Introducing: Built in Venice
Private, uncensored AI in the wild. SDKs, agents, on-chain experiments, dashboards, and more, all shipped by the Venice community.
The CLARITY Act could make this more valuable for Coinbase too.
If passive USDC rewards become harder to offer, Coinbase needs products where yield is tied to explicit activity: lending, collateral, liquidity, or platform usage.
Ethena gives them a way to route yield-seeking USDC users into real borrow demand, rather than just paying rewards for holding USDC.
Reading between the lines for the upcoming product launch referenced.
Coinbase x Ethena is bullish because it can turn Coinbase’s ~$19B USDC base, with an implied ~$13B of reward-earning balances, into a funding rail for Ethena.
If sUSDe yields clear baseline USDC rates, Coinbase can offer better USDC lending yields, loopers can lever the spread, and Ethena gets deeper/cheaper funding than native DeFi alone.
Aave mechanics, Coinbase distribution.
We think Ethena will keep finding compelling ways to source yield across all markets, especially as RWAs and more financial activity move onchain. Coinbase gives it a much larger funding base to apply that yield against.
@Delphi_Ventures are investors in @ethena
Excited to partner with Coinbase for the first time to support their dollar savings products. The upcoming integration next week will be the first time Ethena products are available for their 100m+ user base.
Given the evolving nature of the Clarity Act, we expect further potential tailwinds for onchain native products like USDe from idle balances on exchanges, and Ethena is well positioned to support this transition.
Most investors still frame neurotech around the ~$65B 2035 clinical TAM.
That market matters, but I think it understates the prize. The bigger question is whether BCIs become the next major human-machine interface layer: keyboards → touchscreens → thought.
The reason this feels investable now isn’t that the science suddenly appeared. It’s that the engineering stack is finally catching up: EEG hardware costs have collapsed, neural decoding is moving to low-power edge/implantable chips, and materials like PEDOT + graphene are already making their way into humans.
Neurotech has spent decades looking like “too early” frontier science. Increasingly, it looks like an underwritable platform shift.
Full Delphi thesis from @august_wstein and @DrewAHenderson:
https://t.co/L9f1YZupqO
Ethena and @coinbase have partnered to grow onchain finance and savings products for their 100m+ userbase, with the first growth initiative launching next week.
Alongside this partnership Coinbase Ventures have also made their first investment into Ethena on the open market.
Yan Liberman (@YanLiberman) explains why @grass is the most underappreciated AI infrastructure play in the market:
"GRASS made it. $12M in quarterly revenue as of November 2025. Sparse updates, that's why it's been overlooked."
"Residential opt-in bandwidth network. Users contribute unused bandwidth through a browser extension. GRASS scrapes the internet with it."
Why are the DeFi projects with positive token flows actually significant?
Yan Liberman says two reasons.
First: DeFi apps run incredibly high margins, so they can return capital to token holders earlier than most businesses would.
Second, and more important: they're forced to.
In crypto, there's a gap between equity holders and token holders. When things go well, everyone attributes value to the token. When things go south, who actually has claim to the business?
Projects returning capital early are establishing that the token is the value-accruing mechanism. That's the signal.
Happy Friday.
Can you feel it? That’s the feeling of regulatory green light.
CFTC approved the first US-listed BTC perpetual via Kalshi, Paxos became the only blockchain-native firm cleared by the SEC.
Jeff Sprecher said Hyperliquid is bigger than NASDAQ and the US is lifting its "blockade" of the Strait of Hormuz.
Wild way to close the week. Higher.
Today's lineup:
2:00PM ET @YanLiberman - @Delphi_Digital
2:30PM ET @Bencera - @polsia
See you on the show.
@YanLiberman gave venicestats a shoutout on @Bankless — pointed david to it as the place to follow what's happening with @AskVenice.
we just build the tools and keep the data honest. seeing that land in a room like this genuinely means a lot 🙏🏻
A new episode of The Hivemind is out now!
This week we discuss Venice AI’s momentum, Hyperliquid’s evolution beyond crypto trading, and more.
Timestamps:
00:00 – Crypto has become a stock picker’s market
05:00 – Venice AI, AI tokens & crypto business models
17:00 – ETH, SOL & the collapse of the L1 premium
24:00 – Hyperliquid changed the crypto landscape
33:00 – Crypto capital formation vs traditional equity markets
51:00 – Advice for investors entering markets today
55:00 – Defense tech, drones & the future of warfare
1:04:00 – Robotics, Figure AI & the next major investment wave
1:12:00 – Bitcoin, macro tailwinds & positioning
INTERVIEW: Why $VVV at $17 is cheap, w/ @YanLiberman
If the TAM for inference is ~$200B, what's the TAM for Private Inference?
@TrustlessState chats with @YanLiberman about
- Why @AskVenice is crypto AI’s clearest revenue story
- How he values $VVV
- How Venice serves agentic inference
- Why AI agents need $DIEM
[TIMESTAMPS]
0:00 Intro
1:40 Privacy Guarantees Explained
5:01 Who Needs Private Inference?
9:25 A Privacy Use Case
10:19 Venice’s Privacy Tiers
12:31 Growth Surges After New Plans
20:29 Inference Reseller Economics
23:39 Why Revenue Is Hard to Know
28:11 Token Value & Trust
32:07 @ErikVoorhees & Alignment
33:52 Venice’s Unique Product Edge
37:34 Adult Content Demand
41:18 Permissionless Agentic Inference
44:56 Bear Cases for Venice
49:12 How to Value VVV
51:46 Is VVV Undervalued?
56:10 Closing & Disclaimers
Best piece of research anyone has put out on @AskVenice.
What sets it apart isn't the bull case, it's the honesty. A real bear case, the token-vs-equity caveat spelled out plainly. Most $VVV theses skip both.
That's exactly why the constructive conclusion lands.
Strong work, @YanLiberman. Thanks for keeping it real 👏🏻
Why @yan is bullish on $VVV at ~$800M
Yan thinks VVV's market cap is the better measure than FDV
"market cap and FDV isn't always made equal...
[Venice] is one of those scenarios where market cap is actually probably a better representation of the value. ...so then you're looking at pretty low multiples on that front
...And I could see a world where by the end of the year, the token is deflationary"
Thanks! Agree on profit being lower. Gross margin is the more useful frame here. Net gets dragged by reinvestment, which Venice has said is intentional.
Numbers vary a lot by model on the OSS side. DeepInfra publishes list pricing for what they charge developers, so we can use that as a rough cost basis upper bound for Venice (real cost is probably lower given volume).
Llama 3.3 70B Turbo: DeepInfra list $0.10/$0.32 per M tokens. Venice charges $0.70/$2.80. Implies roughly 85% gross margin if Venice pays list.
Frontier API resale is tighter. Anthropic publishes Claude Sonnet 4.6 at $3.00/$15.00. Venice charges $3.60/$18.00. Closer to 20% margin on that one.
These are upper bounds on cost and lower bounds on margin. Blended OSS margin probably 40-70%. Frontier tighter.