Eddy Deep Dive

Eddy Lazarin, recently elevated to GP at a16z crypto, operates within a single-trigger investment process where one GP can unilaterally commit to a deal without group consensus or a veto mechanism. He applies the Peter Thiel framework that good non-consensus ideas initially look like bad ideas, while noting that non-consensus alone does not imply quality since most non-consensus things are correctly identified as bad.

Eddy grounds his investment thinking in a 2004 William Nordhaus paper finding that 97.8 percent of value created by companies leaks out rather than being captured as profit. Competition drives this leakage by forcing prices down until consumers absorb the surplus, and every price posted is effectively an invitation for competitors to enter. He frames capitalism as a ship sinking but being built faster than it sinks, and argues that VCs obsess over value capture not out of greed but because genuine capture is rare. Value capture is visible on accounting sheets while consumer surplus is nearly impossible to measure, and the entrepreneur's core struggle is retaining enough value inside the business to keep feeding the machine.

Eddy argues that crypto's binding constraints have shifted from technical to regulatory and product-design problems. Blockchain throughput has improved radically, with ordering capacity suspected to reach hundreds of megabytes per second, and zero knowledge proof applications have improved by orders of magnitude. The audience that wanted to own crypto assets already has what it needs, and the goal now is reaching mainstream users and modernizing how normal people and institutions handle finance and capital formation. He attributes the shortfall in innovation after technical constraints lifted partly to regulatory uncertainty and partly to a lack of imagination among builders.

The CLARITY Act, over 300 pages, distinguishes network tokens from company tokens. A network token is the natural ownership asset for a decentralized marketplace where value is captured and distributed automatically, with blockchains described as decentralized marketplaces for compute and verified storage, and projects like Uniswap and Sky as decentralized marketplaces for liquidity and borrowing. The defining dimension is whether a single controller exists rather than whether people are contributing ongoing efforts. Crypto founders previously avoided building real token business models because doing so risked SEC security classification, and revenue was stigmatized in some circles, though Eddy notes that deferring revenue forever may indicate a lack of pricing power rather than a strategic growth choice. If CLARITY passes, Eddy predicts significant experimentation in fully automating things that could not previously be automated.

Tokens introduce monetary policy capabilities unavailable to conventional companies. Buy and burn takes in revenue and retires tokens to reduce supply and potentially drive price appreciation. Mint and spend issues tokens to pay for growth, decoupled from incoming fees. Burning and minting can be fully decoupled from each other in a way stocks cannot replicate, and token networks can mint directly for growth spending whereas conventional companies must sell equity in private rounds to raise cash. One a16z GP expressed skepticism about buy and burn while Eddy is a proponent, and he acknowledges some of his token economic arguments may not hold up once real market experimentation is possible under clearer regulation.

Stablecoins are described as crypto's first killer app and now a consensus topic across conventional fintech. The GENIUS Act created conditions for financial institutions to connect real capital to crypto via stablecoins, and CLARITY is expected to do the same for capital formation around crypto-native assets. Stablecoins solve one specific piece of the payment chain, moving a balance between addresses quickly and cheaply, but do not fix the entire money-movement chain end to end. Adding stablecoin capability to a regular bank enables second-order improvements including visible collateral, lending, and cross-bank interoperability without bespoke integrations. Existing financial intermediaries are optimized for compliance rather than speed and cost, and the shift to stablecoin infrastructure cannot happen in one shot but requires many individual pieces to reconfigure.

Eddy sees crypto and AI as complementary rather than competing. AI agents will require cryptographic constraints covering identity, authentication, money via blockchains, and verification of compute and programs to ensure they perform what users actually want. He describes the current AI phase as an attract phase where users spend roughly five thousand dollars a day in compute costs while paying only two hundred dollars a month, and argues the extract phase that follows is where crypto becomes essential. He sees no compelling story for how AI agent infrastructure plays out without crypto, though he acknowledges conventional systems are technically possible. He predicts AI agents managing personal finances will become standard as people trust the diligence of capable AI, and that improvements in automation will accelerate commerce and capital management broadly.