cannot exist on transparent chains.

Whether their differentiated blockspace has product market fit is yet to be seen and it certainly won’t be an easy road but from where I’m sitting, Aztec seem to be taking all the right steps towards their vision of becoming the “private world computer”.

I’d recommend watching this one very closely. What is my strategy you ask? Market order & chill (h/t @FreddieFarmer).

Daylight

-Charlie

the following does not constitute investment or financial advice

This week I’ve been digging into Daylight Energy and honestly, this might be one of the few DePIN projects where the DeFi integration actually makes sense. Let me explain why.

The Problem (Very Real)

The US power grid has structural issues:

• Some markets see 20%+ electricity price increases year-over-year

• Massachusetts households pay 35 cents per kWh

• The average American household has 8 hours of blackouts per year, double what it was a decade ago

Meanwhile demand is exploding; AI compute needs power, EVs need power, industrials reshoring from China need power, etc.

Naively one would assume utilities can just “build more power plants and transmission lines” but in reality, building this out takes forever, costs billions and no one wants it in their backyard...

On the other hand, distributed energy (home solar + batteries) can deploy in weeks. The problem is then how to actually finance this at scale?

Traditional residential solar has insane customer acquisition costs with the average US homeowner paying $4 per watt for installation while in Australia they pay $1 per watt. The difference here is purely sales and marketing bloat.

This is where Daylight comes in.

The Actual Model

Daylight offers subscription solar. No upfront cost, you pay them monthly for power (less than your utility) and you get backup batteries for when the grid goes down.

But the interesting part isn’t the subscription model, that’s been tried before, it’s how they’re using crypto to solve two massive problems in residential solar:

Customer Acquisition: Instead of door-knocking salespeople, they’re using token incentives (”Sun Points”) to turn it into a pull market. Reward people for signing up, reward them for bringing others and create a flywheel.

Financing: They just raised $75M ($15M of equity and $60M in project financing) but the real play is DayFi; a DeFi protocol that tokenises the electricity revenues from their solar installations.

The USDai Parallel

From my DeFi pattern matching perspective, they’re basically building USDai but for energy infrastructure instead of GPUs.

USDai is tokenizing GPU compute loans to fund AI infrastructure. Their sUSDai token is a yield-bearing token backed by GPU revenues.

Daylight plans to operate similar a structure for distributed energy:

• USDai: GPU loans → sUSDai yield token

• Daylight: Solar installations → sGRID yield token

This week they announced a partnership with Plasma to launch:

• GRID - M0-issued stablecoin

• sGRID - Yield-bearing token backed by electricity revenues

As far as I know, this is the first time you can get on-chain exposure to electricity yields through a tradable token.

Why This Structure Might Actually Work

Most DePIN projects have this vague “maybe fees accrue to token somehow?” value prop. Daylight’s is straightforward: homeowners pay monthly subscriptions for power. Daylight aggregates distributed energy infrastructure and sells stored energy back to grid during peak demand. The protocol generates actual cash flows from electricity, sGRID tokenises those cash flows.

The revenue model is simple:

1. Monthly subscription payments (predictable, recurring)

2. VPP dispatch revenue (sell stored energy back to grid at a premium during peak hours)

Together these create a flywheel. More subscribers → bigger battery network → more VPP revenue → lower subscription costs → more subscribers.

They’re currently active in Illinois and Massachusetts which makes sense given these are high-price markets where the value prop is obvious.

From a DeFi perspective, this would be genuinely compelling if sGRID can offer:

• Double-digit yields backed by electricity revenues

• Transparent on-chain verification

• Liquid token tradable on DEXs

• Composable (Pendle markets, usable as collateral on Aave, etc.)

My Take

For me personally, I’m interested and will look to participate in GRID/sGRID whenever they launch.

The USDai comparison gives me some confidence the structure can work. Daylight has the same setup but for energy, which is arguably a larger and more fundamental market.

What I’m watching:

• GRID/sGRID launch date announcement

• What are the actual tokenomics when they’re revealed?

• Growth in current markets (Illinois and Massachusetts)

Whether it works obviously remains to be seen. But at least the blueprint makes sense in a way that many DePIN projects do not!

Stream Finance

-Charlie

This week, the DeFi ecosystem suffered one of the most significant blow ups we’ve seen in quite some time. Stream Finance — a yield-focused protocol — disclosed a $93 million loss attributed to an external fund manager, triggering a cascade of liquidations, depegs, and frozen withdrawals across the entire ecosystem.

What makes this particularly interesting (and concerning) is not so much the size of the initial loss, but rather the intricate web of interconnected positions that amplified a $93m problem into what analysts now estimate to be at least $285 million in total exposure across lending markets.

The incident has reignited debate about the “risk curator” model that’s become increasingly prevalent in DeFi — a model where third-party vault managers promise double-digit yields on stablecoins through complex leverage loops, while taking management fees and claiming to “curate risk.”

As the dust begins to settle, let’s dissect what happened, who got caught in the blast radius, and what this tells us about the current state of DeFi infrastructure.

What Happened?

On November 3rd, users began noticing that Stream Finance’s stablecoin xUSD was depegging. Despite no official communication from the team, the token started sliding.

Hours later, the team finally acknowledged the situation in an X post:

The announcement confirmed what many had already suspected: Stream had frozen all deposits and withdrawals, trapping an estimated $160 million in user funds. The protocol engaged law firm Perkins Coie LLP to lead a comprehensive investigation into the incident.

Stream Finance operated what’s known as a “recursive looping” yield strategy.

1. Deposit collateral (USDC, ETH, BTC)

2. Mint Stream’s synthetic assets (xUSD, xBTC, xETH)

3. Use those synthetic assets as collateral in lending markets

4. Borrow more assets, then repeat

This allowed Stream to offer a headline yield of 18% on stablecoins — far above what traditional DeFi protocols like Spark could provide.

The problem? Stream’s xUSD was used as collateral across multiple lending platforms (Euler, Morpho, Silo, Gearbox), creating a house of cards.

When confidence evaporated and xUSD began depegging, it triggered liquidations across these platforms. But because many of these markets were “custom” or “isolated” pools with low liquidity and fixed-price oracles, the normal liquidation mechanisms failed to function properly.

As Omer Goldberg from Chaos Labs pointed out, Stream’s reliance on “fundamental value” oracles instead of real-time market pricing meant that liquidation triggers were delayed. While such systems can prevent unfair liquidations during normal volatility, they convert “price discovery into trust discovery” when confidence falters.

The result? Markets went to 100% utilization, meaning lenders couldn’t withdraw their funds even if they wanted to.

The Contagion

DeFi research group Yields and More (YAM) mapped out the exposure, revealing a staggering $284.9 million in interconnected debt positions:

Largest victims:

• TelosC: $123.6 million

• MEV Capital: ~$34 million across four lending markets

• Re7 Labs: $27.4 million ($14.65m in isolated xUSD/USDT0 pool + $12.75m in bad sdeUSD/deUSD collateral)

• Elixir’s deUSD: Backed 65% by Stream positions ($68m loan to Stream)

• Treeve’s scUSD: Exposure through lending market positions

Perhaps the most dramatic secondary effect was on Elixir Protocol’s deUSD synthetic dollar.

Elixir had lent $68 million to Stream Finance — representing roughly 65% of deUSD’s backing collateral. When Stream halted withdrawals, deUSD holders panicked, despite Elixir claiming to have “full redemption rights at $1” as Stream’s only creditor with such privileges.

The reality proved more complex. As Elixir later admitted:

“Stream holds roughly 90% of the deUSD supply (~$75m), while Elixir holds a similar proportion of its remaining backing as a Morpho loan to Stream.”

Translation: Stream Finance controlled most of the deUSD supply and owed Elixir the backing collateral. A perfect circular dependency.

By November 6th, Elixir announced it would sunset deUSD entirely. The team processed redemptions for 80% of holders and took a snapshot of remaining balances for 1:1 USDC redemption through a forthcoming claims portal.

Winners, Losers, and the Risk Curator Debate

The “risk curators” who had embraced Stream’s high yields found themselves on the wrong side of history:

• MEV Capital: $34 million exposure across four permissionless lending markets

• Re7 Labs: $27.4 million total exposure

• TelosC: The biggest loser with $123.6 million

Perhaps most damningly, Re7 Labs faced accusations of holding funds hostage in illiquid markets, unable to process withdrawals even after the depeg.

The incident also exposed questionable risk practices. Re7 Labs later admitted they had identified “centralized counterparty risk” in xUSD through due diligence but chose to list it anyway due to “significant user and network demand.”

In other words: they saw the red flags, but the yield was too attractive to pass up.

Not everyone was caught in the blast radius. Conservative risk managers who had avoided xUSD entirely emerged vindicated:

• Steakhouse Finance: Zero xUSD exposure

• K3 Capital: Not only avoided xUSD, but actually made $250,000 during the crisis managing Neutrl Labs’ $75 million deposit

• Gauntlet: No xUSD exposure, though they still had to pause certain markets due to Elixir contagion

As one DeFi participant put it:

“There are actually some curators that never touched xUSD. Shout-out to them as they actually seem to CURATE RISKS. A very unknown concept for many curators.”

The Stream collapse has forced the DeFi community to confront some uncomfortable truths about the “risk curator” model:

1. Incentive misalignment: Curators earn management fees on deposits regardless of risk. Higher yields attract more deposits, so there’s natural pressure to take on more risk.

2. Opacity masquerading as transparency: While positions are technically “on-chain,” the complex web of recursive loops, cross-chain exposure, and synthetic collateral makes true risk assessment nearly impossible for retail users.

3. Fixed-price oracle risk: Using “fundamental value” oracles instead of market pricing creates a false sense of security. As one analyst noted: “Hardcoded oracles mean in an impairment, you’re potentially riding it to zero.”

4. The race to the bottom: When users can get 18% on stablecoins from one curator and 12% from another, capital flows to the highest yield. This creates competitive pressure to take on more leverage, tighter collateral, and riskier positions.

Lessons and Implications

Having observed this unfold in real-time, a few things are clear:

1. Yield isn’t free

The 18% yields weren’t magic — they were compensation for taking on catastrophic tail risk. When protocols like Spark offer 5% and Stream offers 18%, the 13% difference is the premium for insolvency risk, leverage risk, and contagion risk.

2. Oracle design matters enormously

The use of “fundamental value” oracles rather than real-time market pricing meant that positions didn’t liquidate when they should have, allowing bad debt to accumulate. This is a critical design choice that needs more attention.

3. The circular dependency problem

Stream borrowing against its own xUSD to create leverage, which was then used as collateral elsewhere, created a perfect storm. When confidence broke, there was no way to unwind positions in an orderly fashion.

Closing Thoughts

For those of us who participate in DeFi, this is a sobering reminder that higher yields GENERALLY always come with higher risks. The protocols that survived this crisis unscathed weren’t the ones with the cleverest leverage loops or the highest APYs — they were the ones with boring, conservative risk management.

MegaETH

-Charlie

This week the MegaETH ICO saw:

• $1.39bn committed

• $49.95m raised

• $999m valuation

• 27.8x oversubscribed

• 50k+ total bidders

This has certainly been the highest profile ICO I have personally witnessed (I wasn’t around for the 2017 ICO era). It certainly eclipses the PUMP ICO earlier this year from my perspective.

Empirically there seems to be a clear divergence in the profile of participants in this sale:

• Profile A simply views $1bn as underpriced and a free money opportunity, they opt for no lock up and likely are a seller in the pre market (violating Sonar T&Cs) or at TGE.

We’ve even seen what looks like sybil activity (likely from the Profile A types) to get a greater than max allocation using multiple identities:

• Profile B is in it for the tech, likely opts for the lock up and may be adding to pre-existing Fluffle and/or Echo exposure to MegaETH.

Both are logical for different reasons, Profile A is looking at the pre market and thinking this could be a profitable and low risk quick flip while Profile B is looking into the tokenomics and thinking this might be the first L2 token which actually has a valid raison d’être, taking a longer term view.

And it’s one of those spots where both profiles do have a strong and valid point of view (of course other than hedging against T&Cs which I cannot advocate for).

Naturally from the team’s perspective, Profile B is significantly more “aligned” and represents the kind of holder base you want to foster and encourage. Equally, if you offer 100% of the allocation to Profile B you suffer from no float at TGE and this huge looming cliff unlock event 1 year post TGE.

We likely see allocations go to both lockers and non-lockers in order to balance this dynamic but given the severe over subscription, it’s hard to believe anyone, even the most aligned, get anything close to the max allocation.

MegaETH certainly seem to be one of those teams who are doing all the right things: community first ethos, vibrant native ecosystem and a refreshing approach to KPI-based token vesting (something I have wanted to see forever).

For those who participated in the ICO, let’s see what November 5th has in store, I wish each and every reader luck in their allocation.

For those who didn’t participate, in my opinion this is a fantastic example of what projects should strive for in terms of fundraise design, community involvement and marketing.

Here at parsec.fi, we are looking forward to exploring ways we can strengthen our ties with and contribute to MegaETH and the ecosystem.

A market hungry for yield

-Charlie

the following does not constitute investment or financial advice

One of the most salient characteristics of the on chain environment at present is an insatiable demand for yield, particularly on stables. As the market has matured, the market for on chain yields has become increasingly efficient too.

We are seeing projects become smarter about how they incentivise liquidity/TVL, often opting to bootstrap with capped deposit programs in the early days which often pay a higher retroactive APR. Naturally the market is reacting to this with extremely competitive windows on chain surrounding these deposit program openings.

For example, yesterday saw a project called Stable (Stablecoin chain, Plasma competitor, backed by Bitfinex and USDT0) open up an $825m pre-deposit campaign:

As it happens close to $600m of this cap was already filled prior to the public announcement. Anecdotally, users who saw the announcement and wanted to allocate had about 45 seconds until the entire cap filled.

I am not speculating on whether there is insider activity at play here (I’ll let you be the judge) but rather highlighting the idea that demand for “safe” stablecoin yield is extremely high and the level of competition to get that yield is increasing exponentially.

We have seen similar instances with USDai caps filling almost instantaneously when raised:

It does beg the question as to why? What is driving this increase in demand and therefore efficiency? I think there are probably a few reasons:

a) increased perceived difficulty presented by other sources of returns in the market

b) decreased friction for sharp trading firms/family offices to onboard capital

c) decline in historically dependable and high sources of yield (likely driven by both a and b) as we see funding rates/on chain interest rates exhibit a lower baseline and liquid incentive/reward opportunities quickly normalising to a risk-adjusted baseline as capital rapidly moves to capture outsized yield

This is likely why we see projects like Sky begin to shift backing to include higher yielding corporate credit instruments:

It’s also why I personally find projects like USDai and Daylight interesting, as their promise is to bring exogenous yield on chain in a scalable manner.

A practical guide on positioning for Polymarket

-Charlie

the following does not constitute investment or financial advice

Introduction

In both crypto and traditional markets, the most asymmetric opportunities often occur before assets go public (in private rounds).

For the most part, best-in-class projects tend to TGE at high valuations, limiting opportunities for secondary participants. This holds true in nearly every case except for the occasional Echo deal that breaks the pattern.

While some dismiss airdrop farming or pre-TGE participation as mercenary, I see it as a form of early-stage “seed investing” in protocols I believe in. Historically, engaging with pre-TGE protocols has been an excellent use of both time and capital — especially when you have high conviction in the team or product.

It’s not truly “early” to use Polymarket anymore, but prediction markets are seeing a clear resurgence in interest and usage. Today I want to outline some practical ways the average crypto market participant can position for $POLY.

The Case for $POLY

Let’s first look at the signs pointing toward a likely $POLY token and potential airdrop in the near future.

Exhibit A: Polymarket CEO retweets a tweet that mentions token launch

Exhibit B: Interesting reply guy activity

Exhibit C: Recent filing telling a similar story

Exhibit D: don’t know whether it gets much clearer than this...

It hardly takes a detective to see what’s unfolding here: the $POLY token is coming, and an airdrop to users seems likely. With Polymarket’s most recent fundraise at a $9 billion post-money valuation, it’s safe to assume it’ll be a substantial one.

Strategy

Despite the obvious setup, on-chain data suggests that high-volume, profitable, and active Polymarket wallets remain relatively few (h/t @DidiTrading).

Of 1.35 million total traders:

• $1k+ PnL: top 0.51%

• $50k+ volume: top 1.74%

• 50+ trades: top 22.87%

In other words, the threshold for being considered an “active” or “high-value” user is surprisingly low. Perhaps making this a particularly attractive opportunity for smaller portfolios.

My framework for opportunities like this is simple:

Make economically rational decisions in a vacuum — decisions that make sense even if the airdrop never happens — but which offer large asymmetric upside if my thesis is correct.

Or more plainly: don’t burn money chasing an airdrop that might never come, or one you might not qualify for.

While I can predict that a Polymarket airdrop is highly likely, it’s not guaranteed. From a risk-management perspective, I aim to ensure that my activity is +EV regardless of the outcome.

Guide

So what am I actually doing?

A. Conservative “yield farming”

Betting on high-probability outcomes with a “nothing ever happens” mindset

In these cases, betting No on obvious outcomes has yielded roughly 3–5% returns. On an annualised basis, a portfolio of such bets looks strong — but the return profile is concave:

The payoff resembles a short put: small, steady gains while risking absolute loss (limited to position size). The phrase “picking up pennies in front of a steamroller” comes to mind.

B. Directional edge in crypto-native markets

While there are genuine sharps in political or macro markets, that’s beyond the reach of most participants. I find the crypto-focused markets far more interesting.

For example, bidding Yes below 60c on Lighter market cap (FDV) >$2B one day after launch? looked compelling — and yielded over 50% in a few days.

These setups are usually size-constrained, but they’re great opportunities to engage with Polymarket while earning respectable returns on capital. Success here requires staying in tune with valuations, TGE timelines, and general market sentiment — but it’s arguably easier than guessing who the next president of Venezuela will be.

C. Cross-platform arbitrage

Some traders take opposing positions across platforms (e.g., long Zohran on Kalshi @ 86%, short Zohran on Polymarket @ 87.5%) to generate volume and open interest while maintaining neutrality.

In theory, that sounds great. In practice, it likely burns some capital through fees and spreads unless the mispricing is significant. I haven’t personally tested this, but if anyone has, feel free to ping me on X — I’d love to hear how it’s working in reality.

Closing Thoughts

Whether or not $POLY ultimately launches, actively using Polymarket remains one of the most interesting ways to participate in crypto’s evolving prediction-market ecosystem. My approach is simple: engage with the platform in ways that are self-sustaining, rational, and fun — while leaving the door open for asymmetric upside if and when the token drops.

In the meantime, I’ll continue using Parsec’s Polymarket analytics suite to monitor odds, holders, and market activity: https://parsec.fi/polymarket

Plasma Mainnet

-Charlie

Yesterday saw Plasma mainnet go live along with their native token XPL and what a launch it has been so far...

• ~$3bn in TVL in ~24 hours

• Most insane S tier list of Day 1 integrations, almost every DeFi bluechip you could think of deployed with mainnet launch

• XPL incentives in abundance

Before we get into the state of mainnet and opportunities on offer, let’s back track a little.

Plasma is an L1 focused on stablecoins: payments, yield, DeFi... you name it, if it involves stablecoins they want it on the chain. Backed by Tether, their north star is to make USDT transfers free and instant. Technically, the L1 is an EVM-compatible Bitcoin sidechain running “PlasmaBFT” consensus for rapid finality. It features a native BTC bridge, whitelisted gas tokens, and zero-fee USDT transfers.

Plasma raised:

• $24M from VCs

• $50M from the community via Echo

They’ve further strengthened this community first focus with actions taken during the token launch process with rewards for discord role holders, community contributors etc;

This also included rewarding minimum depositors into their June predeposit with a completely unexpected airdrop upon TGE:

As a highly anticipated project, unsurprisingly, many exchanges listed XPL pre launch perps prior to the spot token launch yesterday. This led to a great deal of contention around how the token would launch into mainnet and where it would find fair value. Safe to say, the pre launch markets were under selling by a quite a significant margin as XPL traded ~+60% yesterday:

Naturally, this price action along with an aggressive post mainnet XPL rewards program has created a bunch of attractive opportunities as summarised by this Merkl dashboard:

APRs were absolutely insane for the first few hours so naturally a significant amount of capital has bridged across to harvest the outsized yields. Despite that, yields remain elevated across the board, especially when we consider the amount of capital in some of these opportunities.

~20% on $2bn+ of capital is no easy feat!

For those who dig a little deeper there are even more attractive opportunities albeit with lower capacity, for example the “Borrow USDT0 from Fluid USDai-USDT0/USDT0 Vault” opportunity created an interesting looping opportunity.

With a 0.87 collateral factor, 0.9 liquidation threshold and collateral hard coded to 1, opportunities like this have offered extremely juicy opportunities for loopers. But as with any very attractive opportunity, capacity was gobbled up very quickly with caps filled at the time of writing:

Ethena have taken things into their own hands and look to be running a Plasma-focused incentive program of their own, subsidising USDe supply on Aave to the tune of ~10% (unlevered):

Plenty of opportunities for both the passive and the active in what has been the first new and exciting chain launch in quite some time. I look forward to monitoring and writing about future opportunities on Plasma but for now, go out there and enjoy some of those yields :)

All about that base

-Charlie

Just recently at “BaseCamp”, Aneri Amin (the Head of Product @ Base) tweeted this:

While this is not a formal announcement, Coinbase is hardly the kind of company to leak a rumour which hasn’t already been approved by legal. Having seen this, I believe a Base token is now a matter of when not if. This puts base in a growing category of teams who have been building for a long time but which the market was not certain would launch a token. In this respect, I’m putting Base in the basket with Polymarket and OpenSea here.

The small “Base basket” I track (consisting of the Aerodrome, Virtual and Brett tokens), has not maintained outperformance vs ETH on a significant basis since the announcement:

So if we assume the opportunity is not in existing Base coins, what can one do from an actionable perspective to expose oneself to a future Base network token? If we assume that some portion of the Base token might be distributed in the form of an airdrop, deploying funds into Base DeFi and organically using the chain might make a lot of sense.

One project I have been playing around with is Limitless, a Base native prediction market which is currently running a Season 1 of their points program.

With Polymarket secondary transactions rumoured to be transacting at a $10bn valuation, there’s a chance that up and coming prediction markets are ascribed some value in which case Limitless might offer a nice opportunity to play with something new on Base while becoming eligible for both a Limitless and Base airdrop simultaneously.

We’ll be keeping our eyes peeled for further information from the Base team regarding timelines, token utility and any potential airdrop information too - it’s time to lock in!

USDH is just a ticker

-Charlie

USDH is just a ticker, it really doesn’t matter as much as people think. Allow me to red pill you on what *actually* matters as we emerge from this little USDH saga.

In customary crypto twitter fashion, much of the discourse surrounding USDH this week has been salty to say the least. Abundant lamenting that there is cabal activity taking place; missing the woods for the trees in what has been a magnificent display of high calibre teams showing interest in building on Hyperliquid.

Yes it seems as though Native Markets might have had the jump on putting a proposal together, but is that really as catastrophic and distressing as some commentators would have you believe? I don’t think so. The price performance of HYPE during this period has been a testament to this.

A wide array of service providers, large and small, have come to the fore with extremely strong proposals and I do believe validators have considered all options in good faith.

While not the most liquid market, the Polymarket on this event hints that the outcome was not entirely “predetermined” as we saw odds vary significantly during the first half of the week:

The single most important thing to remember in all of this is that USDH is just a ticker... the best and most aligned teams will likely do everything they laid out in their proposals anyway and we will simply see a broader selection of high quality quote assets (Stablecoins) being deployed.

For example, we’ve already seen Ethena come out and say that no matter the outcome, they will continue with the plans laid out in their proposal, notably:

• hUSDe native synthetic dollars in partnership with Liminal + other USDe enabled savings and card spending products

• Bringing USDe hedging flows to Hyperliquid

• Ethena-enabled HIP-3 initiatives including:

  • reward-bearing trading collateral

  • modular prime broking

  • perpetual swaps on equities

Equally, I find it highly likely that Paxos (and by extension PayPal) will stay true to much of what they laid out in their proposal too:

• HYPE listing on PayPal/Venmo

• Free on/offramps for their quote stable

• Integrations for PayPal Checkout, Braintree, Venmo, Hyperwallet, Xoom, and Pay With Crypto

• Quote stable migration from USDC, PYUSD, USDG, and supports enterprise credit/debit cards.

This is all especially prevalent given the quote asset network upgrade which looks be coming soon, as announced by the Hyperliquid team yesterday:

The most notable pieces for me are the 200k HYPE stake requirement and liquidity requirements for spot quote assets.

While this specific network upgrade will only support alternative quote assets for spot markets, it is only a matter of time before we see alternative quote assets for perp markets. This will likely act as a big unlock for Unit by enabling non stable quote assets for perps.

I’m consistently impressed by Hyperliquid’s rapid, high-quality development and the exceptional calibre of teams building within its ecosystem. We’re proud to support the ecosystem with Purrsec and look forward to the very bright future ahead.

USD.AI

-Charlie

A couple of weeks ago, on the week of their public launch, I wrote about USD.AI. If you haven’t read that one, go back and read it here before reading this. Today I want to discuss some interesting developments which have transpired over the last 2-3 weeks while the initial article goes deeper into protocol itself and the market opportunity they are addressing.

Loan issuance ramping up

With such a large amount of TVL flowing into the protocol over the last couple of weeks, the majority of this capital is sitting in T-Bills for now. However, it seems as though the loan issuance machine is in motion with some larger loans coming online in due course with some healthy looking APYs inbound for stakers:

Given the apparently high demand to engage with USD.AI and participate in their “Allo” (dual ICO<>Airdrop) campaign, it is likely that we will continue to see TVL growth outpace growth in loan issuance meaning that risk should remain relatively low for sUSDai in the near term. At some point (perhaps if/when team stop raising caps) I would expect this to even out as TVL growth slows and loan issuance picks up, thereby increasing yield paid to sUSDai but also risk to stakers (increase in redemption times + increase in credit risk).

Pendle markets live for USDai and sUSDai

DeFi integrations were still TBD when I wrote the initial piece but we have since seen a raft of integrations including Pendle which has seen a lot of interest on both the USDai and sUSDai markets:

Interestingly, participating in these Pendle markets offers points multipliers for their “Allo” campaign with the multiplier depending on the interaction a user makes:

Buying the YT of either the USDai or sUSDai market offers the best dollar for dollar pure play on either the ICO aligned USDai track or the Airdrop aligned sUSDai track (with 25x Allo multipliers respectively). This dynamic is likely juicing the yield, which is sitting in excess of 20% for both markets:

Binance Labs

Following the public launch we’ve seen YZiLabs (prev. Binance Labs) make a “strategic investment”. While such an investment far from guarantees a Binance listing at TGE, naturally this narrative has taken hold and is likely changing the way some participants are thinking about the $300m FDV price point of the ICO/Airdrop. YZiLabs is joining quite the roster of angels who are already on the cap table here, including @goodalexander who has cemented himself as somewhat of a thought leader at the intersection of Crypto and AI.

Plasma Partnership

Just 2 days post the Binance Labs announcement we saw a partnership announcement between USD.AI and Plasma. If you are not familiar with Plasma, read this one I wrote 3 weeks ago for an overview. While the details of the partnership are not clear, it seems likely that USD.AI will deploy on Plasma and become a core ecosystem project over there. Yield on stables will surely be one of the core use cases which Plasma want to offer users making this partnership look very synergistic. If you expect significant activity and TVL growth on Plasma, this will be very accretive for USD.AI. Demand for anything Plasma farming has been absolutely insatiable so it is not surprising that the combination of Binance Labs and this partnership announcement led to a flurry of USDai mints (far more than at public launch itself) and the initial deposit cap of 110m being hit:

Note the significant uptick in mints/deposits around the 26-28th August right around these announcements.

Deposit Caps lifted and hit

This one is rather self explanatory based on the screenshots above. With the initial caps set at 110m, we saw those caps lifted to 160m and filled within a very short space of time indicating the increasing interest in USD.AI and particularly the ICO/Airdrop campaign they’re running at present. These caps will be raised again by a further 50m next week so set an alarm if you want to deposit!

prediction markets mania

-Charlie

Prediction markets seem to be commanding a bunch of attention at present, seemingly out of nowhere. This doesn’t necessarily seem to be driven by a large increase in volumes on the platforms themselves but rather a renewed sense of excitement among market participants perhaps driven by the outperformance of prediction market related onchain coins and some notable press releases. Let’s dive in...

Polymarket fundraise

While this is not public news, I have heard along the grapevine that Polymarket recently raised at a $1bn valuation with secondaries already trading much higher.

The more public news in this regard has been Donald Trump Jr. making an investment in Polymarket and joining the team as a strategic advisor - certainly not something I had on my bingo card for 2025!

Kalshi on Solana

Recently we’ve seen Kalshi hiring for a “head of crypto” and that role being filled by CT extraordinaire mr John Wang.

John Wang was previously at Ventuals (a Hyperliquid ecosystem project building perps for pre-IPO companies with HIP-3). Part of me was hoping that this might lead down some form of a Kalshi<>HIP-3 type path but perhaps this was just wishful thinking with Kalshi announcing yesterday their commitment to “expanding prediction markets on Solana” with native USDC and SOL deposits.

New age prediction markets

Naturally this excitement and buzz surrounding prediction markets has born some new market entrants. A couple I am watching are Limitless on Base and Outcome on HyperEVM. Limitless are a little further along, currently on mainnet and conducting their first points season. Outcome on the other hand are quite a bit earlier and currently engaging in a testnet trading competition.

This week also saw the following announcement:

FORMER POLYMARKET AND KALSHI TEAM RAISE $15M FROM UNION SQUARE AND COINBASE TO BUILD AN ONCHAIN, PERMISSIONLESS, AND REGULATED PREDICTION MARKET: BLOCKWORKS

Competition is intensifying. In a similar fashion to a newly launched DEX, as a user, liquidity is often the biggest issue I have found with these newly launched prediction markets; the age old cold start problem!

Onchain prediction market coins outperforming

And yes, unsurprisingly, this uptick in attention has been accompanied by prediction markets related coins seeing some very strong moves. With the largest platforms like Polymarket and Kalshi being tokenless (for now!), most of the coins in this sector fall into the small-microcap range and as a natural consequence, tend to be of lower quality.

Unsurprisingly, the onchain runners here seem to be concentrated in both the Base and Solana ecosystems with coins like TRUE (Truemarkets) performing incredibly well;

While today’s weekly has been focused mostly on newer developments within the prediction markets sphere, I would be remiss not to mention our Polymarket functionality at parsec. It remains the premier choice for Polymarket wide analytics and I would strongly recommend checking it out if you are not already familiar: parsec.fi/polymarket

the gpu backed dollar

-Charlie

Following last week’s dive into stablecoins, today I want to look at a newly launched project sitting at the intersection of stablecoins and RWA that’s caught my attention.

Most stablecoins and RWA-backed tokens draw from the same collateral pools and sources of yield. Unlike Ethena, which relies on derivatives markets, or Cap, which in effect aggregates DeFi strategies, usd.ai is betting on AI hardware as collateral — a fundamentally different source of yield.

The problem they identify is that, as it pertains to financing AI infrastructure, pools of traditional capital seem to fall into one of two buckets:

• 9 fig+ minimum deal size - the largest firms on wall street

• zero willingness to lend - the average high street bank

This has created a financing gap for small to medium sized infrastructure operators. While they can often access trade financing from OEM’s to finance the initial GPU purchase, they struggle to access longer term debt for operations/growth via traditional financing methods.

Coreweave can tap capital markets for $7.5 billion in a single capital raise but a small sized operator might struggle to raise $20m. With regional and international banks unwilling to underwrite such infrastructure, many of these small to medium sized operators have been limited to expensive equity financing.

This is what usd.ai propose to solve; by connecting on chain capital to small to medium sized operators with a rigorous underwriting methodology.

In practice, usd.ai is implemented with a dual token model:

• USDai = dollar-pegged, collateralized, liquid stablecoin

• sUSDai = 15-25% yielding staked USDai wrapper which generates yield from: interest on loans made to GPU operators (GPU sale leaseback style arrangements), treasury yield on idle USDai, and redemption queue fees (QEV).

sUSDai relies on three key mechanisms: CALIBER, FiLo and QEV.

CALIBER

CALIBER is a smart legal-financial wrapper transforming GPUs into on-chain NFTs. Borrowers legally transfer ownership of hardware to a “Tokenizing Agent” then retain use via bailment. That NFT represents title, so if they default, it's automatically reclaimed by the lender. In theory it snips away courts and delays, bringing real collateral enforcement into DeFi.

FiLo

FiLo (First-Loss, Last-Out) is the underwriter layer that sits beneath depositor capital. Curators originate loans and post junior capital that absorbs defaults before sUSDai is touched. In exchange they earn higher yields and origination fees. It’s skin-in-the-game underwriting that lets depositor yield scale without compromising safety.

QEV

Redeeming sUSDai isn’t instant (because you can’t sell 100 GPUs overnight), so the protocol uses a redemption queue + auction system. If you need out fast, you bid to jump the line, generating extra yield that gets redistributed. It's liquidity on demand—with a risk-adjusted price tag.

Of course, the risks here aren’t trivial — GPUs depreciate fast, borrowers can default, and ultimately redemption liquidity is only as strong as GPU resale markets. The design (CALIBER, FiLo, QEV) is meant to mitigate that, but real-world risk doesn’t disappear.

USDai mints are approaching 60m with 10m coming in since they opened up to public deposits on the 18th August. This has been helped along by the coincident announcement of their Allo program. Allo is an ICO<>Airdrop program which allocates points depending on how depositors interact with the protocol.

If AI is the new oil then usd.ai are certainly making a bold attempt at creating the AI petrodollar. Whether it is able to scale safely is yet to be seen but it’s certainly one of the more ambitious experiments in the stablecoin space right now and a project I will be keeping firm eyes on.

The Battle of Stablecoin

-Charlie

It’s not controversial to suggest that to date, stablecoins have been Crypto’s killer use case with the most profitable company in the industry being a stablecoin issuer (Tether). We’re entering a new phase of the stablecoin wars: not about who issues them, but who controls the rails they move on.

I recently saw a data point that 35%+ of total blockchain fees are USDT transfers (which is insane):

Meanwhile, stablecoin supply looks like this (on Ethereum mainnet alone):

With that said, it’s no surprise that we’re seeing an influx of projects looking to capitalise on this trend, including a plethora of new L1s, each bidding to be “the stablecoin chain”. Competition is intensifying with both stablecoin issuers and fintechs in the race, today I wanted to break down the most prominent front runners, especially given some recent announcements from both Circle and Stripe.

Plasma

Initially touted as “the USDT chain”, Plasma is a stablecoin L1 designed to make USDT transfers free and instant. Technically, Plasma is an EVM-compatible Bitcoin sidechain running “PlasmaBFT” consensus for rapid finality. It features a native BTC bridge, whitelisted gas tokens, and zero-fee USDT transfers.

The team recently ran a blockbuster token sale with $373 million in commitments (7x larger than the initial $50m target raise). As part of the sale, investors were mandated to commit stablecoin deposits to bootstrap initial TVL, thanks to this mechanism Plasma mainnet will launch with $1b in stablecoin TVL - a healthy head start!

Stable

Prior to the announcement of Stable, it seemed as though Plasma was Tether’s bet for “the USDT chain”. However, with Tether on the cap table for Stable too, it seems clear they are hedging their bets.

While technically different, the idea here seems to be almost identical to Plasma. Stable will be a high-performance, EVM-compatible L1 built for USDT and running StableBFT (CometBFT-based PoS). It uses USDT as native gas and offers gas-free USDT transfers and unwraps for frictionless stablecoin transactions.

While it makes sense for Tether to hedge their bets with respect to execution, the risk here would be liquidity and user fragmentation across both chains. Interoperability will be a key question here and one to monitor from an investment perspective.

Arc (Circle)

Arc was just recently announced on Aug 12th with this blog post, it looks to be Circle’s answer to the L1 bets Tether are making.

In summary, Arc is an EVM-compatible L1 purpose-built for stablecoin finance, using USDC as native gas and sub-second finality via “the Malachite consensus engine”. It features a built-in FX engine, opt-in privacy, and deep integration with Circle’s payment, minting, and cross-chain tools. The target use cases are enterprise-grade payments, FX, capital markets, and tokenisation.

Tempo (Stripe)

And like clockwork, an announcement from Stripe shortly followed! For now we don’t have much context here, just that Stripe is developing a blockchain called "Tempo" in partnership with Paradigm. It is being described as a “high-performance, payments-focused blockchain” - basically the exact same use case/description as every single one of these chains.

Interestingly, Paradigm’s Matt Huang will act as inaugural CEO - Paradigm are putting both significant capital and human capital behind this one. Unlike the aforementioned chains, Tempo will directly benefit from Stripe’s pre-built distribution prowess. Stripe can route merchant payouts and cross‑border flows over Tempo on day one, abstracting the chain behind its API surface. If Stripe end up issuing their own stable (which could be logical given the Bridge/Privy acquisitions), they’d be capturing the entire stack.

Winners and Losers?

With Plasma and Stable being focused on USDT, the logical loser here would be Tron given that Tron wears the “USDT chain” crown at present. Applying the same logic, as a current leader in stablecoin TVL and transfer volume Ethereum mainnet would be another likely loser.

However, this assumes that the introduction of these L1s doesn’t bring net new users and therefore issuance. Tron already has very strong distribution in certain regions and Ethereum mainnet has a very strong head start from a DeFi/Liquidity perspective. It’s not as though the introduction of Solana or any other new cheap and fast L1 has meaningfully cannibalised Tron or Ethereum stablecoin usage so it will be interesting to see whether any of these new L1s changes the calculus.

If Treasury Secretary Scott Bessent’s estimations for stablecoin growth are even directionally correct, there will be plenty of value to be had for all of these players.

My base case is that their distribution or go-to-market strategy makes a much bigger difference than the tech stack (surprise, surprise) and that rather than purely eating market share from incumbents, these new L1s will grow the pie and bring net new users and entities on chain.

Based on the information we have at present, there will likely be more opportunities for us Crypto Natives to be involved with both Plasma and Stable. Plasma token is confirmed (XPL) and I believe it is likely that Stable will go down the token route too.

As a humble retail minnow myself, I will continue to monitor ways that I can interact with these projects both from a usage and investment perspective. Similar to what we saw with the XPL token sale, there’s a high chance that there will be further opportunities for public/retail investors to benefit from this trend. I look forward to discussing such opportunities when they arise, so stay tuned for that!