Pretty smart investment by Solana imo.
The @SolanaFndn announced its investment in @CoinGecko at Breakpoint, with plans to integrate Solana-powered trading directly into the platform.
Consider the distribution theyโre tapping into:
+ 18M registered users
+ 180K applications and businesses using its API
+ CoinGecko, GeckoTerminal and its mobile app
CoinGecko has spent years building a platform where people go to figure out what to buy, only for them to end up trading somewhere else.
You discover a token, look at its historical performance, check the circulating supply, compare valuations and then go to another platform to execute the trade.
CoinGecko makes money selling ads and data while another business collects the trading fees.
Always found that dynamic a bit strange.
If CoinGecko can convert even a small percentage of its existing users into active traders, it could build a fairly large trading business without spending ridiculous amounts on customer acquisition.
Also quite like the economics of this deal from Solanaโs perspective.
Chains regularly spend millions of dollars incentivizing users to bridge over and trade, only for a good portion of that activity to disappear once the incentives end.
Solana is instead investing in a business that already attracts millions of potential traders.
And the best part is that CoinGeckoโs users are already financially motivated. You donโt have to convince them to speculate on crypto.
You just have to make it easier for them to execute.
Personally think acquiring existing distribution is one of the smarter ways for mature blockchains to deploy their capital.
Why pay people to come to your casino when you can own a piece of the place where they go to research their next bet?
You donโt necessarily need to leverage your stablecoins to earn double-digit yields in DeFi.
Here are 4 yield destinations on my watchlist:
1. stUSDS | @SkyEcosystem
APY: ~14.2%
Chain: Ethereum
Supply USDS to fund borrowing against staked SKY. Returns depend on borrowing demand and market utilization.
2. apyUSD | @apyx_fi
APY: ~13.3%
Chain: Ethereum
Deposit apxUSD into the Apyx vault to earn dividends from preferred equity issued by digital asset treasury companies, including STRC and SATA.
3. ONyc | @onrefinance
APY: ~11%
Chain: Solana
Earn from reinsurance premiums and income generated by the underlying collateral. The main risk is underwriting losses.
4. AUTO | @HastraFi
APY: ~11.7%
Chain: Solana
Earn interest from tokenized consumer auto loans originated by Agora Data.
Personally, I like the diversification across these four products.
Youโre earning from four different sources: borrowing demand, preferred equity dividends, reinsurance premiums and consumer credit.
All without manually looping your positions.
PS: Rates as of November 16โ17. Yields are variable.
What if the most profitable AI agent isnโt the one selling the final product?
Imagine an AI agent charging $100 for a DeFi research report.
Behind that report, the agent purchases:
+ Historical onchain datasets
+ Smart contract risk assessments
+ Updated lending market analysis
Each service is supplied by another specialized AI agent.
Now imagine these providers selling their services to hundreds of research agents rather than relying on a single customer-facing application.
This is essentially how supply chains work in traditional industries.
The company selling the finished product doesnโt necessarily capture the highest margins. Sometimes, itโs the supplier of a difficult-to-replicate component.
@termix_ai enables similar commercial relationships through AACP, where agents can operate as both buyers and service providers.
An agent can generate revenue from selling its services while outsourcing specialized tasks to other agents.
From a business perspective, there are several factors that could determine which providers capture the most value:
+ Uniqueness of the service
+ Availability of competing providers
+ Accuracy and reliability of outputs
+ Frequency of recurring purchases
+ Execution and coordination costs
A widely available dataset might face intense pricing competition, while a specialized risk assessment service with limited alternatives could command a premium.
The same economic principles that govern traditional B2B markets could eventually apply to autonomous agent commerce.
I think the AI agent economy will develop its own version of the picks-and-shovels business model.
Beyond consumer-facing AI applications, there could be an entire category of businesses generating recurring revenue by supplying individual capabilities that other agents depend on.
They might not have the largest user bases or the most recognizable brands.
But they could have some of the strongest commercial relationships in the entire agent economy.
One of the more unusual things about @termix_ai is how differently its business is developing across @Base and @BNBChain.
Base accounts for ~72% of recent protocol fees, despite holding only ~10% of tracked TVL.
The numbers are quite revealing.
30-day protocol fees:
+ Base: $54.7K
+ BNB Chain: $21.2K
TVL:
+ Base: $302
+ BNB Chain: $2.81K
Registered AI agents:
+ Base: 1.17M+
+ BNB Chain: ~411K
Base has nearly 3x as many registered agents and generates 2.6x more protocol fees.
BNB Chain, meanwhile, holds over 9x more capital in TermiXโs tracked contracts.
These are different measurements, but the contrast reveals something about how the ecosystem is developing.
TermiXโs primary business is facilitating paid transactions between AI agents and service buyers, rather than accumulating deposits.
AACP handles agreements, escrow, delivery, and settlement, with fees generated through commercial activity.
This makes Baseโs contribution particularly notable.
Itโs already the largest network for both registered agents and recent protocol fee generation.
And with 612K+ settled jobs and over $626K in cumulative fees reported across TermiX, the marketplace is developing a measurable commercial footprint.
For a protocol built around autonomous work, the ability to generate transaction fees across multiple chains is far more consequential than the amount of capital parked in its contracts.
Base is currently leading that side of TermiXโs business.
the next evolution of crypto exchanges may not be about adding more trading pairs. it may be about giving the same capital more ways to work.
thatโs the direction i see @grvt_io moving towards.
over the past few weeks, grvt has been expanding beyond its traditional perpetual trading infrastructure, with developments across yield generation, institutional finance, real world assets and trading execution.
there are a few things worth connecting.
first, the trading infrastructure.
grvt has been working on improving liquidity and execution quality. in its latest liquidity update, it shared a comparison covering 1,356 snapshots across 16 market and order-size combinations.
the reported results placed grvt second overall behind binance, with lower order-book slippage than hyperliquid in 12 of the 16 tests.
then thereโs the move into yield generation.
grvt earn automatically rewards eligible usdt and usdc balances with up to 3.5% apy.
high yield earn takes a different approach, targeting up to 11% apy through diversified sources including:
โ real world asset financing
โ tokenised fixed income
โ crypto credit
โ market-neutral strategies
the mirae asset securities collaboration adds another dimension, with grvt subscribing to tokenised funding notes and exploring further tokenised financial products.
its partnership with axcess finance also introduces an onchain version of grvtโs glp strategy and institutional credit infrastructure.
the trading side is evolving too.
stable funding perps bring real-world assets such as gold, silver and crude oil into a different perpetual trading structure.
the model uses live market-maker quotes, zero trading fees and once-per-weekday funding settlement. traders still need to account for the bid-ask spread and funding costs.
but the development i find particularly important is what grvt is building around capital efficiency.
the planned yield-as-margin functionality is intended to let eligible earn balances support trading positions while continuing to generate yield.
think about the implication.
instead of maintaining completely separate pools of capital for earning and trading, eligible users could eventually put the same balance to work across both activities.
this is also the broader idea behind @grvt_io 2.0: bringing trading, investing and yield generation together within a single self-custodial experience.
there is still plenty to prove. liquidity needs to remain competitive, yield sources need to perform, and the integration of these products has to work in practice.
but i think the more interesting way to evaluate grvt is to look beyond any single feature.
itโs attempting to build a financial platform where capital doesnโt have to sit idle between trading opportunities.
and execution, accessibility and risk management will determine how successful that vision becomes.
not financial advice. yield targets are not guaranteed, and every product carries its own risks.
One of @primus_labsโ less-discussed products isnโt about encrypted balances or confidential transactions.
Itโs about sending crypto to someone without needing their wallet address.
Primus has explored this through its social-identity payment infrastructure, connecting external account verification with crypto payment claims.
The idea is straightforward.
A sender might know someoneโs X account or another supported online identity but have no idea which wallet belongs to them.
Instead of requesting a wallet address, the payment could be associated with that external identity.
The proposed workflow:
1. Payment creation: A sender specifies the recipientโs supported social identity and allocates funds.
2. Payment allocation: The payment system records the funds associated with that identity.
3. Identity verification: The recipient uses Primusโs zkTLS infrastructure to prove control of the relevant external account.
4. Payment claim: Once the verification requirements are satisfied, the recipient can claim the funds through a wallet.
Primusโs public repositories include tip-contracts and fund-js-sdk, providing evidence of development around this payment infrastructure.
And the potential applications extend beyond ordinary transfers.
Creator tipping: Reward a creator using their social identity rather than requesting a wallet address.
Community distributions: Allocate payments to verified community members without collecting wallet addresses beforehand.
Referral rewards: Connect payouts to authenticated external accounts.
Contributor payments: Distribute rewards to people identified through supported online platforms.
What I find compelling is how this changes the role of zkTLS.
In financial verification, Primus authenticates information about balances, positions or external activity.
Here, the same underlying capability could establish whether someone controls the identity associated with a payment.
That creates a connection between existing internet accounts and onchain financial transactions.
There are still practical questions around account recovery, impersonation, supported platforms and unclaimed funds.
And the public repositories alone donโt establish which payment flows are currently live, how many users they serve or how much transaction volume theyโve processed.
But the product direction is notable.
Primus is often discussed as infrastructure for confidential finance and verifiable data.
Its social-identity payment work points toward a more consumer-facing application of that infrastructure.
A user wouldnโt necessarily need to understand zkTLS to benefit from it.
They would simply need a way to receive crypto through an online identity they already control.
s.kaito.ai/Zhl9J3o
a big part of my research is keeping track of tokens and pools i donโt want to lose sight of.
i use @GeckoTerminalโs watchlist for this. iโve added mihao, ANSEM and CASHCAT so i can easily check how theyโre performing without having to search for them every time.
when checking a token, i donโt just look at price. trading volume, transactions, market cap, pool age and security indicators also give me useful context.
i can quickly see whether activity is increasing, slowing down or if thereโs been a change worth looking into.
thatโs why i find the GT watchlist useful. i have the markets iโm following in one place instead of jumping between charts.
hereโs my watchlist:
geckoterminal.com/watchlist
whatโs on yours?
$14.2B disappeared from stablecoin supply in 30 days, while tokenized gold held comparatively steady.
Thatโs the largest monthly stablecoin contraction since the 2022 bear market.
โข Major stablecoin supply: -$14.2B in 30 days
โข $USDT + $USDC: Nearly $11B in combined contraction
โข $USDT alone: -$6.6B over 60 days, per CryptoQuant
โข Gold-backed tokens: $5.61B in market capitalization
Two different asset categories, two very different reactions to market stress.
$USDT and $USDC function primarily as dollar liquidity for trading, collateral and settlement. When their supply contracts, thereโs less dollar-denominated capital circulating through crypto.
Meanwhile, gold-backed assets like $XAUT and $PAXG have maintained a comparatively stable combined valuation.
But thereโs an important caveat.
Gold-token market capitalization depends on both circulating token supply and the price of gold.
A stable valuation doesnโt necessarily mean investors are minting more tokens or that redemption demand has disappeared.
To establish actual capital rotation, weโd need to compare net issuance and redemptions across $XAUT and $PAXG, not just their dollar market caps.
The broader divergence is still revealing.
Dollar liquidity is contracting sharply, while tokenized gold hasnโt experienced the same reported decline in market value.
If gold token circulating supply also holds steady, that would provide stronger evidence of different investor behavior during this period of market stress.
a perpetual position can be right on direction and still become expensive to hold. funding and execution costs have a way of eating into the trade.
traditional perpetual futures typically settle funding every 1, 4 or 8 hours, depending on the exchange and market.
that means traders holding positions for several days have another cost to monitor, alongside price movement, leverage and liquidation risk.
@grvt_io is approaching this differently with its upcoming stable funding perps.
instead of relying on the conventional crypto perp funding mechanism, grvt is introducing a model with:
โ once-daily funding settlement at market close
โ funding rates submitted by market makers, reflecting their tradfi cost of funds
โ separate funding rates for long and short positions
โ live buy and sell quotes, with the spread determining trading costs rather than a conventional trading fee
the distinction is worth understanding.
funding frequency alone doesnโt determine whether a position is cheaper to hold. the actual funding rate, spread, entry price and duration of the trade all contribute to the final cost.
what i find particularly interesting is the funding mechanism itself.
grvt is bringing a more traditional financing approach into perpetual-style trading, rather than simply changing the settlement interval on an existing crypto perp model.
for traders, having a predictable daily settlement schedule could make holding costs easier to monitor and plan around.
but the real test will be how competitive those funding rates and execution spreads are once the markets go live.
less complexity is useful. better trading economics are what ultimately matter.
No more 1, 4 or 8-hourly funding.
No more thin order books.
No more paying up for size.
Stable Funding Perps. Coming soon to Grvt.
Hereโs how it works, in 4 slides ๐
Set up your account today: grvt.io/a/g/whnr39m3
One thing I like about @primus_labs is that the two technologies theyโre building donโt actually need each other to become businesses.
And I think thatโs a pretty underrated part of their setup.
zkTLS is basically a verification business.
A stablecoin issuer wants to prove its exchange reserves, a lender wants to verify someoneโs financial history, or an application needs authenticated data from an external platform.
They can pay for those proofs without ever touching an encrypted vault.
FHE is a different market entirely.
A fund wants to move collateral without exposing position sizes, a company wants to pay employees privately, or a stablecoin application wants encrypted balances and transfers.
None of those necessarily require authenticating external accounts.
So you potentially have two separate customer bases.
One pays for verification.
The other pays for confidential computation and the infrastructure around it.
And when a financial application needs both, like proving reserves while keeping sensitive positions confidential, the two technologies become complementary.
Thatโs where the architecture gets more compelling IMO.
But thereโs a catch.
Building two technologies doesnโt automatically mean youโve built two successful businesses.
zkTLS already has measurable verification fees, although revenue is still small. The Confidential Vault demonstrates working FHE functionality, but commercial adoption and recurring revenue from that side are much less established.
Youโre also dealing with different developers, integration requirements, customer acquisition costs and infrastructure expenses.
So execution becomes harder, not easier.
Still, I prefer this setup to a privacy project whose entire future depends on one application taking off.
Primus doesnโt need every financial institution to adopt encrypted DeFi for its verification business to grow.
And it doesnโt need every verification customer to use FHE either.
The bet is that authenticated financial data and confidential financial computation can each develop their own demand.
If both become recurring businesses, their combination becomes an additional advantage rather than the entire investment thesis.

















