Most important take away

Crypto has shifted from an ideological revolution aimed at replacing the financial system to a pragmatic infrastructure layer that integrates with it, with stablecoins, on-chain finance, and tokenized markets driving real mainstream adoption. The most successful founders in this next era will be product-focused, go-to-market-focused, and pragmatic rather than ideological — and the convergence with AI (especially agents needing programmable money) is creating the next major opportunity.

Summary

The a16z crypto GPs frame Fund 5 ($2.2B) around a clear thesis: regulatory clarity (the Genius Act for stablecoins, the pending Clarity Act for the rest) plus real product-market fit in financial use cases has put crypto at an inflection point, even while sentiment and prices are down. They view this combination — strong fundamentals, capital rotating to AI, and clear builder pathways — as a classic environment for a strong vintage.

Key themes and actionable insights:

• Pragmatism beats ideology. Ali Yahya and Guy Wuollet explicitly say the cypherpunk “code is law, replace the system” era is over. Crypto has “won the revolution” and is now writing its constitution by working with banks, regulators, and Wall Street. Career advice for founders: be product- and go-to-market-focused, not ideological.
• Stablecoins are the wedge. ~$300B issued, transaction volume rivaling Visa, growth decoupled from trading volume. Stripe expanded from dozens to 100+ countries overnight via stablecoins. Lending protocols on top of stablecoins are the obvious next “verb.”
• On-chain finance is real. Perps have moved beyond crypto tokens to equities, commodities, FX. New markets are forming natively on-chain for compute (GPUs), energy (solar, batteries), and even oil price discovery (hyperliquid). Default for new markets is increasingly on-chain.
• Business strategy — go-to-market is the new moat. Guy: five years ago the highest-status role in crypto was researcher; now it’s go-to-market. “You cannot vibe-code USDC or Hyperliquid.” Network-effect coordination businesses are defensible against AI commoditization in a way that pure software is not.
• Privacy as the only durable moat (Ali). Public blockchains make state trivially forkable/migratable. Encrypted state raises switching costs dramatically. Privacy is also a hard requirement for institutional and mainstream adoption (no one wants their salary public). Approaches range from trusted central operators (Tempo, Canton) to TEEs to ZK cryptography (Jolt). ZK has improved 10–100x in a decade and breaks the scalability trilemma.
• AI x crypto convergence. The traditional financial system was not built for agents. Agents have no brand loyalty, will route around Visa’s 16 bps, and prefer pay-per-use over subscriptions — all of which favors stablecoins and programmable money. Eddy: AI now lets non-engineers control programmable money via natural language. Crypto also provides proof-of-personhood (e.g., World) to fight deepfakes, and a coordination layer for crowdsourced compute/data to counter AI’s centralization.
• Career advice for founders. Don’t treat AI vs. crypto as either/or. Crypto is currently overlooked, which is exactly when talented founders should look at it. Build network-effect/coordination businesses that survive model commoditization. Focus on financial use cases first because the bar is lower than competing with mature global software systems.
• Fund 5 success metrics. The GPs converge on: 1B+ daily blockchain users (mostly indirectly via stablecoin neo-banking, payments, tokenized stocks/bonds), majority of world finance moved on-chain, AI agents elevated from tools to first-class economic actors, and on-chain capital markets for compute and energy.

Chapter Summaries

Why raise now. Chris Dixon argues the setup is classic venture: down sentiment but strong fundamentals — $300B in stablecoins, the Genius Act giving stablecoins a regulatory framework, the Clarity Act expected to do the same for the other 90% of crypto, and Wall Street tokenizing stocks/bonds. Capital has rotated to AI, leaving more room.

Revolution to governance. Ali Yahya recounts crypto’s cultural shift since 2017 from cypherpunk “replace the system” to pragmatic integration. Guy Wuollet calls it the “colored shirt era” — crypto won the revolution and is now writing its constitution, taking serious meetings with major banks.

The pragmatist turn. Eddy Lazzarin reframes the shift not as capitulation but as an extension of what’s possible — he describes using AI to one-shot a CLI tool that sent Zcash to Coinbase, simultaneously cypherpunk and trad-compatible.

On-chain finance and new markets. Guy details why credit is moving on-chain (stablecoin yield demand, post-GFC private credit duration mismatches, double-pledging risk) and how perps have generalized beyond crypto to equities, FX, commodities. New markets — compute, energy, even oil price discovery — are being born on-chain.

Counterparty risk = decentralization, rebranded. Trad finance values latency, 24/7 markets, capital mobility, and counterparty/platform risk reduction. The marketing term “decentralization” maps to concrete trust assumptions institutions actually care about.

Finance as foundation for the broader vision. Chris connects this to Read Write Own: get a billion people onto blockchain rails via finance first, then adjacent services follow naturally. Finance is the lowest-hanging fruit because trad systems are weakest globally.

AI x crypto convergence. Ali’s anecdote about being laughed out of Google X for proposing crypto in 2016 frames the historical opposition. Now agents need programmable money — stablecoins are the only payment rail built for them. Visa’s 16 bps is vulnerable because agents have no brand preference. Crypto also enables proof-of-personhood against deepfakes.

The agentic future. A sci-fi-but-plausible vision: autonomous AI agents with crypto wallets paying for their own compute, generating value, hiring other agents — fully self-sustaining economic entities within five years.

LP feedback and the GTM moat. Guy reports LPs are focused on how non-AI software competes in an AI world. Crypto’s network-effect/coordination businesses are the answer — you can’t vibe-code USDC or Hyperliquid. GTM, not protocol cleverness, is now the bottleneck.

Privacy as the only moat. Ali argues privacy is essential for mainstream and institutional adoption and is the one feature that creates real switching costs in a world of fungible block space. Approaches range from trusted operators to TEEs to ZK cryptography. Justin Thaler’s Jolt project at a16z is pushing efficient ZK forward.

ZK and the scalability trilemma. ZK proofs let one node do work that everyone else can verify cheaply, breaking the historical limit where every node redoes every transaction. Theoretically unbounds throughput.

Centralization risks and crypto’s role. Chris notes the internet has consolidated dramatically and AI will accelerate that (Stack Overflow traffic collapsing post-LLMs is the canary). Crypto is the only credible counterweight. Guy adds proof-of-personhood, decentralized compute markets, and crowdsourced training data as concrete antidotes.

Fund 5 success criteria. Eddy: mainstream adoption and intensified competition enabling people and machines to own things. Ali: 1B+ daily users, majority of finance on-chain, agents as economic actors. Guy: dollar-denominated stablecoin neo-banking for every human, plus on-chain compute and energy markets. Chris: real mainstream financial use cases driving the path to a billion users.

Read Write Own coda. Chris stands by the book’s thesis — every technology has an essence that determines its long-run trajectory, even if the specific apps and timing are unpredictable.