Palantir’s CEO said it out loud: enterprises are paying frontier labs for tokens while handing over their alpha for free. For individuals it’s worse — ChatGPT logs are already court evidence, and half of users don’t know their chats can be subpoenaed. Then Bridgewater ran the experiment everyone should study: fine-tune an open model on your own experts’ labels. Result: beat frontier models with 30% fewer errors at 1/13th the inference cost. That’s the real signal in our new private AI report — the alpha that separates you from competitors is exactly what should never hit a model provider’s servers in plaintext. Trust or verify? The report ends with a practical framework for choosing.

Quantum won’t kill crypto overnight. It’s a migration countdown, not a doomsday weapon. The hard part isn’t the math (post-quantum crypto is already standardized) it’s the coordination. And BTC vs ETH face opposite problems: Bitcoin’s is political (should we change the untouchable coins?), Ethereum’s is engineering (how do we rebuild the whole stack without breaking it?).

You can buy NVIDIA with USDT on a CEX now. But do you actually own the stock? We audited the plumbing behind tokenized US stocks. 94% of clearing runs through a single broker (Alpaca). Voting rights evaporate, dividends become contractual IOUs, SIPC protection stops at the issuer. What you hold isn’t equity. It’s a digital IOU tracking a price. New IOSG research 🧵👇

Ethereum doesn’t need another act of technical faith. It needs a Musk-style compromise with commercial reality. Everyone’s first reaction to ETHLabs was “2nd foundation.” Wrong question. Five senior EF researchers, funded by BitMine (largest corporate ETH treasury, 5M+ ETH), SharpLink and Lubin, didn’t fork the foundation. They filled a vacuum the EF deliberately left open. And Vitalik didn’t make this move. The ecosystem made it for him. An organization forming around the founder is itself the market’s no-confidence vote on “govern by not governing, stay small.” Not a forum poll. A capital vote. Musk’s edge was never the rocket. He understood how business works, where money comes from, what users want, then did the dirtiest work himself, and delayed Mars to win the Moon. Real world first, then made the tech reach for it. Vitalik runs the other way: he starts from the purest tech and values, and waits for reality to grow on its own. That worked for ten years because there was no other option. ICOs, DeFi, NFTs all emerged from the community itself. Ethereum got lucky every cycle. But there are far more options now, and luck doesn’t stay on your side forever. Two camps: one says Ethereum lacks a Starlink-grade killer app, the other says it lacks BD and a willingness to compromise with commercial reality. These don’t contradict. What Ethereum lacks isn’t another roadmap. It’s someone willing to get in the arena, understand the business, and grind against real-world applications. So my expectation of Vitalik is specific. Not a better whitepaper. That he studies how Musk built companies and how to read a business, then pulls his energy back to one question: what can Ethereum actually run in the real world? Until that’s clear, every technical narrative is suspended in mid-air. The EF’s real problem came from inside. Dankrad Feist: the people leaving are CROPS believers, the problem isn’t strategy, it’s management, the exodus is genuinely bearish. The cure isn’t a new direction, it’s the organization itself. And an org disease can’t be cured by a founder who stands half a step removed and won’t get his hands dirty. The new model is a smaller EF plus independent steward nodes, ETHLabs and Etherealize each making their own calls. This validates a direction I’ve argued for a while: the single-foundation model can’t hold. But their path is far more radical, and harder. Multiple nodes deciding on their own creates a coordination problem money can’t buy. Distributing the checkbook is easy. Distributing direction without fragmenting is the real challenge. My answer: cohesion can only come from ETH as a shared reference asset, not from any organization or person. Which is why Vitalik’s May post said the foundation was never Ethereum’s center, just one node holding 0.16% of ETH, and defending ETH’s value needs the larger holders to step up. I agree. But capital alone doesn’t bind cohesion. The precondition is a real-world narrative everyone can understand and bet on together. And the independence, right now, is a claim, not a fact. Lubin is both ConsenSys CEO and SharpLink chairman, so funder and beneficiary are the same people, and the research is bankrolled by parties holding billions in directional ETH. The structure is reasonable, but trust is earned over years, not in a press release. Last, the deepest layer. The roadmap is still stuck in infra: MEV elimination, default privacy, ETH Pay. All necessary, none of it answers the real threat. Ethereum’s competitor was never Solana. It’s attention migrating to AI. That window is maybe 12 to 18 months. Infra doesn’t win attention back. A founder willing to get in the dirt like Musk, and a ten-year story strong enough to make a top graduate walk away from Anthropic. Vitalik’s idealism hasn’t gone dark. But for that light to reach the ground, what’s needed isn’t another look at the stars. It’s a willingness to bend down and get in. And time is the only opponent that doesn’t negotiate.

Conviction that shows up early and stays. IOSG backed Centrifuge in 2021, when RWAs onchain looked too early to most, and has only added since. Now we build the next chapter together across Asia, where institutions are looking for a proven path onchain.

- Coinbase chooses Centrifuge as tokenization partner - Ethena chooses Centrifuge as tokenization partner - IOSG chooses Centrifuge as tokenization partner - OKX chooses Centrifuge as tokenization partner ... who's next?

Five years ago, @IOSGVC was the first institution to back @Centrifuge, and one of the few willing to touch real-world assets at all. Bringing RWAs onchain looked too early, and too unclear, to most people at the time. I’ve watched a lot of projects rise and fall since. Centrifuge is one of the few that truly endured. Today it has grown into one of the leading RWA tokenization and issuance platforms out of the US, licensed by S&P to issue onchain index funds and now expanding into onchain equities. We first invested in 2021, added to our position in 2024, and have continued to increase it since. Not because of a narrative cycle, but because we have always believed Centrifuge would become core infrastructure for institutional onchain finance. This time, we are bringing it to Asia together, and to the traditional financial institutions across the region that are increasingly looking for a credible, compliant path onchain. Grateful that IOSG gets to keep building something long-term, alongside a team worth betting on.

Quote-tweeting the IOSG team because I want to say something personal about this. Most VC events I’ve sat through over 8 years in this industry are noise. This one wasn’t. 16 founders. 1 legislator. 6 hours straight. The thing that surprised me most wasn’t what got said in front of @DuncanChiu — it was what founders said to each other afterwards, over dinner. Hong Kong’s Web3 story for 2026-2032 is being written right now. If you’re a founder thinking about HK and want a real conversation (not a marketing pitch), we’re doing this again. DM me.

Clarification on the rumor circulating about IOSG Ventures selling UNI / COMP on Binance. The wallet cited is not an IOSG address. The Arkham label is incorrect, and Arkham itself marks both the "IOSG Ventures" and "Binance Deposit Change Address" tags with a question mark (?), meaning unverified. We have not sold any UNI or COMP positions recently. 关于近期市场传言 IOSG Ventures 向 Binance 存入 UNI / COMP 的澄清：相关钱包并非 IOSG 地址，Arkham 标签错误，且平台自身已对 "IOSG Ventures" 与 "Binance Deposit Change Address" 标签均标注 "?"，即未经确认。IOSG 近期未持有和卖出任何 UNI / COMP 仓位。 Official IOSG statement to follow.

𝗧𝗵𝗲 𝗺𝗼𝘀𝘁 𝗱𝗮𝗻𝗴𝗲𝗿𝗼𝘂𝘀 𝗽𝗿𝗶𝘀𝗼𝗻𝗲𝗿'𝘀 𝗱𝗶𝗹𝗲𝗺𝗺𝗮 𝗶𝗻 𝗗𝗲𝗙𝗶 𝗵𝗶𝘀𝘁𝗼𝗿𝘆. $230M in bad debt, unresolved. Aave's Collector holds over $200M in liquid assets. LayerZero Labs has raised $300M+ at a $3B valuation, backed by a16z, Sequoia, and Tether. Both are fully capable of covering this. Yet neither has responded. Neither has made a statement. They are playing game theory against each other while the entire DeFi ecosystem bleeds out. Over a $200M dispute, Aave has already lost $8.45B in TVL. The broader DeFi market shed $13.2B in 48 hours. Whales pulled $6B in 24 hours. ETH and stablecoin pools hit 100% utilization, trapping depositors' funds. The cost of inaction has already dwarfed the bad debt itself. 𝗪𝗵𝗲𝗿𝗲 𝗶𝘀 𝘁𝗵𝗲 𝗗𝗲𝗙𝗶 𝘀𝗽𝗶𝗿𝗶𝘁 𝗼𝗳 𝟮𝟬𝟮𝟬? On March 12, 2020, ETH crashed to $80. On-chain liquidation auctions failed, with ETH briefly pricing at zero. MakerDAO faced systemic collapse. The MKR Foundation stepped up, proposed auctioning MKR to buy back ETH and cover the shortfall. Community whales showed up to bid. That moment — people taking responsibility, taking action in the face of crisis — made every one of us proud to be part of Ethereum and DeFi. Today? 𝗦𝗶𝗹𝗲𝗻𝗰𝗲. Many of my peers aren't just pulling funds from Aave — they're withdrawing from Spark, from every DeFi protocol they can. They're not panicking. They're voting with their feet: if this ecosystem can't even resolve a $200M problem without someone stepping up, why would anyone keep capital here? 𝗧𝗵𝗲𝘀𝗲 𝗳𝘂𝗻𝗱𝘀, 𝗼𝗻𝗰𝗲 𝗴𝗼𝗻𝗲, 𝘄𝗶𝗹𝗹 𝗻𝗲𝘃𝗲𝗿 𝗰𝗼𝗺𝗲 𝗯𝗮𝗰𝗸. Aave has been through many governance crises. The community has never been short of disagreements. But today is not the day for debating governance philosophy. Look back at DeFi Summer. Look at what we built together. We didn't come this far to let it collapse in silence. The future of DeFi should not be destroyed by inaction. This is not Aave's problem alone. Spark, MakerDAO, every DeFi protocol on Ethereum should be part of the coordination. Trust collapse doesn't respect protocol boundaries — if this isn't resolved, all of DeFi gets repriced. Everyone loses. 𝗧𝗶𝗺𝗲 𝗶𝘀 𝗿𝘂𝗻𝗻𝗶𝗻𝗴 𝗼𝘂𝘁. Aave can commit to backstopping now and coordinate the specifics later — that alone would stop the bank run. If project teams remain silent, @VitalikButerin should step in to coordinate. Not to pay — just one sentence: "This will be resolved." That's enough to stop the bleeding. Every hour of silence means more capital permanently lost. I'm calling on more DeFi founders and builders to speak up. Set aside the disagreements. Come together and solve this — now. This isn't about the survival of one protocol. This is about the future of DeFi as an industry. @StaniKulechov @VitalikButerin @AaveDAO @KelpDAO @LayerZero_Core @RuneKek — communicate publicly. Give the market a clear signal. 𝗦𝗶𝗹𝗲𝗻𝗰𝗲 𝗶𝘀 𝘁𝗵𝗲 𝘄𝗼𝗿𝘀𝘁 𝗼𝗽𝘁𝗶𝗼𝗻.

Reading Vitalik's AI research, he's arguably one of the most AI-literate people in the entire blockchain industry. So the question is: Can Ethereum capture the AI opportunity? The Ethereum Foundation's answer: In September 2025, they launched the dAI Team with one mission — make Ethereum the base layer for the AI economy. Two core focuses: Enable AI agents to transact and coordinate on-chain, without intermediaries Build a decentralized AI stack to prevent Big Tech monopolization

After 9 years in crypto VC, we’re more bullish than ever. 2026 marks a structural shift: stablecoins surpassed Visa at $18.4T volume, prediction markets proved real utility to mainstream, institutions are finally building on-chain.