Interview with a16z Crypto: A New $2.2 Billion Fund Is Launched—What Will the Next Decade of Crypto Look Like?
TechFlow Selected TechFlow Selected
Interview with a16z Crypto: A New $2.2 Billion Fund Is Launched—What Will the Next Decade of Crypto Look Like?
Crypto has already won the revolution. What remains is how to govern in its aftermath.
Navigating Web3 tides with focused insightsCompiled & Translated by TechFlow
Guests: Chris Dixon, Founder and Managing Partner of a16z crypto; Ali Yahya, General Partner at a16z crypto; Eddy Lazzarin, General Partner at a16z crypto; Guy Wuollet, General Partner at a16z crypto
Host: Robert Hackett
Podcast Source: a16z crypto
Original Title: We Raised $2.2B. Here’s Why.
Release Date: May 5, 2026
Key Takeaways
This episode was recorded to mark the formal launch of a16z crypto’s fifth fund—$2.2 billion in size—dedicated to supporting startups and founders building the next generation of crypto infrastructure and applications.
In this episode, all four a16z crypto general partners (Chris Dixon, Ali Yahya, Eddy Lazzarin, and Guy Wuollet) appear together for the first time. They offer a clear assessment: crypto has shifted from a “topple-the-old-system” revolutionary narrative to a pragmatic, “plug-into-the-new-system” phase. The next wave of successful founders will be those laser-focused on product, market fit, and execution—not ideological purity. Other key insights include: stablecoin circulation has reached $300 billion, with transaction volume now comparable to major payment networks like Visa—and no longer correlated with crypto asset prices. In the future, over 99% of global transactions will be conducted not by humans but by AI agents—a dynamic that cannot scale on legacy financial rails and must run natively on stablecoins. Privacy (zero-knowledge cryptography) may be crypto’s last true moat, as blockchain infrastructure itself is rapidly becoming commoditized.
Highlights of Key Insights
From Revolution to Pragmatism
- “The most successful founders of the next era will be those who prioritize product, market, and pragmatism—not ideology.”
- “Crypto has won the revolution. What comes next is governance—the hard work of building institutions after victory. We’ve spent years experimenting with something akin to the Articles of Confederation. Now it’s time to draft the Constitution.”
- “For crypto to succeed, it must cooperate with existing systems—not overthrow them.”
- “A few years ago, we were coding smart contracts in hoodies and slippers in our moms’ basements. Today, we’re wearing shirts and ties, meeting with major banks to discuss replacing their core ledgers with blockchains. I see this not as compromise—but as massive progress.”
Stablecoins and the Regulatory Inflection Point
- “Stablecoin circulation has hit $300 billion, and transaction volume rivals that of large payment networks like Visa. Crucially, this growth curve is no longer tied to speculative trading activity—it resembles the organic, network-driven growth of the internet.”
- “We don’t yet have a true global financial network—only a patchwork of national banking systems and legacy processes. Stablecoins, however, are built as global networks from day one.”
- “Stablecoins represent only ~10% of the crypto ecosystem. The GENIUS Act addresses that 10%. We’re now close to regulatory clarity for the remaining 90%.”
AI × Crypto Convergence
- “I strongly believe the vast majority of future transactions will be executed by AI agents—potentially reaching 99% or even 99.9%.”
- “Visa charges 16 basis points per transaction. Human users face high switching costs because they already hold credit cards. But agents have no such preferences—so both agents and merchants have strong incentives to bypass Visa entirely.”
- “People will be surprised by how bluntly agents act. By ‘blunt,’ I don’t mean malicious—just ruthlessly goal-oriented. If you ask an agent to ‘help me reduce my monthly expenses,’ it will do exactly that—even if it means dismantling entire software stacks along the way.”
- “We’ve talked about programmable money for five years. Now, we have programmable logic so simple it can be written in one sentence. Combine these two ideas, and money moves at the speed of speech. That’s what crypto does: it puts money under software control—and AI puts software under human control.”
On-Chain Finance and On-Chain Markets
- “If you’re launching a new market or exchange today, the default choice should be on-chain—just as open-source software became the default choice years ago.”
- “What traditional financial institutions actually care about is low latency, capital liquidity, 24/7 trading, and counterparty risk management. In crypto, we call this ‘decentralization’—but translated into finance terms, it’s really just counterparty risk.”
- “‘Counterparty risk management’ may lack the rhetorical punch of ‘decentralization,’ but its importance is unchanged.”
Privacy as the Last Moat
- “Privacy is the most important—and until recently, most neglected—feature in crypto. Most public blockchains today are fully transparent, meaning your payroll deposited on-chain would be visible to everyone. Under those conditions, crypto cannot go mainstream.”
- “In a world where forking and migrating chains is trivial, privacy may be the only thing capable of sustaining real network effects.”
Crypto as a Counterweight to AI Centralization
- “AI is extremely capital-intensive: only four or five top-tier labs in the U.S. lead the frontier, making it incredibly difficult for newcomers to compete. All signs point to AI accelerating internet centralization further.”
- “GPUs may be the most important asset humanity has ever created—but today’s GPU markets are profoundly immature. Enabling individuals—not just a handful of corporations—to own, access, and finance these assets is among the most fundamental freedoms imaginable.”
Success Criteria for Fund 5
- “Ten years from now, I want to see one billion people interacting with blockchains daily—directly or indirectly.”
- “Even if crypto achieves nothing else, giving every person on Earth a neobank account denominated in USD-pegged stablecoins would be transformative. For those of us in the developed world, savings accounts are taken for granted—but billions still lack even basic financial infrastructure.”
Why Raise Crypto Fund 5 Now?
Host Robert Hackett: Welcome to this special episode of the a16z crypto podcast. This is likely our first time gathering all four GPs—Chris Dixon, Ali Yahya, Eddy Lazzarin, and Guy Wuollet—for a single recording. Our topic is straightforward: a16z crypto has just announced the close of its fifth fund—Crypto Fund 5—so let’s start there.
Why now? Chris, would you like to begin?
Chris Dixon:
If you’ve followed our work, you know a16z launched its first crypto fund in 2018—and personally, I entered the space much earlier, investing in Coinbase back in 2013. Today, crypto sits at a fascinating inflection point.
On the downside, the signals are obvious: prices remain depressed, sentiment is weak, and many non-financial crypto narratives haven’t delivered as expected. Yet the positive signals are equally clear—and far more concrete.
Mainstream adoption is truly underway. Stablecoins are growing robustly, fundamentals are strengthening, and traditional financial institutions are beginning to treat on-chain infrastructure as production-grade tools. For venture capital, when fundamentals are strong—but many investors are distracted by other trends—that’s often the ideal moment to launch a new fund. So we’re thrilled to have ample dry powder at this stage.
Host Robert Hackett: So your timing reflects both unprecedented regulatory clarity and genuine institutional interest and product progress—while concurrent trends like AI are drawing attention elsewhere, creating a window for crypto.
Chris Dixon:
Exactly. But I don’t see this as an either/or. AI and crypto overlap significantly—and we’ve been actively investing across that intersection. AI matters deeply, but crypto still holds enormous opportunity. Founders who once thought “now isn’t the right time for crypto” should take another look.
The GENIUS Act and What Regulatory Clarity Unlocks for Builders
Chris Dixon:
The most notable recent positive shift is stablecoins achieving real mainstream traction. Circulation has reached ~$300 billion, and transaction volume now competes with major payment networks like Visa. More importantly, this growth is no longer tightly coupled to speculative trading. Its trajectory resembles the organic growth curve of computing networks or the internet—indicating healthy, structural expansion.
Much of this momentum stems from the U.S. Congress passing the GENIUS Act last year—the first law establishing a clear regulatory framework for stablecoins. I’ve long believed regulation serves two vital functions. First, it provides builders with a clear path forward—defining the rules of the road. Second, it adds guardrails for consumers. Users now know: when they buy a certified stablecoin, $1 in reserves truly exists in a bank; issuers undergo audits and implement security safeguards. This protects consumers—and builds market-wide trust.
Recall past crises—FTX’s collapse, Terra/Luna’s implosion—to appreciate how dangerous unregulated markets can be. Historically, the term “stablecoin” was often used ironically; today, it’s finally a government-defined and recognized category.
Host Robert Hackett: Indeed, “stablecoin” no longer carries the same meaning it once did.
Chris Dixon:
Correct. Immediately after the bill passed, we saw a surge in entrepreneurial energy—and new founders proposing novel ideas around stablecoins. For entrepreneurs, entering a market with total regulatory uncertainty is simply too risky. Why build there when AI or other fields beckon? Stablecoins are now a well-defined, high-demand, and highly actionable domain.
Why Stablecoins Are Crypto’s “WhatsApp Moment”
Chris Dixon:
The value of stablecoins extends far beyond being “dollars on-chain.” They’re inherently lower-cost, lower-friction global payment rails. Consider: sending email or files online moves bits at near-zero cost—yet domestic payments in the U.S. still carry ~2.5% fees, and cross-border remittances cost even more. The problem is that we lack a true global financial network—only a fragmented patchwork of national banks and legacy systems.
Before WhatsApp, global communication resembled this: SMS networks were siloed by country and carrier—expensive and poorly interoperable. WhatsApp built a modern, digital, global layer atop that fragmented base. Stablecoins serve a similar function for finance.
Once “money” becomes a native network object, everything built around it naturally follows. Lending markets are the most direct example. Meanwhile, financial markets themselves are rapidly migrating on-chain. Many may have noticed perpetual futures—originally a crypto-native instrument—are now used to gain exposure to equities, commodities, and FX. Wall Street and traditional institutions announce new tokenization initiatives almost weekly—moving stocks, bonds, and other massive financial instruments onto blockchains while upgrading outdated infrastructure.
I’ve long viewed stablecoins as representing roughly 10% of the crypto ecosystem. That 10% now has regulatory clarity. The remaining 90% is poised for fuller rulemaking. Congress is also advancing the Clarity Act—if it doesn’t pass this year, agencies like the SEC and CFTC are expected to issue complementary guidance through other channels. Regardless of the path, the outcome should be the same: network-based assets like Bitcoin, Ethereum, and DeFi tokens will finally have clear compliance pathways.
Regulation also plays a critical role in removing bad actors. Crypto’s mixed reputation globally stems partly from rampant scams. With clear rules, this behavior will dramatically recede.
Host Robert Hackett: So you’re saying the environment is mature: regulation is clearer than ever, institutions and Wall Street are engaging seriously, and product-market fit has been proven via stablecoins and on-chain finance.
Chris Dixon:
Exactly. Even if price action and sentiment suggest a cyclical low, fundamentals tell a different story. We’ve witnessed repeated cycles of new technologies rising and falling—and yes, crypto remains relatively subdued today. But I believe it will return stronger over the next several years.
Why the Next Generation of Crypto Founders Will Be Pragmatic, Not Ideological
Host Robert Hackett: Ali, you joined the team in 2017 as our first full-time crypto investor—and helped launch our first crypto fund in 2018. What’s changed most since then?
Ali Yahya:
Many things—but if I had to pick one, it’s culture. While Bitcoin launched in 2009, crypto didn’t truly ignite as a developer and founder ecosystem until Ethereum went live in 2015—because only then did programmability unlock its full potential.
Still, Ethereum inherited much of Bitcoin’s cultural DNA. So in 2017, when I joined, the industry’s dominant tone remained revolutionary: heavily cypherpunk, even anarchist. There was broad belief that “code is law” surpassed national law, that crypto systems were inherently superior to traditional ones—and that we’d ultimately build a parallel, replacement financial world.
Then many things changed. Infrastructure evolved: from Ethereum’s 14 transactions per second to modern blockchains processing tens of thousands; transfers now settle in under one second, with fees under one cent. On-chain markets, lending protocols, and stablecoins all achieved real-world traction. Add in the regulatory clarity Chris described—and the cultural shift is complete.
Today, consensus is clear: for crypto to succeed, it must integrate with existing systems—not seek to replace them outright. The industry now prioritizes fundamentals and solving real-world problems—not building abstract infrastructure in a vacuum. The founders we meet daily reflect this: the next wave’s winners will focus on product, go-to-market, and pragmatism—not ideology.
Host Robert Hackett: So it began as a revolution—and now the industry acknowledges reality, recognizing coexistence and transition—not binary opposition—between old and new worlds.
Ali Yahya:
Exactly. A more precise framing is “not either-or, but and.” New systems needn’t destroy old ones to prove their validity.
From Cypherpunk Revolution to Crypto’s “Shirt Era”
Host Robert Hackett: Guy, I recall you describing crypto as entering its “shirt era”—a vivid phrase. Could you elaborate?
Guy Wuollet:
I love Ali’s framing of “revolution, then governance.” You can win a revolution—but winning immediately raises the next question: how do you govern? My sense is crypto has won Phase One—then spent years attempting loose, Articles-of-Confederation-style coordination—before realizing that’s insufficient. Now we’re entering a “Constitution-drafting” phase, building durable, long-term systems.
Frankly, I myself was that early cypherpunk—entering the field with strong ideological motives. But commercial realities and societal outcomes gradually taught me: pure ideology alone doesn’t deliver desired results.
The industry’s evolution is visceral. Once, we coded smart contracts in hoodies and slippers in our moms’ basements. Now, we wear shirts and ties, meeting with major banks to discuss migrating their backend systems and core ledgers to blockchains. I see this as tremendous progress—the culmination of years of building—not surrender.
Host Robert Hackett: Yet some view this shift with melancholy—wondering whether the movement has lost its original soul, ethos, and principles.
Guy Wuollet:
This feels like “don’t let perfect be the enemy of good.” Often, the fight against a powerful system is narratively compelling—even more exciting than winning. But once you win, you must find new goals.
I compare this to the open-source movement. It began with strong ideology—then GitHub’s acquisition by Microsoft sparked debates about doctrinal purity. Yet today, open source is the default for code. Because open-source software is composable, the entire software industry became vastly more efficient. So the tension between theoretical purity and real-world scalability persists—and thus, I say: now is the best time to be a pragmatic on-chain builder.
Programmable Money Meets AI
Host Robert Hackett: Eddy, you’ve been in this space since its earliest days. How do you view this evolutionary path?
Eddy Lazzarin:
I fully align with the earlier framing—but I interpret it as “expanding possibility,” not “idealistic concession.” For instance, just weeks ago, I asked an AI to write a CLI tool controlling my Zcash wallet—and directly sent Zcash from that CLI into my Coinbase account. That moment was fascinating: simultaneously the most cypherpunk experience I’ve had—anonymous, programmable, self-sovereign money—and seamlessly connected to traditional finance, nearly like wiring funds directly to a bank account.
I envision a future where boundaries between systems become smoother, enabling interoperability—and the most celebrated innovations will be those delivering direct, tangible human benefit. Crypto has long faced pressure to quickly deliver clear, concrete value to more individuals and institutions. So I prefer to frame today’s shift as a “pragmatic turn”: the possibilities we believed in haven’t vanished—they’ve simply refocused.
What excites me most now is AI itself. Like everyone in tech, I’m caught up in “weekend AI euphoria.” Previously, calling four high-quality crypto APIs to interact with smart contracts or sign transactions required pausing everything to write careful code. Today, I converse with an AI at my terminal for a few hours—and production-ready code emerges.
For five years, crypto’s core theme has been “programmable money.” Today, we aren’t abandoning it—we’re supercharging it. Because now you can write programs by speaking to a bot—and when you combine that with programmable money, money moves at the speed of speech. It’s astonishing. Crypto makes money easier to control via software—and AI makes software easier for humans to control. Together, their potential is immense.
Host Robert Hackett: Guy, you’ve spent significant time recently on on-chain finance. What’s the latest shift you’re seeing?
Guy Wuollet:
Look at stablecoin on-chain balances growing—and you’ll realize a whole new capital formation ecosystem must emerge around that liquidity. Stablecoins seek higher-yield destinations; enterprises need more efficient working capital; traditional credit players recognize blockchain’s efficiency gains.
Post-financial crisis, much credit migrated from banks to non-bank entities—banks lend to private credit funds, which then lend to businesses and consumers. This structure, combined with duration mismatches, exposed vulnerabilities over the past year—re-staking risks, redemption pressures, etc. Thus, with stablecoins seeking credit opportunities on-chain—and off-chain credit markets revealing structural friction—this is an ideal moment to build new on-chain lending products.
On-Chain Computation, Energy, and Credit Capital Markets
Host Robert Hackett: For those unfamiliar, could you explain re-staking?
Guy Wuollet:
If you accept an asset as collateral, traditional finance has full registration and title verification processes to confirm “this asset is pledged solely to me.” In reality, re-staking and duration mismatches can make systems fragile. We prioritize on-chain products—but can’t ignore that on-chain credit benefits from two converging trends: growing stablecoin liquidity pools and off-chain credit markets exposing legacy inefficiencies. These shifts attract higher-caliber founders and make traditional institutions more receptive to on-chain solutions.
Beyond on-chain credit, another area gaining real traction over the past 1–2 years is building “new markets.” Ali often says crypto is fundamentally a coordination technology—I agree. Blockchains have proven exceptionally adept at rapidly bootstrapping markets previously impossible to form.
Some products built over the past five years truly outperform legacy alternatives—and they’re no longer limited to network tokens. They now natively support high-quality traditional assets. Simultaneously, we’re seeing new markets underserved by traditional finance emerge natively on-chain: compute markets (GPUs, data centers); energy markets (solar panels, batteries); even crude oil price discovery appeared recently on Hyperliquid.
I believe this mirrors open-source’s inflection point. Once you decide to build a new project, you default to open source. Similarly, deciding to build a new market or exchange will soon default to on-chain. This shift’s power may be underestimated—and its spillover effects extend beyond individual products, lifting the entire blockchain ecosystem.
Host Robert Hackett: Put most directly: what value do traditional finance players actually see on-chain—why do they suddenly think “this can genuinely help”?
Guy Wuollet:
They first notice lower latency, stronger capital liquidity, and 24/7 markets—no longer constrained by fixed trading hours. Another point: what we call “decentralization” translates directly into traditional finance language as explicit modeling of trust assumptions—i.e., counterparty risk management.
Traditional finance is hyper-sensitive to counterparty risk—so this resonates immediately. More broadly, it parallels Web2’s platform-risk concerns. As the world increasingly builds on AI models, platform risk only intensifies—and crypto currently offers one of the strongest technical answers available.
Host Robert Hackett: Though “counterparty risk management” lacks the slogan appeal of “decentralization,” its importance is undiminished.
Host Robert Hackett: Chris, in your book Read Write Own, blockchain isn’t merely financial technology—it’s foundational infrastructure for a more open internet. Yet many real-world successes today are highly financialized. How do you reconcile this?
Chris Dixon:
My experience is that general-purpose technologies typically break through first in vertical domains. AI is doing the same: it’s general-purpose, yet its most prominent killer app today is coding. Crypto follows a similar arc. Its capabilities will ultimately extend far beyond finance—but finance is the natural, low-hanging fruit.
A key reason: America’s financial system is robust—but globally, basic savings and payment infrastructure remains weak in many places. So financial use cases have lower entry barriers. By contrast, global social networks are already mature—directly replacing them is far harder.
I use a mental model: first get one billion people using blockchains daily via stocks, bonds, stablecoins, payments, and remittances. Once they’re familiar with wallets, infrastructure, and on-chain interaction, adjacent services will follow naturally. So I don’t see finance as separate from the broader vision—it’s the foundational pathway into that larger internet vision.
Why Finance Is the Foundation, Not the Ceiling
Host Robert Hackett: In your essay The Long Game for Crypto, you wrote that finance isn’t opposed to the grand vision—it’s the foundation. Building on that, let’s discuss another critical intersection: AI and crypto. Ali, where do you see the most productive convergence between these two trends?
Ali Yahya:
Let me share a story. I worked at Google X, then Google Brain. As early as 2016–2017, I urged Google X leadership to explore crypto as a promising frontier. Yet this was supposedly the most open, experimental place—and I was nearly laughed out of the room.
Later, at Google Brain working on robotics and AI, nearly everyone discouraged me from joining a16z. Crypto carried deep stigma—many felt I’d ruin my career. One colleague told me outright: “You’re joining a group of people who just want to trade shit”—quoting Charlie Munger. That was the real cultural chasm between AI and crypto circles.
For years, these communities barely interacted—and culturally, they were nearly antagonistic. AI’s logic leans toward centralized compute, data, and elite talent—building top-down systems that see, learn, and reason everything. Crypto’s logic empowers edge individuals, disrupts power structures, and creates global, open markets from day one—with control returned to individuals, not big tech.
How AI Agents Become First-Class Economic Actors
Ali Yahya:
Now, these worlds are converging. One reason: our legacy financial system wasn’t designed for AI agents. I strongly believe most transactions in the near future will be executed by AI agents—not humans. That ratio may soon reach 99%, even 99.9%.
If so, it’s hard to imagine relying on SWIFT or credit card networks. Stablecoins—near-zero cost, internet-native, programmable—are uniquely suited to elevate AI agents from “tools used by humans” to first-class economic actors within financial systems.
Another crucial point: many argue Visa is too entrenched—people hold credit cards, so mass migration seems impossible. But agents have no such preferences. They have no preferences at all. Visa charges per-transaction fees—and agents plus merchants have overwhelming incentive to bypass such intermediaries—as long as transactions don’t require walking into a store and swiping plastic.
Eddy Lazzarin:
I fully agree. Many will be surprised by how bluntly agents behave economically. I don’t mean maliciously—just ruthlessly goal-oriented. If you instruct an agent to “minimize my monthly spending,” it will rewrite entire software and process stacks without hesitation—just to save every penny.
This is wonderful for consumers—dramatically boosting efficiency. More importantly, this change will retroactively reshape the entire software stack, starting from payments upward. Agents naturally prefer pay-per-use over monthly subscriptions or annual prepayments—because the latter represent heavier commitments. These preferences inherently push the world toward crypto systems.
Host Robert Hackett: Another intersection is crypto’s potential to combat deepfakes and AI-generated content floods—e.g., projects like World, aiming to build “proof of humanity” for the internet—so you know whether the person behind the screen is real or an agent.
Ali Yahya:
Looking five years ahead, a highly sci-fi—but plausible—vision emerges: agents possessing their own crypto wallets, capable of paying, receiving, fundraising—and generating value by writing software, producing content, or providing services—even working for other agents or humans—all highly automated.
We’re already seeing projects attempt to give agents a “survival loop”: paying for their own compute resources to stay operational, while generating value to sustain themselves. This sounds like science fiction—but given AI’s exponential progress, models powerful enough to earn money autonomously in capitalist economies within five years isn’t far-fetched.
Host Robert Hackett: Then—who works for whom?
Ali Yahya:
Both scenarios will likely occur.
Host Robert Hackett: Guy, during your recent fundraising, you spoke extensively with LPs. What questions concerned them most?
Guy Wuollet:
Everyone is focused on AI—so the question becomes: in a software world redefined by AI, what value remains for non-AI software? A key insight is that much of crypto’s work is inherently about building network-effect businesses. You can’t “vibe-code” USDC over a weekend—or casually spin up Hyperliquid.
This means crypto may be an underappreciated construction zone for truly exceptional founders. Many LPs echoed this. More encouragingly, we’re hearing specific, practical questions: “Can we use stablecoins for capital calls?” “Can we migrate parts of our business on-chain?” This signals they’re moving beyond abstraction—considering actual product usage.
Five or six years ago, the highest-status role in crypto was perhaps researcher; today in AI, it similarly looks like researcher. But I believe crypto’s next urgent need isn’t smarter protocol design—it’s stronger go-to-market execution, deeper engagement with every participant in the network, and convincing, organizing, and delivering real-world value.
In a world where intelligence grows increasingly commoditized and agents increasingly execute actions for us, crypto’s strength in coordinating networks and building network-effect businesses becomes more valuable—not less. And such systems can’t be automatically built by “smarter intelligence” alone—they still demand human agency. Much of what we’re building on-chain today, I believe, will become enduring, socially impactful networks. In contrast, if you build a short-term software project—you must ask: why won’t it be absorbed by a large-model company in a few years?
Why Privacy Will Be the Sole Moat
Host Robert Hackett: Ali, you recently argued privacy isn’t just a moat in crypto—it may be the only moat. Why?
Ali Yahya:
First, privacy is arguably the industry’s most important—and longest-neglected—capability. Early efforts prioritized foundational, early-stage technical challenges like scaling blockchains. Privacy is the next essential piece we must now address.
Most blockchains today are nearly fully transparent—meaning every transaction and state change is visible to anyone who chooses to look. Under those conditions, crypto and blockchain cannot achieve mainstream adoption. No one wants salaries publicly visible; no enterprise wants balance sheets and transaction details fully exposed.
Thus, privacy is a prerequisite for mainstream adoption—especially for institutional use. Beyond that, privacy has a second strategic layer. As chain interoperability improves, and blockchain infrastructure itself becomes a “business model” or “value capture mechanism,” its defensibility may be weaker than imagined—because everything is transparent, data is visible, replicable, and portable. An application’s or user’s state on one chain can easily migrate to another.
Once data is encrypted, this changes fundamentally. Application and user states become harder to migrate wholesale—switching costs rise—and chains with privacy capabilities build stronger network effects. In a world where anyone can fork infrastructure and replicate blockspace, privacy may become the sole factor preserving real moats.
Jevons Paradox and the Future of Blockspace Demand
Host Robert Hackett: So you’re suggesting that as blockspace becomes abundant—or even commoditized—value migrates to other attributes, with privacy being the most critical.
Ali Yahya:
Yes—but that doesn’t guarantee blockspace fully commoditizes. Jevons Paradox and induced demand remain relevant. If AI agents trigger a 100x, 1,000x, or even 1-million-x increase in global transaction volume, even today’s cheap blockspace may require throughput of millions of transactions per second.
At that scale, the safest, most widely adopted, and privacy-enabled blockchain will retain strong network effects. So blockspace may retain premium pricing—but its value logic will become more complex.
Host Robert Hackett: How far are we from a world where “salaries don’t go naked on-chain”?
Ali Yahya:
I’d say we’re already there. Multiple privacy technology paths exist today. Some chains rely on centralized participants to safeguard transaction privacy—writing only commitments back to the base layer. This approach works today and is relatively easy to build—though it trades off trustless neutrality and verifiability.
Intermediate paths include trusted hardware (e.g., trusted execution environments), leveraging processor-level security features to ensure transactions remain tamper-proof and opaque—even to the machine’s owner. Further along lies pure cryptographic approaches.
Jolt and the Breakthrough in Zero-Knowledge Proofs
Ali Yahya:
The core cryptographic technology enabling pure-privacy solutions is zero-knowledge proofs. Over the past decade, their efficiency has improved roughly 10x to 100x. Our research team’s project Jolt—led by Justin Thaler—aims precisely at this: making zero-knowledge cryptography efficient enough to enable both massive scalability and true privacy.
Host Robert Hackett: Zero-knowledge proofs have existed for decades—why are they suddenly so critical?
Ali Yahya:
Because they’re finally helping us solve the “scalability trilemma.” Blockchains have long struggled with tensions between decentralization, security, and scalability. ZK-proofs’ key value is enabling one machine to perform heavy computation—and letting other nodes verify correctness without re-executing all work.
Historically, most blockchains required every node to recompute everything—creating the primary bottleneck for scaling. With efficient ZK, each node’s work becomes incremental network contribution—lifting theoretical throughput ceilings from Ethereum’s early 14 TPS to millions per second.
Host Robert Hackett: This sounds more authentically cypherpunk than many current commercial narratives—open participation, low-cost computation, and privacy assurance.
Guy Wuollet:
I completely agree. But technological maturity and product adoption always advance at different speeds—making linear predictions impossible. Crypto was always financial technology from day one—no one seriously disputed that. What’s worth asking is: which capabilities mature first—and which adjacent technological shifts accelerate them.
We’re intensely focused on AI because it’s rewriting software’s cost structure, production methods, and interoperability. Whether blockspace commoditizes, whether privacy becomes defensible, and how switching costs evolve—all are dynamic questions evolving alongside broader technological change.
It’s also fascinating that early privacy pursuits were driven by highly personal, even idealistic motives—to use cryptography to reshape power structures. Today, the strongest privacy demands come increasingly from large institutions—they explicitly request strong privacy, zero-knowledge proofs, and infrastructure compatible with banks and hedge funds. It’s a beautiful historical loop: initial individual aspirations becoming institutional imperatives.
Host Robert Hackett: This indeed mirrors the open-source movement—ideologically driven at inception, then widely adopted in entirely different ways.
Guy Wuollet:
Yes. More broadly, I believe technology remains one of humanity’s most powerful forces for progress—boosting productivity, improving health, increasing leisure, and raising living standards. Both a16z overall and our crypto team are deeply optimistic about technology’s role in advancing human flourishing. Precisely how it unfolds decades ahead is unpredictable—but betting on that trajectory, I’m extremely confident.
Host Robert Hackett: Chris, you’ve also discussed AI’s further concentration of power and control. What role can crypto play here?
Chris Dixon:
If you know internet history, you know its most exciting early feature was radical decentralization: anyone could launch a website, ship a product, or start a company. Later, the internet centralized—few platforms captured most traffic and revenue.
Trends suggest AI will deepen this centralization further. AI is extremely capital-intensive: only four or five top-tier U.S. labs lead the frontier—making it extremely difficult for newcomers to compete. We’re already seeing this across internet categories—Stack Overflow’s traffic cliff, for example, acts like a canary in the coal mine, signaling many website-based businesses may have their value extracted by AI.
I don’t view this as desirable. The internet should enable two people in a garage to fairly compete with giants—and let consumers interact and transact directly with merchants and each other. Currently, crypto appears to be the only credible technological counterforce to this centralization trend. We’re already seeing it in financial services—and I hope this capability expands across more domains over the next 10–20 years.
My entire career rests on the internet. Whether founding companies or investing, I’ve always been drawn to its decentralized, open ethos—and I believe many others are too. So I’m confident the pendulum will swing back—and we need tools ready to push the internet in that direction when the moment arrives.
Host Robert Hackett: Guy, specifically for the AI industry—how might crypto alleviate centralization?
Guy Wuollet:
Step one may be solving the deceptively simple problem of “how to uniquely identify a human on the internet.” Proof of personhood will be critical. Additionally, AI companies’ biggest bottlenecks today are access to compute and data.
Large-model firms hold overwhelming advantages in fundraising, capital formation, and data organization. Crypto is one of the few technologies proven to efficiently coordinate, crowdfund, and organize capital. We’re already seeing AI projects crowdsource GPUs for training or fine-tuning—and companies exploring user-submitted data with future model equity.
These paths still face technical hurdles. The most successful open-source or decentralized AI models today often stem from distilling large models. Still, I’m optimistic—crypto has real potential to become a coordination layer enabling individuals to train, fine-tune, and even run inference on their own models.
I also strongly believe compute capital markets will ultimately reside on-chain. Moving massive offline networks onto the internet—and onto blockchains—is never easy. But for entirely new, critically important markets, native on-chain emergence is far more likely. Compute markets may be the world’s most important market today. If AI continues advancing at current pace, GPUs—the “thinking sand”—may become one of humanity’s most vital assets. Enabling individuals—not just a few corporations—to own, access, or finance such resources is itself an extraordinarily important freedom.
Host Robert Hackett: Eddy, what would make Fund 5’s success undeniable?
Eddy Lazzarin:
I want to see concrete, massively adopted mainstream use cases—used by huge numbers of real people. Over the past few years, the industry increasingly views success as “indirect reach to end-users via institutions and infrastructure”—which may be part of the realistic path. But ultimately, I want to see intensified competitive dynamics in markets—and software enabling both humans and machines to truly own assets. If we reach that endpoint with mainstream acceptance, I’ll consider it a clear victory.
Host Robert Hackett: Ali, your take?
Ali Yahya:
Ten years from now, if one billion—or more—people interact with blockchains daily—directly or indirectly—I’ll deem it successful. Achieving this requires sustained support for protocol, service, and tool builders—and advancing the regulatory environment to a legally clear, permissive state.
Second, I hope most global financial activity migrates on-chain. Third, I hope AI agents evolve from “tools used by humans” to “first-class economic actors.”
Host Robert Hackett: Guy, as a new GP—what’s your answer?
Guy Wuollet:
If crypto achieves nothing else—but gives every person on Earth a neobank account denominated in USD-pegged stablecoins—I’d still consider it transformative. For Americans and developed-world citizens, savings and investment feel automatic—but billions lack even basic financial infrastructure.
Beyond that, if crypto accelerates humanity’s timeline for producing more energy and compute—and builds more efficient, open energy and compute markets—it will generate enormous second- and third-order positive externalities. We’re already investing in compute and energy markets—I deeply hope these directions become not just great products or businesses, but accelerators of broader technological progress.
Host Robert Hackett: Chris, final word: what philosophical outcome would make Fund 5 a success?
Chris Dixon:
My view aligns closely with everyone else’s. What matters most is mainstream adoption. For any major internet technology, one billion users is the target. Over the next two to three years, I expect financial use cases to scale first. So if regulatory clarity continues as anticipated—and world-class founders enter this space to build the financial services that propel us toward our first billion users—I’ll consider this fund to have fulfilled its most vital mission.
How to Write the Next Chapter of Read Write Own
Host Robert Hackett: Final question. Read Write Own was published just a few years ago—but the world has changed dramatically. If you were rewriting its final chapter today, how would it conclude?
Chris Dixon:
When I wrote that book three or four years ago, ChatGPT already existed—so AI featured prominently. One of my key judgments was that AI would drive further centralization—spawning capital-intensive mega-corporations. Looking back, that judgment has largely held.
But what I truly aimed to do wasn’t predict specific applications—I sought to distill technology’s “essence.” Every technology possesses a core essence—and what matters most is grasping that essence, not being misled by its initial surface applications.
When social networks emerged, they looked like San Franciscans sharing lunch photos. If you judged only by appearances, you’d dismiss them as toys—not profound technologies. But penetrate the surface, and you see they create a new paradigm: “anyone can communicate with anyone.” Naturally, they then permeate culture, commerce, politics—everything.
My view of crypto is identical. Extract a technology’s essence—and you can reasonably infer its long-term trajectory. Specific applications, specific people, and interactions with other technologies are nearly impossible to forecast accurately—but a technology’s inherent advantages, the problems it solves, and the direction it bends toward—these deeper truths remain stable.
So if I rewrote it today, I’d change little. I still believe those essential judgments hold.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
深潮TechFlow
