# DeFi — X 热门讨论 (2026-09-30 14:13 UTC)
## @amirhp771 (amir(❖,❖) inkog.base.eth🟣🟢) · 09-30 13:16 · ♥223 ↻17 💬3 Rally just landed on Robinhood Chain, and the first campaign I want to see is for a dollar most people here hold without ever seeing it.
USDe.
There is about $343M of it on this chain right now.
Roughly 7% of all USDe in existence, about triple what it was in mid July, on a chain with 26M addresses.
Only 7,233 of them hold it.
So who is using it?
A lot of it works for people who never chose it.
Robinhood Earn takes your USDG, deposits it into a Morpho vault curated by Steakhouse, and lends it to borrowers posting collateral from Spark, Ethena and Maple.
You deposit one dollar. Another one sits quietly under your yield.
My pick: @ethena
Here is what sold me after digging through their dashboards.
The story most people repeat about USDe is out of date.
"Crypto collateral hedged with shorts" was the whole product a year ago.
Today Ethena's own numbers show $4.90B of supply against $4.97B of backing (101.36%), and crypto basis is only $930M of that, roughly 19%.
The rest is DeFi lending (33%), liquid stables (29%), institutional lending (13%) and RWA (6%).
Supply peaked north of $14B last autumn. It is a different animal now, and the timeline meme has not caught up.
And it is still moving.
On Sep 25 Ethena announced it is extending the basis trade to equity perpetuals with Binance, with tokenized stocks as spot collateral.
On a chain built for tokenized stocks, that one is worth watching.
The redemption door matters just as much.
Mint and redeem run against USDC, USDG, USDtb, PYUSD and USDT, with USDG at 0 bps both ways, but only for onboarded users.
Everyone else exits through the market.
Add up the redemption-ready stablecoins on Ethena's dashboard and you get roughly $1B, about a fifth of supply.
My honest take: I trust Ethena more after reading it than before.
A 101.36% backing ratio, a $62.1M reserve fund, custodian attestations, proof of reserves, all public.
Most protocols ask for blind faith in the peg. This one hands you the receipts.
That is exactly why it deserves a Rally campaign.
Rally scores accuracy, and Ethena is a project where claims can be checked against live public data.
A post about it cannot survive on vibes.
The uncomfortable part: transparent is not the same as understood.
On a chain where the front door is a brokerage app, a synthetic dollar can end up carrying real weight in the yield without depositors ever learning what backs it.
Not because the design is flawed. Because nobody asked them to look.
The timing is sharp too.
Ethena said token incentives on USDe end this month.
From October 1, the yield is only what the protocol actually earns.
What a campaign paid in ENA could reward:
◈ Explaining in plain words what a dollar backed 19% by a hedged crypto trade and 33% by DeFi lending actually is ◈ Following one Earn deposit from USDG to a borrower, and what a liquidation looks like ◈ USDe next to USDG: what backs each, how each redeems, what can break each ◈ Reading the funding-rate table next to the basis yield and saying how thin it is right now
All checkable against Ethena's public dashboards. Scored on accuracy, not follower count.
Which Robinhood Chain project deserves a Rally campaign next, and what is the one thing you think its users hold without really understanding? > 引用 @RallyOnChain: RALLY IS LIVE ON ROBINHOOD CHAIN 🚨
Robinhood built an AI-native chain, so we brought the AI-native attention layer to it.
Starting today, any project on @RobinhoodCrypto can fund a Rally campaign with its own token and put 150K+ creators to work on its story.
Welcome 💚 https://t.co/FkWc6rbgAe https://x.com/amirhp771/status/2105285608046227540
## @mdabdurrahim98 (Vexora ☣) · 09-30 12:54 · ♥111 ↻2 💬122 The KBW booth is not the part I would focus on with @zerufinance.
The bigger thing is that zScore is already sitting on Etherscan.
Now ZeruAI is building ways to actually use that wallet history through Zaps, screening and credit signals.
If you have an older DeFi wallet, connect it to Zaps and check what your past activity has already earned.
https://t.co/sPyl9iFiuw https://x.com/mdabdurrahim98/status/2105280190826037335
## @Vivaldi100x (Vivaldi) · 09-30 09:08 · ♥79 ↻0 💬99 You sign a message and that is the whole handshake.
Up to ten wallets. One name.
No transfer. No spend approval. The old DeFi wallet went in with no new fill on it.
wait.
The starting zaps were already sitting on that history.
Why do we need history to become a point, and the point to become standing.
Because if allocation ever looks at standing, it cannot start on the week you found the app. It has to start on the fills you already paid fees on.
@zerufinance reads those wallets. It does not hold them.
Connect the ones you actually trade with and check the number.
https://t.co/R6UaDiONzf
https://t.co/uUFR42pAal > 引用 @zerufinance: ZeruAI is on the ground at KBW. 🇰🇷
@Ashutosh_Sahoo and the team are at the booth, bringing behavioral intelligence into practice with zScore on @etherscan, wallet screening APIs, and Zaps.
Come find us, say hi, and there's $50,000 on the table. https://t.co/qdVvgbh5Ck https://x.com/Vivaldi100x/status/2105223296144851384
## @cryptocevo (cevo) · 09-30 13:38 · ♥135 ↻14 💬2 $NPC https://t.co/xUClrpBwwB > 引用 @Aurelius0121: @cryptocevo @tikoxbt Ahahahha I have $NPC and $ANYONE in a DeFi wallet. Inshallah, God willing, I’ll add more publicly as well. I’ve been holding them since the last cycle https://x.com/cryptocevo/status/2105291265281872249
## @chjokka (chjokka) · 09-30 13:17 · ♥73 ↻18 💬8 From Thursday onwards, all eyes will be on $NET and its upcoming MMORPG / DeFi Kingdoms game.
Last chance ever to get in below $500. > 引用 @NetNetCap: Shareholders,
We will be hosting a spaces on Thursday, October 1st at 4pm EST to discuss all things NetCorp, our novel MMORPG currently in Beta.
Excited to share more, including full demos and information on getting early access.
More Soon. https://t.co/1lmeKDrTZi https://x.com/chjokka/status/2105285977459863573
## @gandreou007 (Giannis Andreou) · 09-30 11:30 · ♥75 ↻4 💬2 🚨 COINBASE SELECTED $LINK FOR TOKENIZED STOCKS
Coinbase has chosen Chainlink as the official oracle infrastructure for its tokenized stocks on Base.
Chainlink Data Feeds provide continuous pricing for tokenized shares, including Apple, Nvidia, Meta and Alphabet.
This gives developers the data needed to bring tokenized equities into DeFi lending and trading. https://x.com/gandreou007/status/2105258862567735308
## @MSBIntel (MSB Intel) · 09-30 09:27 · ♥69 ↻7 💬2 BREAKING: Standard Chartered sees potential for Ethena’s USDe to grow roughly 8x over the next two years as DeFi and TradFi converge. https://t.co/Djvqptklqd https://x.com/MSBIntel/status/2105227936005796012
## @KevinWSHPod (MR SHIFT 🦁) · 09-30 11:00 · ♥63 ↻0 💬7 Who Owns AI? Michael Heinrich on 0G and Building an Economy That Pays People
In this episode of DROPS, I sit down with @michaelh_0g, Co-Founder and CEO at @0G_labs, to discuss what it could mean for ordinary people to actually own a piece of AI. We unpack the idea of compute finance, why AI compute could eventually become a financial asset, and how people might participate in the AI economy through their data, hardware, expertise, or ownership. But this conversation is also about what happens when things do not go according to plan, from launching a multibillion-dollar token before the ecosystem was fully ready to navigating burnout and a brutal bear market.
Building for a Mission
Michael describes himself as a “mission-driven entrepreneur,” someone who believes financial outcomes should follow excellent decisions rather than being the sole reason for making them. He traces that outlook partly to his time at Bridgewater and his experience building a Y Combinator-backed company that scaled to roughly 700 employees before COVID forced the business to close. The company was eventually sold, but Michael is honest about the fact that it did not reach the potential he originally imagined.
That mindset shaped how he approaches 0G. Rather than treating that chapter as a clean success or failure, he sees it as information. “Failure is loaded… But for me, it’s really just information.” That explains why his current mission is so ambitious: he wants to make AI a public good and, in his words, build technology that allows AI to “serve humanity.”
Launching Before Everything Is Ready
The team was operating from a war room in South Korea, two weeks before 0G’s token generation event, when they learned that critical parts of the DeFi ecosystem would not be ready. A related treasury strategy they expected to support the ecosystem was also delayed. The team had a choice: postpone everything, or launch and build the missing utility afterward. They chose to launch.
Michael estimates that the decision was “60% momentum, 40% conviction.” Exchanges, investors, marketing, and expectations were already moving, and the token ultimately debuted around a $3 billion valuation with enormous first-day trading activity.
But the celebration did not last long. The team was exhausted, with some people even needing medical leave. Then deteriorating market conditions hit, causing the token price to decline, and Michael noticed that people were beginning to tie their sense of success to a market price they could not control. He immediately responded by pulling the company back towards building, something they could actually control.
Compute Finance: Turning AI Into Something You Can Own
Imagine holding an asset that gives you access to AI compute, the underlying resource that makes AI possible. Instead of paying a fixed subscription to one AI provider, you receive units of computing capacity that you could use, trade, or potentially hedge against future price changes. Michael argues that if AI agents eventually operate more independently, they will still need to pay for the computing power required to keep themselves running, transforming compute from a simple technology expense into an underlying economic resource.
The idea becomes particularly interesting when AI agents themselves have economic activity. Revenue or transaction fees generated by an agent could theoretically finance the compute required to keep that agent operating. In other words, the ambition is to make intelligence itself ownable and economically programmable.
The Bigger Question: Who Benefits From AI?
Michael believes the most important divide in AI shouldn’t be about systems being technically centralized or decentralized. It is whether ordinary people can participate economically in what AI creates.
Today, users provide enormous amounts of data, expertise, and behavior that help improve digital systems. In a more open model, Michael argues that someone who contributes useful data could consent to its use and be compensated. For instance, someone with specialized knowledge could contribute training data to an expert model and potentially participate in the revenue it generates. Others could provide hardware or compute.
So the question is: are people becoming participants in the AI economy, or are they simply being displaced by systems owned by a small number of organizations?
For Michael, decentralization matters because ownership matters, and participation does not necessarily require being an AI researcher. He compares it to the early crypto networks, where ordinary people eventually learned how to mine, stake, run infrastructure, or simply own the underlying assets. He expects AI networks to create similar entry points around data, hardware, expertise, and compute.
AI Abundance Still Needs Human Values
Beyond finances, Michael imagines AI eventually creating abundance in time, thought, and the freedom to focus on things that make us human. Personal AI assistants could remove routine work, and over a longer horizon, robotics could extend that assistance into the physical world.
However, as AI systems become better at cybersecurity, autonomous action, and solving problems humans cannot, questions about what they should do become just as important as what they can do.
Compute finance sounds like a new financial concept, but underneath it sits a much older question about technology: who owns it, who benefits from it, and what kind of world are we trying to build with it? Michael bets that AI does not have to be something that just happens to people. If ownership and participation are designed differently, people can contribute to the system, own part of the infrastructure behind it, and earn from the value that AI creates.
👉If you enjoyed reading the summary, head over to When Shift Happens on YouTube or your favorite podcast platform to access the full convo. https://x.com/KevinWSHPod/status/2105251379422068812