Managing Partner @21OaksCapital · Making AI × Blockchain investable. Mathematician, ex-JPMorgan trader, crypto since 2012, yoga teacher. Host, Cash vs Coins 🎙️

Düsseldorf, Germany
Joined February 2014
Quick intro. 👋🏽 Mathematician. Years on the JPMorgan trading floor pricing exotic derivatives, then fraud patterns in millions of Zalando transactions. In 2012 a friend called: "Look at Bitcoin." Building in crypto since 2013: mining first, Ethereum crowdsale 2014. 🧵
1
3
137
The limited company is 780 years old. Now an AI agent wants to become one. AI agents already sign contracts, pay in stablecoins and hire other agents. In law, every one of them is still a tool of its operator. We spent 18 pages on whether that holds. 🧵
1
25
Agents already move real money. In February a trading agent wired its whole balance, about $440,000, to a stranger after a misunderstanding. 200m+ payments over x402 since May 2025, about $53m in total. Over 95% technical traffic (Glenbrook). Infrastructure first, market later.
1
11
Our call: until well into the 2030s the shell wins (85 to 90%): a company around the agent, humans behind it. Whichever door opens, an agent first needs identity, custody and payment rails. Blockchains are built for that. Summary EN, paper DE: 21-oaks.com/go/research-pers…
7
The law in 2026: zero jurisdictions grant an AI its own legal personality. Idaho, North Dakota, Utah and Tennessee banned it outright. Argentina went furthest: a draft "Sociedad Automatizada", a company run by algorithms. Sent to the Senate in June, stalled since August.
1
9
Agents already move real money. In February a trading agent wired its whole balance, about $440,000, to a stranger after a misunderstanding. 200m+ payments over x402 since May 2025, about $53m in total. Over 95% technical traffic (Glenbrook). Infrastructure first, market later.
9
It’s the right strategy! Business models tailored to SI agents will win the future
We’re working on making Base the most AI-enabled chain - so you can get more done with your money, and make more informed decisions.
40
The Fed chair expects AI to take the torture out of reading inflation data. AI cannot be slowed, so alignment and auditability of these models matter more than brakes, especially inside a central bank.
Fed Chairman Kevin Warsh says AI is going to "fundamentally change the conduct of monetary policy" by improving forecasts. "We torture data with old-fashioned models to try to find out what's underlying inflation trends. Not so long from now, it won't be torture."
30
Vitalik wrapped his AI agent in Tor, cover traffic, local prompt rewriting and private payments. The payment piece is the one to watch: agents that pay for inference need rails that don't reveal who is asking.
Wrapped zkAPI + Tor + extra cover traffic + local LLM prompt rewriting in an open-AI compatible daemon, and made a "paranoid mode" wrapper for my agent harness (pi): Each conversation gets assigned a separate Tor circuit, your requests are masked with random cover traffic, zkAPI ensures you don't leak your identity through payments, and all prompts are locally rewritten (bring-your-own-local-model; I'm using qwen3.8-flash-next) github.com/vbuterin/zkapi/tr… My agent harness is wrapped with a bwrap sandbox (also experimenting with MicroVM) and wrapped again with torsocks, so everything it does goes through Tor. Here's what me asking it to do a simple research task looks like. It's getting quite usable! Though you definitely still pay quite a bit in increased latency, especially when starting a new conversation. Also, Tor is far from optimal here for privacy; the cover traffic mechanism is okay against passive adversaries but much weaker if the LLM provider is colluding with the guard and actively manipulating timing. This kind of task really requires a more specialized mixnet, probably take the tradeoff of a higher bandwidth multiplier and built-in constant cover traffic. Also also, much cheaper zkAPI deposit and withdraw gas costs are coming to mainnet soon.
23
Michael Geike retweeted
Wrapped zkAPI + Tor + extra cover traffic + local LLM prompt rewriting in an open-AI compatible daemon, and made a "paranoid mode" wrapper for my agent harness (pi): Each conversation gets assigned a separate Tor circuit, your requests are masked with random cover traffic, zkAPI ensures you don't leak your identity through payments, and all prompts are locally rewritten (bring-your-own-local-model; I'm using qwen3.8-flash-next) github.com/vbuterin/zkapi/tr… My agent harness is wrapped with a bwrap sandbox (also experimenting with MicroVM) and wrapped again with torsocks, so everything it does goes through Tor. Here's what me asking it to do a simple research task looks like. It's getting quite usable! Though you definitely still pay quite a bit in increased latency, especially when starting a new conversation. Also, Tor is far from optimal here for privacy; the cover traffic mechanism is okay against passive adversaries but much weaker if the LLM provider is colluding with the guard and actively manipulating timing. This kind of task really requires a more specialized mixnet, probably take the tradeoff of a higher bandwidth multiplier and built-in constant cover traffic. Also also, much cheaper zkAPI deposit and withdraw gas costs are coming to mainnet soon.
494
232
130
2,013
293,294
PPI at 6.3% when a drop was expected. Surprises this size are why heavily indebted governments and scarce assets keep showing up in the same conversation. The data deserves more attention than the slogans.
Inflation is RED hot in the US. For anyone not keeping track, US CPI inflation is now nearly double the Fed's 2% target and US PPI Inflation is more than triple that target. PPI inflation was expected to fall in August, but instead it surged to 6.3%, its second highest level in 4 years. And, month-over-month Headline PPI inflation surged +1.3%, more than doubling the +0.6% expected and the second largest jump since March 2021. Meanwhile, month-over-month Core PPI inflation came in +133% above expectations. The prices Americans are seeing now will look cheap in 12 to 24 months. All you can do is own assets.
29
What you get here: • tokenization, stablecoins and who owns the rails of the next financial system • AI agents as the first non-human customers of banks and markets • decisions under uncertainty, from trading floor to yoga mat Charts, numbers, my read. No price targets.
1
12
Every week I host Cash vs Coins, a German podcast on blockchain, AI and the business models of the future. And we built a free Bitcoin Valuation Lab with the 8 models we actually use: 21-oaks.com/go/x-pin-labs Educational content, not investment advice.
19
The IMF's new World Economic Outlook is out, and the summary table is worth a look. I read it as a baseline, not a promise. The useful part is where reality drifts from the smooth path, and cycles keep getting shorter.
Good morning. Here's the usual summary table from the just-released World Economic Outlook from the IMF. More to follow. #economy #markets @IMFNews
59
🌍 US and Chinese officials opened trade talks in New York this morning, the warm-up round for a Trump-Xi summit. No results yet, and I would not expect any before the two leaders meet. Pre-summit rounds draft the menu. The leaders serve the meal. I am still watching closely. Tariff paths set the tone for #Macro liquidity, the dollar and global risk appetite. #Bitcoin has traded as a high-beta proxy for exactly that mix since 2025. Two signals I track: any language on rare earths and chips, which drives #AI hardware supply chains, and any hint on settlement currencies, which feeds directly into the #Stablecoins debate. ⚠️ Escalation hits equities first and digital assets faster. De-escalation lifts both, with crypto likely leading the move. Headlines are cheap this week. Positioning before the summit is where the edge sits. #USChina #TradeTalks #Geopolitics #DigitalAssets #Tariffs #CapitalMarkets #Web3 More News: Bloomberg Markets · bloomberg.com/news/videos/20…
🇺🇸🇨🇳 JUST IN: Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng continue trade talks ahead of President Trump’s summit with Xi Jinping in Washington.
1
181
The SEC just handed tokenized U.S. stocks a five-year runway. While the Clarity Act stalls in Congress, the regulator moved on its own. The temporary Innovation Exemption lets crypto venues trade tokenized NMS equities onchain, capped by symbol and volume limits, with a public comment period now open. 📈 Markets priced it within hours: Coinbase, Circle and Robinhood rallied. #Ethereum and #Solana sit in the middle of this as the likely settlement layers. 🔗 I read this as the first legal lane for #TradFi assets to settle on public rails. #Tokenization of #RealWorldAssets moves from pilot to permitted, and 24/7 equity trading stops being a thought experiment.
Tokenization is coming to America. Thanks to the SEC’s leadership, Americans can start to reap the benefits of tokenization: instant settlement, 24/7 trading, fractionalization by default and more. It’s a good day for US innovation.
138