a16z's Major Annual Report: The 17 Most Exciting Ideas in the Web3 Industry for 2026

Original Title: 17 things we're excited about for crypto in 2026

Compiled by: Jiahua, Chaincatcher

Editor’s Note: a16z released its annual "big ideas" this week, a collaboration among various teams (Apps, American Dynamism, Bio, Crypto, Growth, Infra, Speedrun). Below are observations from a16z crypto partners and guest contributors about the future—topics cover agents and AI; stablecoins, tokenization, and finance; privacy and security; prediction markets, SNARKs, and other applications, as well as how we will build.

1. Better and Smarter On/Off Ramps for Stablecoins

Last year, the trading volume of stablecoins was estimated to reach $46 trillion, continually setting historical records. To put this number into perspective: it is more than 20 times the volume of PayPal; nearly 3 times that of Visa (one of the largest payment networks in the world); and is rapidly approaching the trading volume of ACH (the electronic network for financial transactions in the U.S. used for direct deposits).

Today, you can send stablecoins for less than a cent and in under a second. However, the unresolved issue is how to connect these "digital dollars" to the financial rails that people use daily—in other words, the on/off ramps for stablecoins.

A new generation of startups is filling this gap by connecting stablecoins to more familiar payment systems and local currencies. Some companies use crypto proofs to allow people to privately convert local balances into digital dollars. Others integrate regional networks, utilizing QR codes, real-time payment rails, and other features to facilitate interbank payments; while others are building a truly interoperable global wallet layer and card issuance platforms that allow users to spend stablecoins at everyday merchants. These approaches collectively expand the participants in the digital dollar economy and may accelerate the direct use of stablecoins as a mainstream payment method.

As these on/off ramps mature, and digital dollars directly connect to local payment systems and merchant tools, new behaviors will emerge. Workers can receive wages in real-time across borders; merchants can accept global dollars without a bank account; applications can settle value instantly with any global user. Stablecoins will fundamentally transform from niche financial tools into the foundational settlement layer of the internet.

------ Jeremy Zhang, a16z crypto engineering team

2. Thinking About RWA Tokenization and Stablecoins in a More "Crypto-Native" Way

We see strong interest from banks, fintech companies, and asset management firms in putting U.S. stocks, commodities, indices, and other traditional assets on-chain. As more traditional assets go on-chain, current tokenization often tends to be "skeuomorphic"—rooted in the concepts of existing real-world assets without leveraging crypto-native functionalities.

However, synthetic representations like perpetual futures (Perps) allow for deeper liquidity and are often easier to implement. Perps also provide easily understandable leverage, so I believe they have the strongest product-market fit (PMF) among crypto-native derivatives. I also think emerging market stocks are one of the most interesting asset classes for "perpification." (The liquidity of "zero-day-to-expiration" or 0DTE options for certain stocks often exceeds that of the spot market, making it a fascinating experiment in perpification.)

It all comes down to the question of "perpification vs. tokenization"; but in any case, we should see more crypto-native RWA tokenization in the coming year.

Along similar lines, in 2026, as stablecoins enter the mainstream in 2025, we will see more "native origination, not just tokenization"; the outstanding issuance of stablecoins will continue to grow.

However, stablecoins without a strong credit infrastructure look like "narrow banks," holding specific liquid assets deemed particularly safe. While narrow banks are an effective product, I don't believe they will become a long-term pillar of the on-chain economy.

We have already seen many new asset management firms, curators, and protocols begin to facilitate on-chain asset-backed lending based on off-chain collateral. These loans are often issued off-chain and then tokenized. I believe the benefits of tokenization here are minimal, except perhaps for distribution to already on-chain users. This is why debt assets should be natively issued on-chain, rather than issued off-chain and then tokenized. Native on-chain issuance reduces the cost of loan servicing, backend structural costs, and improves accessibility. The challenging part here will be compliance and standardization, but builders are already working to address these issues.

------ Guy Wuollet, a16z crypto general partner

3. Stablecoins Unlocking Bank Ledger Upgrade Cycles—and New Payment Scenarios

The software that ordinary banks run is unrecognizable to modern developers: in the 1960s and 1970s, banks were early adopters of large software systems. The second generation of core banking software began in the 1980s and 1990s (e.g., through Temenos's GLOBUS and Infosys's Finacle). But all this software is aging and upgrading too slowly. Thus, the banking industry—especially the critical core ledgers (the key databases that track deposits, collateral, and other obligations)—still often runs on mainframes, programmed in COBOL, and uses batch file interfaces instead of APIs.

Most of the world's assets exist on these same decades-old core ledgers. While these systems are battle-tested, trusted by regulators, and deeply integrated into complex banking scenarios, they also hinder innovation. Adding key features like real-time payments (RTP) can take months, or more likely years, and requires dealing with layers of technical debt and regulatory complexity.

This is where stablecoins come in. Not only have stablecoins found product-market fit and entered the mainstream over the past few years, but this year, traditional financial (TradFi) institutions have embraced them at an unprecedented level. Stablecoins, tokenized deposits, tokenized government bonds, and on-chain bonds allow banks, fintech companies, and financial institutions to build new products and serve new customers. More importantly, they can do this without forcing these organizations to rewrite their legacy systems—systems that, while aging, have been reliably running for decades. Thus, stablecoins provide institutions with a new way to innovate.

------ Sam Broner, investment partner

4. The Internet as a Bank

With agents arriving at scale, and more business happening automatically in the background rather than through user clicks, the way money (i.e., value!) moves needs to change.

In a world where systems act based on "intent" rather than step-by-step instructions—where AI agents identify needs, fulfill obligations, or trigger outcomes by moving funds—value must flow as quickly and freely as today's information. This is the role of blockchains, smart contracts, and new protocols.

Smart contracts can already settle a one-dollar payment globally in seconds. However, by 2026, emerging primitives like x402 will make that settlement programmable and responsive: agents will pay each other instantly and permissionlessly for data, GPU time, or API calls—without invoices, reconciliations, or batch processing. Developers will release software updates bundled with built-in payment rules, limits, and audit trails—without fiat integration, merchant onboarding, or banks. Prediction markets will self-settle in real-time as events unfold—odds update, agents trade, and settle globally in seconds… without custodians or exchanges.

Once value can move this way, "payment flows" will no longer be a separate operational layer but will become network behavior: banks become part of the internet's basic plumbing, and assets become infrastructure. If money becomes data packets that the internet can route, then the internet is not just supporting the financial system… it becomes the financial system.

------ Christian Crowley and Pyrs Carvolth, a16z crypto listing team

5. Wealth Management for Everyone

Personalized wealth management services have traditionally been limited to high-net-worth clients of banks: providing customized advice and personalized portfolios across asset classes is costly and operationally complex. But as more asset classes are tokenized, strategies enabled by crypto rails—using AI recommendations and co-pilots for personalization—will be able to execute and rebalance instantly at very low costs.

This is not just about robo-advisors; everyone can access active portfolio management, not just passive management. In 2025, traditional finance increased its allocation to cryptocurrencies in portfolios (either directly or through ETPs), but this is just the beginning; in 2026, we will see platforms built for "wealth accumulation"—not just "wealth preservation"—as fintech companies (like Revolut and Robinhood) and centralized exchanges (like Coinbase) leverage their tech stack advantages to capture more market.

At the same time, DeFi tools like Morpho Vaults automatically allocate assets to lending markets with the best risk-adjusted returns—providing core yield distribution for portfolios. Holding remaining liquid balances as stablecoins instead of fiat, and holding them as tokenized money market funds instead of traditional money market funds, further expands yield possibilities.

Finally, retail investors now have easier access to more illiquid private market assets, such as private credit, pre-IPO companies, and private equity, as tokenization helps unlock these markets while still maintaining compliance and reporting requirements. As various components of balanced portfolios are tokenized (along the risk spectrum from bonds to stocks to private assets and alternative investments), they can be automatically rebalanced without cumbersome wire transfers.

------ Maggie Hsu, a16z crypto listing team

6. From "Know Your Customer" (KYC) to "Know Your Agent" (KYA)

The bottleneck of the AI agent economy is shifting from intelligence to identity.

In financial services, the number of "non-human identities" now exceeds human employees by 96 to 1—yet these identities remain unbanked "ghosts." The missing key primitive here is KYA: Know Your Agent.

Just as humans need credit scores to obtain loans, agents will need cryptographic signature credentials to transact—linking agents to their principals, their constraints, and their liabilities. Until this exists, merchants will continue to block agents at the firewall. The industry that spent decades building KYC infrastructure now has only a few months to figure out KYA.

------ Sean Neville, Co-founder of Circle and architect of USDC; CEO of Catena Labs

7. We Will Use AI for Substantive Research Tasks

As a mathematical economist, it was difficult for consumer-grade AI models to understand my workflow back in January; but by November, I could give the model abstract instructions as if it were a PhD student… and sometimes they would return novel and correctly executed answers. Beyond my experience, we are beginning to see AI being used more broadly in research—especially in reasoning, where models now directly assist in discovery and even autonomously solve the Putnam problem (perhaps the hardest university-level math exam in the world).

This remains an open question: which areas will benefit most from this research assistance, and how will it assist? But I expect AI research will enable and reward a new polymath research style: one that favors the ability to speculate on relationships between ideas and quickly infer from even more speculative answers. Those answers may be inaccurate but can still point in the right direction (at least under some topology). Ironically, this is somewhat like leveraging the power of model hallucination: when models are "smart enough," giving them abstract space to collide can still produce nonsense—but sometimes it can unlock the door to discovery, just as people are often most creative when they do not work in a linear, directed manner.

Reasoning in this way will require a new style of AI workflow—not just agent-to-agent, but more agent-wrapping-agent—where model layers help researchers evaluate early models and gradually refine them. I have been using this approach to write papers, while others conduct patent searches, invent new art forms, or (unfortunately) discover new types of smart contract attacks.

However: operating this research-wrapping reasoning agent ensemble will require better interoperability between models and a way to identify and appropriately compensate each model's contributions—both of which cryptocurrencies can help solve.

------ Scott Kominers, a16z crypto research team and Harvard Business School professor

8. The "Invisible Tax" of Open Networks

The rise of AI agents is imposing an invisible tax on open networks, fundamentally undermining their economic foundations. This disruption stems from the growing mismatch between the context layer and execution layer of the internet: currently, AI agents extract data from ad-supported websites (the context layer) to provide convenience to users while systematically bypassing the revenue streams that fund content (like ads and subscriptions).

To prevent the erosion of open networks (and retain the diverse content that powers AI itself), we need to deploy technological and economic solutions at scale. This could include next-generation sponsored content models, micro-ownership systems, or other new financing models. Existing AI licensing agreements have also proven to be financially unsustainable "band-aids," often compensating content providers for lost revenue due to AI siphoning traffic by only a small fraction.

The network needs a new technological economic model where value can flow automatically. A key shift in the coming year will be from static licensing to real-time, usage-based compensation. This means testing and scaling systems—potentially leveraging blockchain-enabled micropayments and complex attribution standards—to automatically reward each entity contributing information for the successful tasks of agents.

------ Elizabeth Harkavy, a16z crypto investment team

9. Privacy Will Be the Most Important Moat in Cryptocurrency

Privacy is a key feature of the world's financial shift to on-chain. It is also a feature that almost all existing blockchains lack. For most chains, privacy is an afterthought.

But now, privacy itself is significant enough to distinguish one chain from all others. Privacy also does something more important: it creates a lock-in effect for the chain; if you will, a kind of privacy network effect. Especially in a world where merely competing on performance is no longer sufficient.

Thanks to cross-chain bridge protocols, moving from one chain to another is trivial as long as everything is public. But once you make things private, it is no longer the case: bridging tokens is easy, bridging secrets is hard. When entering and exiting private areas, there is always the risk that those monitoring the chain, memory pools, or network traffic may figure out who you are. The boundaries between private chains and public chains—even between two private chains—will leak various metadata, such as correlations of transaction times and sizes, making it easier to track someone.

Compared to many indistinguishable new chains (where fees may drop to zero due to competition—the block space has become the same everywhere), blockchains with privacy can have stronger network effects. The reality is that if a "general-purpose" chain does not have a thriving ecosystem, killer apps, or unfair distribution advantages, there is almost no reason for anyone to use it or build on it—let alone remain loyal to it.

When users are on public blockchains, they can easily transact with users on other chains—what chain they join does not matter. When users are on private blockchains, on the other hand, the chain they choose becomes more important, as once they join, they are less likely to leave and risk exposure. This creates a winner-takes-all dynamic. And because privacy is essential for most real-world use cases, a small number of privacy chains may capture a large portion of the cryptocurrency market.

------ Ali Yahya, a16z crypto general partner

10. The (Near) Future of Messaging is Not Just Quantum-Resistant, It is Decentralized

As the world prepares for quantum computing, many cryptography-based messaging applications (Apple, Signal, WhatsApp) have taken the lead and done an excellent job. The problem is that each major messaging application relies on private servers run by a single organization that we trust. These servers are easy targets for government shutdowns, backdoors, or coercion to hand over private data.

What good is quantum encryption if a country can shut down your server; if a company has the keys to a private server; or even if a company owns a private server? Private servers require "trust me"—but no private server means "you don't have to trust me." Communication does not need to be mediated by a single company. Messaging needs open protocols where we do not have to trust anyone.

The way to achieve this is through decentralized networks: no private servers. No single application. All open-source code. Top-notch encryption—including against quantum threats. With an open network, no single person, company, nonprofit, or nation can strip us of our ability to communicate. Even if a country or company shuts down one application, 500 new versions will pop up the next day. Shut down one node, and the blockchain and other economic incentives will immediately replace it with a new node.

When people own their messages like they own their money—i.e., own their private keys—everything changes. Applications may come and go, but people will always maintain control over their messages and identities; end users can now own their messages, even if they do not own the application.

This is not just about being quantum-resistant and encrypted; it is about ownership and decentralization. Without these two, all we are doing is building unbreakable encryption that can still be shut down.

------ Shane Mac, Co-founder and CEO of XMTP Labs

11. Secrets as a Service

Behind every model, agent, and automation lies a simple dependency: data. But today, most data pipelines—whether inputting or outputting data from models—are opaque, mutable, and un-auditable. This is fine for some consumer applications, but many industries and users (like finance and healthcare) require companies to keep sensitive data private. This is also the main barrier currently hindering institutions from tokenizing real-world assets.

So how do we enable secure, compliant, autonomous, and globally interoperable innovation while maintaining privacy? There are many ways, but I will focus on data access control: who controls sensitive data? How does it move? Who (or what) can access it?

Without data access control, anyone wanting to keep data confidential must use centralized services or build custom setups—this is not only time-consuming and expensive but also hinders the features and benefits of fully unlocking on-chain data management for traditional financial institutions. Moreover, as agent systems begin to autonomously browse, trade, and make decisions, users and institutions across industries will need cryptographic guarantees, rather than "best-effort trust."

This is why I believe we need "secrets as a service": a new technology that provides programmable, local data access rules; client-side encryption; and decentralized key management, enforcing who can decrypt what under what conditions, and for how long… all enforced on-chain. Combined with verifiable data systems, "secrets" can then become part of the internet's basic public infrastructure—rather than application-level patches (where privacy is often an afterthought)—thus making privacy a core infrastructure.

------ Adeniyi Abiodun, Chief Product Officer and Co-founder of Mysten Labs

12. From "Code is Law" to "Specs are Law"

Recent DeFi hacks have struck well-tested protocols that have strong teams, diligent audits, and years of production experience. These events highlight a disturbing reality: today's standard security practices remain largely heuristic and case-by-case.

To mature, DeFi security needs to shift from a Bug model to design-level attributes and from a "best effort" to a "principled" approach:

• In the static/pre-deployment aspect (testing, auditing, formal verification), this means systematically proving global invariants rather than verifying manually selected local invariants. Several teams are building AI-assisted proof tools that can now help write specifications, propose invariants, and share the burden of much of the manual proof engineering that has been prohibitively expensive in the past.

• In the dynamic/post-deployment aspect (runtime monitoring, runtime enforcement, etc.), those invariants can become real-time "guardrails": the last line of defense. These guardrails will be directly encoded as runtime assertions that every transaction must satisfy.

So now, we no longer assume every bug is caught; instead, we enforce key security properties of the code itself, automatically rolling back any transactions that violate them.

This is not just theoretical. In practice, nearly every attack to date has triggered one of these checks during execution, potentially stopping the hacker. So the once-popular "code is law" has evolved into "spec is law": even new types of attacks must satisfy the same security properties that maintain system integrity, thus leaving the remaining attacks either trivial or extremely difficult to execute.

------ Daejun Park, a16z crypto engineering team

13. Prediction Markets Becoming Bigger, Broader, and Smarter

Prediction markets have entered the mainstream, and in the coming year, as they intersect with cryptocurrencies and AI, they will only become bigger, broader, and smarter—while also presenting new significant challenges for builders.

First, more contracts will be listed. This means we will be able to access real-time odds, not just for major elections or geopolitical events but for a variety of nuanced outcomes and complex, cross-cutting events. As these new contracts surface more information and become part of the news ecosystem, they will raise important societal questions: about how we balance the value of this information and how we can better design them to be more transparent and auditable—something made possible by cryptocurrencies.

To handle the larger volume of contracts, we need new methods to reach consensus on the truth to resolve contracts. Centralized platform resolutions are important, but controversial cases like the "Zelensky Suit Market" and "Venezuela Election Market" demonstrate their limitations. To address these edge cases and help prediction markets expand into more useful applications, new decentralized governance and LLM oracles can help determine the truth of disputed outcomes.

AI opens up new possibilities for prediction markets—including agents automatically betting based on real-time data, synthesizing new contracts, and dynamically adjusting markets based on agent behavior. This will make prediction markets smarter, more responsive, and potentially unlock new use cases, such as real-time risk assessment, automated hedging, and AI-driven predictions.

However, as scale increases, builders will face new challenges: ensuring market resistance to manipulation, handling the complexities of dispute resolution, and balancing information transparency with privacy. These challenges will drive innovations such as advanced cryptographic proofs and decentralized arbitration systems.

------ Andy Hall, a16z crypto research advisor and Stanford University professor of political economy

14. The Rise of "Staked Media"

The cracks in traditional media models—and their so-called objectivity—have been apparent for some time. The internet has given everyone a voice, and more operators, practitioners, and builders are now speaking directly to the public. Their viewpoints reflect their stakes in the world, and counterintuitively, audiences often respect them not despite these stakes, but because of them.

What is new here is not the rise of social media, but the arrival of cryptographic tools that allow people to make publicly verifiable commitments. As AI makes generating infinite content (whether real or fictional, and capable of claiming any viewpoint or persona) cheap and easy, relying solely on what people (or bots) say feels insufficient.

Tokenized assets, programmable locks, prediction markets, and on-chain histories provide a more solid foundation of trust: commentators can make arguments and prove they put their money where their mouth is. Podcasters can lock tokens to prove they will not engage in pump-and-dump schemes. Analysts can tie predictions to publicly settled markets, creating auditable records.

This is what I envision as the early form of "Staked Media": a media form that not only embraces the idea of "skin in the game" but also provides proof. In this model, credibility comes neither from pretending to be above it all nor from baseless claims; rather, it comes from having stakes that you can make transparent and verifiable commitments to. Staked media will not replace other forms of media; it will complement what we already have. It provides a new signal: not just "trust me, I am neutral," but "this is the risk I am willing to take, and here is how you can check if I am telling the truth."

------ Robert Hackett, a16z crypto editorial team

15. Cryptographic Technology Provides a New Primitive Beyond Blockchain

For years, SNARKs—cryptographic proofs that allow you to verify computations without re-executing them—have primarily been a blockchain technology. Their overhead has been too high: proving a computation can take up to 1,000,000 times more work than simply running it. This is worthwhile when distributed among thousands of validators, but impractical anywhere else.

That is about to change. By 2026, the overhead of zkVM provers will drop to about 10,000 times, and memory usage will be in the hundreds of megabytes—fast enough to run on mobile phones and cheap enough to be ubiquitous. Here is a reason why 10,000 times might be a magical number: high-end GPUs have parallel throughput about 10,000 times higher than laptop CPUs. By the end of 2026, a single GPU will be able to generate proofs of CPU executions in real-time.

This could unlock a vision from old research papers: Verifiable Cloud Computing. If you are going to run CPU workloads in the cloud anyway—because your computation is not heavy enough to require GPU acceleration, or you lack expertise, or for legacy reasons—you will be able to obtain cryptographic proofs of correctness at reasonable overhead. Provers are already GPU-optimized; your code does not need to be.

------ Justin Thaler, a16z crypto research team, Georgetown University computer science associate professor

16. Trading is Just a Waystation, Not the Endpoint of Crypto Business

It seems that today every well-run crypto company (except for stablecoins and some core infrastructure) has transformed or is transforming into a trading platform. But if "every crypto company becomes a trading platform," where does that leave everyone else? So many participants doing the same thing will eat away at the public's mind share, leaving only a few big winners. This means that those companies that pivot too quickly to trading miss the opportunity to build a more defensible, lasting business.

While I have great sympathy for all those founders trying to turn their financial situations around, chasing immediate product-market fit (PMF) comes at a cost. This issue is particularly pronounced in crypto, where the unique dynamics around tokens and speculation may lead founders down the path of instant gratification in their quest for PMF… if you will, it is a kind of "marshmallow test" (delayed gratification test).

Trading itself is not wrong—it's an important market function—but it does not have to be the final destination. Those founders who focus on the product part of product-market fit may ultimately become the bigger winners.

------ Arianna Simpson, a16z crypto general partner

17. Unlocking the Full Potential of Blockchain

For the past decade, one of the biggest barriers to building blockchain networks in the U.S. has been legal uncertainty. Securities laws have been overly stretched and selectively enforced, forcing founders into a regulatory framework built for "companies" rather than "networks." For years, mitigating legal risks has supplanted product strategy; engineers have been forced to yield to lawyers.

This dynamic has led to many strange distortions: founders have been told to avoid transparency; token distributions have become legally arbitrary; governance has turned into a charade; organizational structures have also been optimized for legal cover. Tokens have been designed to avoid economic value/no business model. Worse still, non-compliant crypto projects often outperform those sincere builders.

However, the potential for regulatory clarity around crypto market structure—the likelihood of the government passing such legislation is greater than ever—has the potential to eliminate all these distortions in the coming year. If passed, this legislation would incentivize transparency, establish clear standards, and replace "enforcement roulette" with clearer, more structured paths for financing, token issuance, and decentralization. Following the GENIUS Act, the proliferation of stablecoins has exploded; legislation around crypto market structure will be a more significant shift, but this time aimed at networks.

In other words, such regulation would enable blockchain networks to operate like networks—open, autonomous, composable, trustlessly neutral, and decentralized.

------ Miles Jennings, a16z crypto policy team and general counsel