Ethereum's Inflection Point: DATs, CROPS, AI Exploits, and Quantum Threats

Why This Matters

This episode of The Chopping Block arrives at a moment of deep bearishness for Ethereum. ETH/BTC is at multi-year lows, the Ethereum Foundation is undergoing its most significant restructuring in years, and Manuel Araoz — the founder of OpenZeppelin — has publicly declared that "all of DeFi is now unsafe." Into this storm walks Joe Lubin, one of Ethereum's co-founders and the chairman of SharpLink, one of the largest Ether DATs.

1. Saylor's 25 BTC and the Death Spiral Question

The episode opens against a backdrop of market jitters. Bitcoin had been bleeding for a week and a half when news broke that Michael Saylor's Strategy (formerly MicroStrategy) had sold 25 BTC — not a lot of Bitcoin, and sold months ago, but it was the first sale in four years. The market reacted by "legging down 6%."

Haseeb Qureshi framed the psychology succinctly: "The one buyer in the market seems to be capitulating." Saylor's company owes roughly $1.4 billion per year in cash dividends from its preferred stock structure (STRK/STF-C), with only about $700 million in cash on hand. The gap must be closed somehow — through new share issuance that dilutes existing shareholders, or through selling BTC.

Tarun Chitra was characteristically blunt about the structure: "If I have 10 layers of preferreds on top of a common stock that I'm dumping constantly — why even have that as a vehicle to hold tokens? It does feel a little like an algorithmic stablecoin with too many steps." He called it an "unstable equilibrium" and noted that "it's an event whose probability seems higher the lower the common stock goes."

Joe Lubin was unbothered: "It's way overblown. They have so much Bitcoin and they had to implement the strategy. I'm not worried at all about Michael Saylor and strategy."

2. Ether DATs: Productive Capital vs. Leverage Trap

The conversation pivoted to a structural comparison. Bitcoin DATs like Strategy's are built on an unproductive asset that must sell itself to service obligations. Ether DATs — like SharpLink and BMNR — are fundamentally different.

"Our token Ether yields approximately 3% risk-free yield. We can layer on sleeves on top of that. Virtually all of our tokens have been staked when we buy them. So it's super solid. No leverage, and we represent long-term essentially permanent capital that we can use to grow valuable things in our industry." — Joe Lubin, Chairman of SharpLink

The distinction is crucial: productive vs. unproductive asset, no leverage vs. a convertible-bond tower, permanent capital vs. an unstable equilibrium. In Lubin's framing, Ether DATs are designed not just to hold value but to deploy it into the ecosystem's growth.

3. The Ethereum Foundation at a Crossroads

The Ethereum Foundation's restructuring has been one of the most discussed topics in the ecosystem. Recent departures, internal messaging that "wasn't handled well" (as Bastian admitted), and Aya Miyaguchi's affirmation that the EF mandate may not match community expectations — all point to a transition.

Lubin's inside view: the EF needed to (A) reduce its budget, (B) set itself up for long-term viability, and (C) do what other ecosystems do — but "the right way. The natural way."

"We should be thinking of the EF as one node, a major node, a node that sets the north star. Focus on the cypherpunk vision, focus on crops. Crops are components and these things need to be protected and continually developed." — Joe Lubin, citing Vitalik's recent post

The institutional initiatives that Tomasz built — an interface for companies, governments, and standards bodies — will continue. What's externalizing is biz dev and commercialization, which are moving to organizations like SharpLink, BMNR, and Consensys. Lubin is explicit: "We want to be what the crypto bros in other ecosystems are" — but without the pay-to-play conflicts of interest that would compromise credible neutrality.

"We can't build a credibly neutral platform for the world without having credibly neutral organizations as some of the major nodes in the ecosystem." — Joe Lubin

4. Why Not a Second Foundation?

The concept of a "Second Foundation" has been circulating for years — since at least 2023 with Konstantin. But Lubin pushes back: "Ethereum Foundation just has a very special place. There's only one foundation."

Tarun offered a structural explanation: the EF was "the first of the Stiftung era" — it created the whole model. Every foundation since is a descendant. Copying the same pattern isn't appealing, and it feels like "usurping or copying the original."

Haseeb cut to the core of community frustration: "What the community needs to see is that Ethereum is listening. So much of the frustration from people in the Ethereum outer orbit is that Ethereum sits in an ivory tower, it looks from up high down below, and it's only really going to respond when it sees the untouchables starting to scale the walls."

The resolution: not a second foundation, but a network of nodes — the EF as north star on CROPS and cypherpunk vision, and other organizations handling commercialization, scalability, and UX.

"People seem to want to anoint a monarch to lead us. Vitalik is a leader and a steward and just an incredible asset for the ecosystem and the world, but he doesn't want to dictate. He's always made suggestions. We're moving into maybe an intent or goal-centric decentralized world." — Joe Lubin

5. Quantum Threat: Google's ZK Breadcrumbs

The quantum story of the week is both alarming and darkly ironic. Google published a paper showing they had created a circuit that massively speeds up Shor's algorithm — making it easier to break ECDSA keys with a quantum computer. But instead of publishing the circuit, Google published only a ZK proof that the circuit exists, claiming it was "too dangerous" to reveal.

"It's so important that we can see it but you can't," Haseeb summarized — and that's where things got interesting. The community treated Google's ZK proof as a reinforcement learning environment, crowdsourcing LLM subscriptions to brute-force circuits until they reproduced the same proof.

The result: amateurs improved on Google's results by 12-13% — purely from the breadcrumbs Google left behind. Tarun called this "an ironic form of Proof of Work." The ZK proof designed to protect the secret became the attack vector.

Ethereum's response is explicit and aggressive: migrate to post-quantum crypto by 2029. The Lean VM construction will rip out BLS signatures, KZG aggregation, and ECDSA signatures — all three going "the way of the dodo." This roadmap, Haseeb noted, may partly explain why ETH is outperforming other majors: "ETH/BTC is the quantum index," as Nick Carter once observed.

6. AI-Driven Exploits: "All of DeFi Is Unsafe"

OpenZeppelin founder Manuel Araoz set the crypto world on edge with a tweet: he believes all of DeFi is now unsafe and is recommending family and friends withdraw everything. The reaction was intense — but the panel largely agreed with the underlying concern, if not the absolutism.

Tarun framed the situation as a Stackelberg security game: the defender writes code and chooses parameters, the attacker plays a move, the defender adapts. The question is what equilibrium they reach.

作者概括: Tarun compares the attacker-defender dynamic to training a chess bot — the first five million epochs, you lose almost everything. Then suddenly the system finds what it kept messing up and starts winning. He expects DeFi security to follow a similar but much faster trajectory, with a painful 1-2 year transition period.

The speed asymmetry is the core concern: ASIC-accelerated models now output 10,000 tokens per second. "Multiple days of work happening within minutes. Attackers are happening at AI speed and the defenders are happening at human speed." Defenders will eventually adopt the same tools, but the transition lag is dangerous.

7. Vulnerabilities in the Cracks

"Most of the vulnerabilities in our ecosystem sit either between the chair and the keyboard — dumb people doing dumb things, making dumb choices, writing dumb things into parameter files — or in between protocols." — Joe Lubin

Lubin described Consensys Diligence's approach: they built their own AI with agents that "swarm on protocols." Because Ethereum is composable, the real danger lies in the cracks between protocols — "three subtle things in these three different situations" combining into a sophisticated attack chain. This is where machine intelligence excels: finding patterns across boundaries that human auditors miss.

The silver lining, in Lubin's view, is formal verification. Ethereum is building pipelines to formally verify specifications and implementations: "There's a decent chance that we go in the direction of a golden age for software, where the builders get access to formal verification and software that will help them think through everything."

8. DeFi's Insurance Gap

Thorchain provided a live case study: hacked two weeks prior and still offline — "literally it's like so broken." When Zellic's V12 AI scanner found vulnerabilities and offered to report them, Thorchain refused to pay. V12 disclosed them publicly. It's emblematic of a systemic problem: "We have learned a lot of lessons in the industry but we've not learned them uniformly."

Tarun drew a historical parallel to the late 19th-century railroad boom — the original Great Depression. Private money, rampant scams, companies issuing their own bonds and dollars. "Everything blew up. And afterwards you saw the creation of the largest insurers."

作者概括: Tarun argues that composability is DeFi's insurance problem. In traditional finance, a single entity provides a single point of coverage. In DeFi, every product has a supply chain — bridge, oracle, issuer, RWA — and the insurance sits only at the endpoint. The innovation he predicts: distributing insurance costs across the system rather than fixing them at one place.

9. ZK Bridge-Free Architecture

Joe Lubin outlined what he calls the "extended Metropolis" — Layer 1 plus L2s plus private permissioned networks, connected through zero-knowledge proofs rather than bridges.

Linea can now run ZK proofs across multiple Linea networks and into private permissioned networks. There are roughly 200 Base networks running worldwide, including Citi's Token Services. ZK proofs from these environments can be aggregated into a single proof and settled on a Layer 2 or Layer 1 block.

"We can create these atomic execution zones where we can have a bunch of transactions happen across a bunch of different networks. We can have tokens move across these networks without the need for bridges because you're just doing zero knowledge proofs. That is an essentially bridge-free architecture that removes a lot of the vulnerabilities." — Joe Lubin

Lubin acknowledged that "it's really hard to get inside a zero knowledge prover and hack that thing" — though he added a cautious "we'll see if that's true." The broader point: ZK proofs bring composability back into an ecosystem that intentionally introduced fragmentation through L2s. "You can do real-time unification of different liquidity fragments across these different networks."

10. CFTC Opens US Perps

In a landmark move, the CFTC approved Kalshi's BTC perpetual — the first US-regulated perps product. The door is now open for further reviews, though the CFTC clarified that equities and agricultural products are probably off the table for now.

Hyperliquid pumped on the news, hitting $75 and a fully diluted valuation that flipped Solana, making it the ninth most valuable crypto asset. Tarun predicted Hyperliquid will eventually create a regulated US entity — "Polymarket style" — with an offshore product and a domestic licensed product. Joe Lubin noted the competitive landscape: Lighter (an Ethereum perp), Ostium, Variation, and "a little company called ICE, New York Stock Exchange" are all "tooling up."

Lubin connected this to a broader thesis: "Most of the world is financially illiterate and are regularly led to the slaughter by the financial industry." Perps and prediction markets, as an "entertainment layer," can draw people in, get them asking questions of AI assistants, and eventually build numerical literacy — a pathway to financial empowerment.

11. The Hyperliquid Distribution Flywheel

Tarun identified Hyperliquid's key innovation: "The key thing, the key innovation for Hyperliquid that no one has copied is distribution." Through integrations with MetaMask, Fantom, and HIP3/HIP4 asset listings, Hyperliquid achieved something no L1 has done. An L1 has too many counterparties to coordinate; Hyperliquid found a simpler path.

Perhaps more telling: "The generalized platform part of Hyperliquid is not what's worked. What's worked is the market-specific platform." HyperEVM adoption has lagged, but the constrained platform — "a little spot in our order book, here you go" — attracts enormous entrepreneurial activity. Tarun called it "the bucket shopization of an exchange" — partitioning an exchange into smaller components and selling off pieces to grow faster.

The result is a "Polymarket effect" — the number of people betting on Hyperliquid is dwarfed by the number who know about it, look at it, and use its data. Free marketing at a scale you can't buy. Tom called it "distribution begets distribution."

Core Quotes

Joe's Closing Vision

Asked to give a reason for optimism, Lubin pointed to convergence: AI and Ethereum are both about empowerment. AI provides "some total knowledge of humanity, best practices, wisdom, access to the domain experts in all the domains in the world, available to virtually everybody." Ethereum provides "economic, social, political, financial agency for everybody who has control of their own assets." Together, they're "a recipe for a much flatter world."

Consensys is in deep discussions with FMIs and major financial institutions. Besu has been contributed to Hyperledger Foundation; Linea to Linux Foundation Decentralized Trust. "A hundred well-announced projects have built businesses on Besu. There's another hundred not so well announced." All of them can be brought "into the Ethereum economy proper" and linked via composability into an extended ecosystem.

And then the teaser: MetaMask has an announcement coming "within a small number of days" — something that will facilitate the hybrid human-machine economy Lubin describes. The agents, he insists, are coming.