Things to know 👀
Live in Greenwich, London, at the National Maritime Museum. Vlad Tenev, wearing a swanky pirate-like outfit, took the stage to announce Robinhood Earn, a 7% APY DeFi yield product. Stock Tokens in 120+ countries, trading 24/7 and usable as collateral in DeFi lending pools. And commodity, ETF, and FX perpetual futures in Europe (gold, oil, EUR/USD) at up to 10x leverage
🧠 The 7% earn product is fascinating. Where does the yield come from? A Morpho vault curated by Steakhouse invests in spUSDG, USDe, and SyrupUSDG, which gives overcollateralized lending to institutions. The yield is a mix of half native and half incentives.
🧠 Put another way, roughly 3.5% of the yield is marketing spend, for now. However, as the campaign grows, USDG volume grows, and it gains float (as it buys more treasuries). USDG shares its float with distributors, in this case Robinhood. So if all goes well, the other 3.5% should start to come from the stablecoin itself.
🧠 This is for all 27m users not just DeFi folks. The 7% yield product will be available in the core Robinhood app to its 27 million users (powered by Privy, the embedded wallets)
🧠 It's also insured by Lloyds of London against DeFi risks. So think about hacks, or smart contract exploits (which do happen!)
🧠 Stock tokens are super interesting. While not technically shares, and beneficial ownership. Robinhood is a registered broker dealer, so it can ensure 1:1 tracking against those shares. It also enables securities lending. So a user could take that stock token and borrow against it to pay for a down payment on a house or something.
🧠 Now combine the stock tokens with the yield product. If you have a very large “lending to traders” capability that you can earn yield on natively, and via your stablecoin, now you can lend to people who want to borrow against those stock tokens. And what you end up with looks a lot like a decentralized prime broker. Fascinating.
🧠 Their own chain has economic benefits too. They keep all fees for transactions instead of paying them. Those fees are negligible on most modern chains.
🧠 The contrast with Coinbase is notable. Coinbase has announced they'll be doing "stock tokens" soon, offers lower yield on USDC, and it offers perps through its advanced and institutional platforms but not core consumer product. Robinhood with USDG has more control over the stablecoin, and as a tenured, licensed broker, more scale and experience getting stocks and ETFs in place to back their products.
🧠 Valued just shy of $100bn, can Robinhood be the trillion-dollar finance company? There are no trillion-dollar finance companies. Not JP Morgan, not Visa, not BlackRock. This company is expanding Geo's products and into entirely new areas like prediction markets and AI faster than anyone else. They're arguably the finance company to catch (or tied with Revolut and Nubank).
🧠 Damn, Robinhood's marketing is slick. They do these events for their most engaged users, and they have TV-quality camera setups, a whole theme, and they announce maybe 10 different products.
We recorded a full interview with Johann live from their event and you can find it here on Tokenized:
The company is said to have held preliminary discussions with banks about a potential public listing, according to Bloomberg. Plaid was valued at $13.4B during the 2021 FinTech boom before resetting to $6.1B in 2025. Earlier this year, it rebounded to an $8B valuation as investor sentiment toward FinTech began to recover. If Plaid moves forward, it could become one of the most important FinTech IPOs of the year.
🧠 Plaid has successfully caught a little of the AI tailwind. They've launched integrations with both ChatGPT and Perplexity for people to access and view their own account data. This brings what was plumbing into secure plumbing for AI agents. Which could be much bigger than for apps or fintech companies.
🧠 They're also doing interesting things with LLM custom transformer models. I spoke to them last week about their new sequential foundation model, which was more effective at categorizing payments, predicting fraud or missed payments than previous machine learning attempts. That turns into more lending or less fraud for customers, which makes the product better, which grows revenue.
🧠 Plaid has successfully diversified its revenues. Its lending and fraud products now drive meaningful revenue for the company. The jury is still out on "pay by bank," but if that ever did happen, Plaid is reasonably well positioned there too.
🧠 As Plaid goes, Fintech goes. While Stripe, Robinhood, and Revolut are much bigger companies, Plaid is the bellwether for the industry. They acquired This Week in Fintech. In many ways, Plaid is the very heart of "Fintech" from 20019 to now. When they IPO, it’s graduation day. I just hope the window stays open long enough for them to make it. There are some nerves out there at the moment.
Erebor is at $4.05bn in deposits, per Bloomberg. Their March call report said $1.1bn. Nearly 4x in a single quarter, for a bank that only got its full national charter in February. They're now in talks to raise at an $8bn+ valuation, up from $4.35bn in December.
Founded by Palmer Luckey. Backed by Founders Fund, 8VC and Lux Capital. They went from OCC application to full charter in roughly 9 months. Paxos, Ripple, Circle and Stripe have all filed for trust charters. None of them operates a full depository bank.
🧠 They added roughly 400 customers this quarter while deposits grew ~$3bn. Even if every new dollar came from new customers, the average account is north of $7m. A small number of very large depositors, nearly all above the FDIC cap. Sound familiar?
🧠 Demand for crypto-backed lending came in below their expectations. The product winning is the boring one: a deposit account that stays open. Turns out the "chokepoint proof" bank might be a major feature for a certain clientele.
🧠 Luckey went out of his way to say none of this quarter's growth came from his own companies. When the CEO pre-empts that criticism, you know it landed the first time.
🧠 A fast-growing, concentrated, uninsured, correlated deposits are exactly the shape SVB was. The difference so far: per the Q1 call report, Erebor holds zero loans. It's operating closer to a narrow bank than a lender. And profitable without that lending.
Before Meta CEO Mark Zuckerberg directed employees to build a standalone prediction market app, he proposed buying Kalshi. NPR reports there was a meeting, but talks never advanced beyond that. In June 2025, about $28 billion was traded every month on Kalshi and Polymarket. A year later, monthly volume on the sites is nearly $220 billion, driven mostly by sports-related betting, according to The Block.
While the acquisition talks never advanced, Meta did strike a partnership with Kalshi in March, allowing for easy integration of Kalshi markets on Meta's social media app Threads.
🧠 Meta seems to be better at M&A than R&D. WhatsApp and Instagram may well be all-time great acquisitions. But with questionable AI and "Metaverse" investments, the ill-fated "Facebook Workplace," you have to wonder where next for the big tech firm, beyond better ad monetization.
🧠 Last week I wrote about Meta trying again at payments in India. After Facebook Credits, Libra, and countless other attempts, maybe by hiring the founder of CRED, they can add another major monetization stream.
4 Companies 💸
1. Novaquant - AI Decision Governance Layer
Novaquant sits between internal systems and AI models to observe every interaction, and ensure that those interactions are auditable and policy-aware. The policies, decisions, and information flow can then be put together into a complete decision-making history.
🧠 If you're regulated and sending data to models, you need to do this. I first wrote about the idea in the agentic oversight framework for Sardine back in early 2025. Companies have 100s, maybe even 1000s of policies; the link between those, your ops process, and what the model did can be incredibly hard to trace. Baking in a layer between them makes sense. Although, I wonder if this is a feature of bigger "enterprise AI control planes" that are emerging, rather than a product?
2. F2 - AI Native Deal Execution for Private Markets
F2 reads all historic memos, models, and reviews to build a structured, queryable history of every deal a firm has created. This intelligence can then be used to screen new deals, perform deep diligence that's ready for an investment committee, build a model in Excel, and then monitor the entire portfolio.
🧠 F2 are private credit nerds. There are folks like Hebbia and Rogo that are designed for a broader set of tasks and workflows. F2's differentiator is a deep understanding of credit markets, baked into the product. Time will tell if this means they become the first choice.
3. Wealthreach - The AI content and web engine for RIA's
Wealthreach helps RIA's create websites without a template that rank well for SEO. Companies can continue to edit and maintain their site, change copy, publish blog posts, and use the platform like a dedicated SEO agency and web agency combined. Living sites perform much stronger in SEO audits. The platform aims to create new content in the RIA's tone of voice based on inputs you give it, and retarget visitors to your site who didn't fill in a form with email and LinkedIn outreach. Living sites start at $300/mo, the SEO at $500/mo, and retargeting at $500/mo.
🧠 AI is producing companies operate in a niche within a niche. Imagine building this company 5 years ago pre-AI? Would be tricky. With AI, generating pages and copy isn't hard. Making it sound like an RIA and perform in SEO for that segment is.
4. Trustapp - Escrow for agents
Trustapp started as an online escrow solution to solve risk in marketplaces, which often suffer from high levels of fraud, with fake sellers or bad buyers. This results in high chargeback rates and high costs. They've pivoted their top-line message to be more focused on AI agents and a processor for any large-value transaction or B2B payment. The world needs processors who are specialists in complex transactions. And I wonder if being that is the best outcome. They're also focusing on helping these merchants get better at having AI discover them.
🧠 I must see two or three pitches to make your "e-commerce AI agent ready." But Car and B2B marketplaces need agent cataloging and discovery too. That vertical focus could be the wedge for Trustap.
Good Reads 📚
Microsoft CEO Satya Nadella argues that the value from AI won't come from "who has the best model," but from "who can build the learning loop around various models". Can you bring in all of your data, your context, and private IP that you don't want to share and push that to different models depending on the task? The skill of the next decade is building private evals and private reinforcement learning environments that help your AI get better at what it is you do, each time it does it.
🧠 This is clearly Satya's pitch to own the enterprise harness, control plane, and context that helps those companies build private evals and RL'd models. And given what lives in the Microsoft suite, they have a reasonable shot at it. I also strongly agree with the pitch and vision. As we wrote in our operating model playbook, the harness, the control plane, and your ability to make evals and private models is what separates the AI haves from the have-nots.
Tweets of the week 🕊
That's all, folks. 👋
Remember, if you're enjoying this content, please do tell all your fintech friends to check it out and hit the subscribe button :)
Want more? I also run the Tokenized podcast and newsletter.
(1) All content and views expressed here are the authors' personal opinions and do not reflect the views of any of their employers or employees.
(2) All companies or assets mentioned by the author in which the author has a personal and/or financial interest are denoted with a *. None of the above constitutes investment advice, and you should seek independent advice before making any investment decisions.
(3) Any companies mentioned are top of mind and used for illustrative purposes only.
(4) A team of researchers has not rigorously fact-checked this. Please don't take it as gospel—strong opinions weakly held
(5) Citations may be missing, and I've done my best to cite, but I will always aim to update and correct the live version where possible. If I cited you and got the referencing wrong, please reach out
