Author: a16z New Media
Translation: TechFlow
Over the past two days, we have shared the challenges and opportunities that builders will face in 2026, as seen by teams focused on infrastructure, growth, life sciences and health, Speedrun, applications, and American vitality.
Today, we share 17 insights about the future, summarized by several a16z crypto partners (and some special contributors). These topics cover everything from smart agents and artificial intelligence (AI), stablecoins, tokenization and finance, privacy and security, to prediction markets, SNARKs (zero-knowledge proof technology), and other applications... as well as how we will build in the future. (To stay updated on trends, builder guides, industry reports, and other crypto resources, be sure to subscribe to the a16z crypto newsletter.)
Tomorrow, we will conclude the week with a special announcement and an exclusive invitation from a16z—don’t miss it!
Here are today’s highlights:
Privacy Will Become the Most Important Moat in Crypto
Privacy is one of the key features driving global finance on-chain, yet it is a missing piece in almost all current blockchains. For most blockchains, privacy is merely a secondary or even neglected function.
However, today, privacy itself is attractive enough to make a chain stand out among many competitors. More importantly, privacy can create a “chain lock effect,” or what could be called a “privacy network effect.” In a world where performance competition alone is no longer sufficient, this effect is especially important.
Due to the existence of cross-chain bridge protocols, as long as everything is public, migrating from one chain to another becomes extremely simple. However, once privacy is introduced, this convenience disappears: migrating tokens is easy, but migrating secrets is hard. When moving from a privacy chain to a public chain, or between two privacy chains, there are always risks, such as third parties observing on-chain transactions, mempools, or network traffic that could identify you. Crossing the boundaries between privacy and public chains, or even between two privacy chains, leaks various metadata, such as the correlation of transaction times and sizes, making tracking easier.
Compared to the many homogeneous new chains—whose fees may drop to zero due to competition (as the essence of block space has become similar across chains)—blockchains with privacy features can form stronger network effects. In fact, if a “general-purpose” blockchain does not have a thriving ecosystem, killer app, or asymmetric distribution advantage, there is almost no reason to attract users or developers, let alone user loyalty.
On public blockchains, users can easily transact with users on other chains—it doesn’t matter which chain they join. However, on privacy blockchains, the choice of chain becomes especially important, because once users join a chain, they are less willing to migrate to avoid exposure risks. This creates a “winner-takes-all” dynamic. And since privacy is a necessary condition for most real-world scenarios, a few privacy chains may capture most of the crypto market.
——Ali Yahya, a16z Crypto Partner
Prediction Markets: A Bigger, Broader, Smarter Future
Prediction markets have moved from niche to mainstream, and in the coming year, at the intersection of crypto and AI, they will become larger, broader, and smarter, while also presenting new challenges for builders.
First, more contracts will be listed. This means we can not only get real-time probabilities about major elections or geopolitical events, but also learn about the outcomes of various details and complex intersecting events. As these new contracts reveal more information and gradually integrate into the news ecosystem (a trend already underway), they will also raise important social questions, such as how to balance the value of this information, and how to better design these markets for transparency and auditability—issues that can be addressed with crypto technology (SC will link to our related articles>).
To handle the larger scale of contracts, we need new ways to reach consensus on the truth to resolve contract issues. The solutions of centralized platforms (did an event really happen? How do we confirm?) are crucial, but controversial cases like the Zelensky lawsuit market and the Venezuela election market expose their limitations. To address these edge cases and help prediction markets expand to more useful applications, new decentralized governance and LLM-based oracles can help determine the truth of disputed outcomes.
AI’s application in oracles can go beyond LLMs. For example, AI agents trading on these platforms can search for global signals, providing an edge for short-term trading, thereby revealing new insights about the world and predicting what might happen in the future. (Projects like Prophet Arena already demonstrate the potential of this field.) Besides serving as complex political analysts for our queries, when we study these agents’ strategies, they may also reveal the fundamental predictive factors of complex social events.
Will prediction markets replace polls? No; they will make polls better (and poll data can be input into prediction markets). As a political scientist, I am most interested in how prediction markets will work alongside a rich and vibrant polling ecosystem—but this will require new technologies, such as AI to improve the polling experience, and crypto to provide new ways to prove that poll/survey participants are humans, not bots, among other features.
——Andy Hall, a16z Crypto Research Advisor (and Stanford University Professor of Political Economy)
Tokenization and Stablecoins: A More “Crypto-Native” Perspective on Real-World Assets
We are seeing banks, fintech companies, and asset managers showing strong interest in putting US stocks, commodities, indices, and other traditional assets on-chain. However, as more traditional assets go on-chain, this tokenization is often “skeuomorphic”—based on current perceptions of real-world assets, without fully leveraging crypto-native features.
But synthetic representations like perpetual contracts (perps) not only provide deeper liquidity but are often easier to implement. Perpetuals also have an easy-to-understand leverage mechanism, making them the most suitable derivatives for crypto markets, in my view. I also believe that emerging market equities are among the most promising asset classes for “perpification.” (For example, some stocks’ “0DTE” options markets have higher liquidity than spot markets, providing a very interesting experiment for perpification.)
This boils down to the issue of “perpification vs. tokenization”; but in any case, we should see more crypto-native tokenization of real-world assets (RWA) in the coming year.
Similarly, in 2026, after stablecoins go mainstream in 2025, we will see more “issuance, not just tokenization,” and the outstanding issuance of stablecoins will continue to grow.
However, stablecoins without strong credit infrastructure are more like “narrow banks,” holding only specific liquid assets considered especially safe. While narrow banks are an effective product, I don’t think they will become the backbone of the on-chain economy in the long run.
We have already seen many new asset managers, curators, and protocols facilitating on-chain asset-backed loans based on off-chain collateral. These loans are usually initiated off-chain and then tokenized. But I believe the benefits of tokenization in this case are limited, perhaps only making it easier to distribute assets to users already on-chain. Therefore, debt assets should be originated directly on-chain, not off-chain and then tokenized. On-chain origination can reduce the cost of loan servicing and back-end structuring, and improve accessibility. The challenge lies in compliance and standardization, but developers are already working to solve these issues.
——Guy Wuollet, a16z Crypto General Partner
The Crypto Business Waystation: Trading Is Not the Final Destination
Today, aside from stablecoins and some core infrastructure, almost every well-performing crypto company has pivoted or is pivoting to trading. But if “every crypto company becomes a trading platform,” what does the future hold for the industry? When too many players do the same thing, not only does it dilute market attention, but it also means only a few big companies will win. This also means companies that pivot to trading too early miss the opportunity to build more defensible and lasting businesses.
While I sympathize with founders striving to make their company finances work, chasing short-term product-market fit also comes at a cost. This issue is especially pronounced in crypto, where the unique dynamics of tokens and speculation lead founders to favor instant gratification in the search for product-market fit... it’s a kind of “marshmallow test” (referring to the test of delayed gratification). Trading itself is not wrong—it’s an important market function—but it doesn’t have to be the ultimate business destination. Founders who focus on the “product” part of product-market fit may ultimately become bigger winners.
——Arianna Simpson, a16z Crypto Partner
From “Know Your Customer” (KYC) to “Know Your Agent” (KYA)
The bottleneck in the agent economy is shifting from intelligence to identity.
In financial services, “non-human identities” now outnumber human employees 96 to 1—yet these identities remain ghostlike, unable to enter the banking system. The missing key infrastructure here is KYA: Know Your Agent.
Just as humans need credit scores to get loans, agents need cryptographically signed credentials to transact—credentials that link agents to their principals, constraints, and liabilities. Until this mechanism is established, merchants will continue to block agents at the firewall. The financial industry spent decades building KYC infrastructure, but now it has only months to solve KYA.
——Sean Neville, Circle Co-founder and USDC Architect; Catena Labs CEO
The Future of Stablecoins: Better, Smarter Onramps and Offramps
Last year, stablecoin transaction volume was estimated at $46 trillion, continually hitting new highs. To put this in perspective, that’s more than 20 times PayPal’s volume; nearly three times that of Visa, one of the world’s largest payment networks; and rapidly approaching the transaction volume of the US Automated Clearing House (ACH)—the electronic network that handles direct deposits and other financial transactions in the US.
Today, you can complete a stablecoin transaction in less than a second, at a cost of less than a penny. However, the unresolved issue is how to connect these digital dollars to the financial systems people use every day—in other words, how to build stablecoin onramps and offramps.
A new generation of startups is filling this gap, connecting stablecoins to more familiar payment systems and local currencies. Some use cryptographic proofs to let people privately convert local balances into digital dollars. Some integrate with regional networks, using QR codes, real-time payment rails, and more to enable interbank payments... Others are building truly interoperable global wallet layers and card platforms, allowing users to spend stablecoins at everyday merchants. Together, these approaches broaden participation in the digital dollar economy and may accelerate stablecoins’ adoption as a mainstream payment method.
As these onramps and offramps mature, digital dollars will directly connect to local payment systems and merchant tools, spawning new behaviors: cross-border workers can receive wages in real time; merchants can accept global dollar payments without a bank account; apps can settle instantly with users anywhere, anytime. Stablecoins will evolve from a niche financial tool to the foundational settlement layer of the internet.
——Jeremy Zhang, a16z Crypto Engineering Team
Stablecoins: Unlocking the Bank Ledger Upgrade Cycle and Enabling New Payment Scenarios
Today, many banks still use software systems that modern developers can barely recognize: in the 1960s and 70s, banks were early adopters of large software systems; by the 80s and 90s, second-generation core banking software (like Temenos’s GLOBUS and InfoSys’s Finacle) emerged. However, these systems have aged, and upgrades are too slow. As a result, the banking industry—especially the critical core ledger databases that track deposits, collateral, and other obligations—still largely runs on mainframe computers, using COBOL, and relies on batch file interfaces rather than modern APIs.
The vast majority of global assets still depend on these decades-old core ledgers. While these systems have been validated by long-term practice, trusted by regulators, and deeply integrated into complex banking scenarios, they also hinder innovation. For example, adding key features like real-time payments (RTP) to these systems can take months or years, with layers of technical debt and regulatory complexity.
This is where stablecoins shine. In recent years, stablecoins have not only found product-market fit and gone mainstream, but this year, traditional financial institutions (TradFi) have embraced them at new heights. Stablecoins, tokenized deposits, tokenized treasuries, and on-chain bonds are enabling banks, fintechs, and financial institutions to build new products and serve new customers. More importantly, these institutions can innovate without fully rewriting their legacy systems—which, though aging, have run reliably for decades. Thus, stablecoins provide institutions with a whole new path to innovation.
——Sam Broner
Decentralization Is the Future of Messaging—More Important Than Quantum Encryption
As the world moves toward the quantum computing era, many crypto-based messaging apps (like Apple, Signal, WhatsApp) are leading the way and achieving great results. However, the problem is that almost all major messaging apps rely on private servers operated by a single organization. These servers are easy targets for governments to shut down, backdoor, or forcibly access private data.
If a country can shut down your server, if a company holds the keys to private servers, or even if a company just has a private server, what’s the point of quantum encryption? Private servers require users to “trust me,” but if there are no private servers, it means “you don’t need to trust me.” Communication doesn’t need a middleman company to operate.
Messaging needs open protocols so users don’t have to trust anyone. The way to achieve this is decentralized networks: no private servers, no single app, all code open source, and using top-notch encryption—including quantum-resistant cryptography.
With open networks, no individual, company, nonprofit, or country can take away our ability to communicate. Even if a country or company shuts down an app, 500 new versions will appear the next day. Even if a node is shut down, economic incentives from technologies like blockchain will prompt new nodes to immediately take its place.
When people control their messages with keys, just as they control their money, everything changes. Apps may come and go, but users will always control their messages and identities. Even if an app fails, end users can still own their messages.
This is not just about quantum resistance and encryption, but about ownership and decentralization. Without both, what we build is just an unbreakable but still shutdown-able crypto system.
——Shane Mac, XMTP Labs Co-founder and CEO
From “Code Is Law” to “Specification Is Law”—A New Evolution in DeFi Security
Recent DeFi hacks have targeted battle-tested protocols, run by strong teams, strictly audited, and live for years. These incidents reveal an unsettling reality: current security best practices still rely mainly on heuristics and case-by-case handling.
For DeFi security to mature further, we need to shift from patching vulnerability patterns to property assurance at the design level, from “best effort” to “principled approaches”:
In the static/pre-deployment phase (testing, auditing, formal verification, etc.), this means systematically verifying global invariants, not just hand-picked local ones. Today, multiple teams are building AI-assisted proof tools to help write specifications, propose invariants, and share the burden of what was once expensive and time-consuming manual proof engineering.
In the dynamic/post-deployment phase (runtime monitoring, runtime enforcement, etc.), these invariants can become real-time “guardrails”—the last line of defense. These guardrails are directly encoded as runtime assertions, ensuring every transaction must satisfy them.
So now, we no longer assume every vulnerability is caught in advance, but embed critical security properties directly into code, automatically rolling back any transaction that violates them.
This isn’t just theory. In practice, almost every past attack would have triggered these checks during execution, potentially stopping the hack. Thus, the “code is law” concept is evolving into “specification is law”: even novel attacks must satisfy security properties that maintain system integrity, leaving only minor or extremely difficult attacks possible.
——Daejun Park, a16z Crypto Engineering Team
Crypto Technology Beyond Blockchain: Ushering in a New Era of Verifiable Computation
For years, SNARKs (succinct non-interactive arguments of knowledge)—a cryptographic proof technology for verifying computation without re-execution—were almost exclusively used in blockchains. This was due to high computational costs: generating a proof could require 1,000,000 times the work of running the computation directly. Such costs are worthwhile when spread across thousands of verifiers, but impractical elsewhere.
This is about to change. By 2026, zkVM (zero-knowledge virtual machine) provers will have overheads down to about 10,000 times, with memory usage of just a few hundred megabytes—fast enough to run on a phone, cheap enough for many scenarios. Why is “10,000 times” a magic number? Because high-end GPUs have about 10,000 times the parallel throughput of a laptop CPU. By the end of 2026, a GPU will be able to generate proofs of CPU computations in real time.
This breakthrough could realize the vision of early research papers: verifiable cloud computing. If you already run CPU workloads in the cloud—whether due to insufficient computation for GPUs, lack of expertise, or legacy system constraints—you’ll be able to get cryptographic proofs of correctness at reasonable cost. And these provers are already GPU-optimized, requiring no extra code changes.
——Justin Thaler, a16z Crypto Researcher & Associate Professor of Computer Science at Georgetown University
AI Will Become a Research Assistant
As a mathematical economist, in January this year, I still struggled to get consumer-grade AI models to understand my workflow; by November, I could already give models abstract instructions as I would to a PhD student... and sometimes they even provided novel and correct answers. Beyond my personal experience, we’re also seeing AI applied more widely in research, especially in reasoning—models now not only participate directly in discovery but can autonomously solve Putnam problems (possibly the world’s hardest undergraduate math exam).
It’s still unclear which fields this research assistance will be most effective in, or exactly how it will work. But I expect AI research will foster and reward a new “versatile” research style: emphasizing the ability to speculate on relationships between ideas and quickly extrapolate from more hypothetical answers. These answers may not be entirely accurate, but can still point in the right direction (at least in some topology). Ironically, this approach is a bit like harnessing model “hallucinations”: when models are “smart” enough, giving them an abstract space to explore may generate some nonsense, but may also spark discoveries—just as humans are often more creative when working in nonlinear, ambiguous directions.
This reasoning style requires a new AI workflow—not just “agent-to-agent,” but “agent-wrapping-agent” structures. In this setup, different model layers help researchers evaluate early model approaches and gradually extract valuable content. I’m already using this method to write papers, while others use it for patent searches, creating new art forms, and even (regrettably) finding new smart contract attack vectors.
However, to efficiently run such a reasoning-agent-based research system, better model interoperability is needed, as well as a way to identify and fairly compensate each model’s contribution—problems that crypto can help solve.
——Scott Kominers, a16z Crypto Research Team Member & Harvard Business School Professor
The “Invisible Tax” of Open Networks: Economic Imbalance and Solutions in the AI Era
With the rise of AI agents, open networks are facing an invisible tax that is fundamentally undermining their economic foundation. This stems from a growing mismatch between the internet’s “context layer” and “execution layer”: currently, AI agents extract data from ad-supported content sites (context layer) to provide convenience for users, while systematically bypassing the revenue sources (like ads and subscriptions) that support this content.
To prevent further erosion of open networks and protect the diverse content ecosystem that fuels AI, we need to deploy technical and economic solutions at scale. This could include next-generation sponsored content models, micro-attribution systems, or other novel funding models. However, existing AI licensing agreements are proving financially unsustainable—they often only compensate content providers for a small fraction of the revenue lost to AI traffic diversion.
The network needs a new techno-economic model for automated value flow. The key shift in the coming year will be moving from static licensing to real-time usage-based compensation. This means testing and scaling systems—possibly using blockchain-powered nano-payments and advanced attribution standards—to automatically reward every entity that contributes information to AI agents’ successful task completion.
——Liz Harkavy, a16z Crypto Investment Team
The Rise of “Staked Media”: Reshaping Trust with Blockchain
Cracks in the traditional media model of “objectivity” have been evident for some time. The internet has given everyone a voice, and now more operators, practitioners, and builders are expressing their views directly to the public. Their perspectives reflect their interests in the world, and surprisingly, audiences often respect them for these interests, not despite them.
The real new change is not the rise of social media, but the arrival of crypto tools that enable people to make publicly verifiable commitments. In an era where AI makes it cheap and easy to generate infinite content—from any perspective, real or fake—relying solely on what people (or bots) say is no longer enough. Tokenized assets, programmable lockups, prediction markets, and on-chain history provide a stronger foundation for trust: commentators can prove they “put their money where their mouth is” when expressing opinions; podcasters can lock tokens to show they won’t “pump and dump”; analysts can link predictions to publicly settled markets, creating auditable records.
This is what I call the prototype of “Staked Media”: a media form that not only accepts the idea of “skin in the game,” but also provides proof. In this model, credibility no longer comes from feigned detachment or baseless claims, but from clear, transparent, verifiable commitments. “Staked Media” won’t replace other media forms, but will complement existing models. It offers a new signal: not just “trust me, I’m neutral,” but “here’s the risk I’m willing to take, and how you can verify if I’m telling the truth.”
——Robert Hackett, a16z Crypto Editorial Team
“Secrets-as-a-Service”: How Privacy Protection Becomes Core Internet Infrastructure
Behind every model, agent, and automation system lies a simple but crucial factor: data. Yet most of today’s data pipelines—the flows of data into or out of models—are opaque, mutable, and unauditable. This may be fine for some consumer apps, but for many industries and users (like finance and healthcare), enterprises need to ensure the privacy of sensitive data. For institutions trying to tokenize real-world assets, this is an even bigger barrier.
So, how can we achieve secure, compliant, autonomous, and globally interoperable innovation while protecting privacy? While there are many approaches, I focus on data access control: who controls sensitive data? How does data flow? Who (or what) can access it?
Without data access control, anyone wanting to protect data confidentiality must currently rely on centralized services or custom solutions—which are time-consuming, costly, and hinder traditional financial institutions and other industries from fully leveraging on-chain data management. As agent systems begin to autonomously browse, trade, and make decisions, users and institutions across industries need cryptographic guarantees, not just “best effort” trust.
Therefore, I believe we need “Secrets-as-a-Service”: a new technology that provides programmable native data access rules, client-side encryption, and decentralized key management, clearly specifying who can decrypt data under what conditions, and for how long... all enforced on-chain. Combined with verifiable data systems, “secrets” can become part of the internet’s basic public infrastructure, rather than a privacy feature added at the application layer after the fact. This will make privacy a core part of internet infrastructure.
——Adeniyi Abiodun, Chief Product Officer and Co-founder, Mysten Labs
Wealth Management for Everyone
Personalized wealth management services have traditionally been reserved for high-net-worth clients, as providing customized advice and portfolio allocation across asset classes is expensive and complex. However, as more asset classes become tokenized, crypto infrastructure enables AI-recommended and assisted personalized investment strategies to be executed and adjusted instantly at extremely low cost.
This is not just an upgrade to “robo-advisors”: everyone can enjoy active portfolio management, not just passive. In 2025, TradFi has already shifted 2-5% of portfolios into crypto (via direct bank investment or ETPs), but this is just the beginning; by 2026, we’ll see more platforms focused on “wealth accumulation” rather than just “wealth preservation”—fintechs (like Revolut and Robinhood) and centralized exchanges (like Coinbase) will leverage their tech advantages to capture more market share.
Meanwhile, DeFi tools like Morpho Vaults can automatically allocate assets to lending markets with the best risk-adjusted returns, providing core yield allocation for portfolios. Additionally, holding surplus liquidity in stablecoins instead of fiat, and investing in tokenized money market funds instead of traditional ones, can further expand yield possibilities.
Finally, ordinary investors now have easier access to more illiquid private market assets, such as private credit, pre-IPO companies, and private equity. Tokenization unlocks these markets while still meeting compliance and reporting requirements. As all parts of a balanced portfolio become tokenized (from bonds to stocks to private and alternative assets), these assets can be automatically rebalanced without the hassle of bank transfers and other manual operations.
——Maggie Hsu, a16z Crypto Go-to-Market Team
The Internet Becomes a Bank: The Future of Value Flow
With the widespread adoption of AI agents and more transactions happening automatically in the background rather than via user clicks, the way money—i.e., value—flows must change. In a world where systems operate based on intent rather than step-by-step instructions, funds may move because AI agents identify needs, fulfill obligations, or trigger outcomes. At that point, value needs to flow as quickly and freely as information does today, and blockchain, smart contracts, and new protocols are key to achieving this.
Today, smart contracts can already settle dollar payments globally in seconds. By 2026, emerging foundational tools (like x402) will make such settlement programmable and reactive. Agents can instantly and permissionlessly pay each other for data, GPU time, or API calls—no invoices, reconciliation, or batch processing; developers can release software updates with built-in payment rules, limits, and audit logs—no fiat integration, merchant onboarding, or bank involvement; prediction markets can settle automatically in real time as events unfold—no custodians or exchanges, odds update in real time, agents trade, and payments settle globally in seconds.
When value can flow this way, “payment flow” will no longer be a separate operational layer, but part of network behavior. Banks will become part of internet infrastructure, and assets will become infrastructure. If money can be routed through the internet like data packets, then the internet won’t just support the financial system—it will become the financial system itself.
——Christian Crowley and Pyrs Carvolth, a16z Crypto Go-to-Market Team
When Legal Architecture Matches Technical Architecture: Unlocking the Full Potential of Blockchain
Over the past decade, one of the biggest obstacles to building blockchain networks in the US has been legal uncertainty. Securities laws have been expanded and selectively enforced, forcing entrepreneurs into a regulatory framework designed for companies, not networks. For years, mitigating legal risk replaced product strategy; engineers’ roles were replaced by lawyers.
This dynamic led to many strange distortions: entrepreneurs were told to avoid transparency; token distribution became legally arbitrary; governance evolved into superficial “theater”; organizational structures were optimized for legal protection; token design was forced to avoid economic value, sometimes with no business model at all. Worse, crypto projects that skirted the rules often outpaced honest builders.
However, the US government is now closer than ever to passing crypto market structure regulation, which could eliminate all these asymmetries next year. If passed, this legislation will incentivize transparency, set clear standards, and replace “enforcement roulette” with a more structured path for financing, token issuance, and decentralization. Driven by GENIUS, stablecoin adoption has already exploded; legislation around crypto market structure will be an even bigger shift, but this time, it’s built for networks.
In other words, this regulation will allow blockchain networks to truly operate as networks—open, autonomous, composable, credibly neutral, and decentralized.
——Miles Jennings, a16z Crypto Policy Team and General Counsel
