Original author: a16z
Original compilation: KarenZ, Foresight News
Editor's note: Based on the opinions of a16z partners in the fields of artificial intelligence, American Dynamism, bio/health, Crypto, enterprise, fintech, gaming, and infrastructure, a16z has released a comprehensive list of "Big Ideas" that technology builders may need to address in the coming year. The following is an overview of some highlights that a16z's crypto partners are excited about for the coming year. For an outlook on policy, regulation, etc. for 2025, please refer to this article published in November 2024.
TL;DR
An AI needs its own wallet to act autonomously;
Decentralized autonomous Bots emerge;
As more people use AI, we will need unique identity proofs;
From prediction markets... towards comprehensive information efficient aggregation;
More and more companies will accept stablecoin payments;
Countries explore tokenizing government bonds;
"DUNA" (Decentralized Unincorporated Nonprofit Associations) will see broader adoption in the US blockchain ecosystem;
Online liquid democracy moves offline;
Builders will reuse infrastructure, rather than just reinventing infrastructure;
Crypto companies will focus on the end-user experience, not just the underlying infrastructure;
"Hidden rails" help usher in killer Web3 apps;
The crypto industry has its own app stores and discovery features;
Crypto asset owners become crypto asset users;
Various industries may start tokenizing "non-standard" assets.
An AI needs its own wallet to act autonomously
As AI transitions from non-player characters (NPCs) to protagonists, they will start to act as agents. However, until recently, AIs have not been able to truly act autonomously. And they still cannot participate in markets in a verifiable autonomous way (i.e., not under human control) - exchanging value, expressing preferences, coordinating resources.
As we've seen, AI agents (like @truth_terminal) can transact using cryptocurrencies, opening up opportunities for all sorts of generative content. But AI agents have even greater potential to become more useful - both helping humans achieve their intents, and becoming independent network participants. When AI agent networks start to custody their own crypto wallets, signing keys, and crypto assets, we'll see some interesting new use cases emerge. These could include AIs operating or validating nodes in a DePIN (Decentralized Physical Infrastructure Network) - for example, assisting distributed energy projects. Other use cases could involve AI agents becoming true, high-value game players. We may even eventually see the first AI-owned and operated blockchain.
—Carra Wu
Decentralized autonomous Bots emerge
Beyond AIs having wallets, there are also AI chatbots running in a trusted execution environment (TEE). A TEE provides an isolated environment in which an application can execute, allowing for the design of more secure distributed systems. But in this case, the TEE is used to prove the bot is autonomous, not controlled by a human operator.
Taking this further, the next big idea here is what we call decentralized autonomous chatbots (DACs, not to be confused with DAOs). These chatbots can build fan communities by publishing engaging content (whether entertainment or informational). They'll build fan bases on decentralized social media; earn revenue in various ways from their audiences; and manage their assets in crypto. The relevant keys will be managed in the TEE running the chatbot software - meaning no one else can access those keys besides that software.
As the risks evolve, some regulatory "guardrails" may be needed. But the key point here is decentralization: running on a permissionless set of nodes, coordinated by a consensus protocol, these chatbots may even become the first truly autonomous, multi-million dollar entities.
—Dan Boneh, Karma, Daejun Park, and Daren Matsuoka
As more people use AI, we will need unique identity proofs
In a world filled with online impersonation, scams, multiple identities, deepfakes, and other highly convincing yet deceptive AI-generated content, we urgently need a "persona certificate" to help us verify that we are interacting with real people. However, the new problem here is not the false content itself, but the entirely new capability to produce content containing all the cues we use to judge whether something is "real" at an extremely low marginal cost.
Therefore, we need methods to privately associate content with individuals in a digital way more than ever before. The "persona certificate" is an important building block for constructing digital identities. But here, it becomes a mechanism to raise the marginal cost of attacking individuals or undermining the integrity of the network: for humans, obtaining a unique identity identifier is free, but for AI, it is costly and difficult.
As a result, the privacy-preserving "uniqueness" property becomes a crucial next big idea for building a trustworthy network. It not only solves the problem of proving personhood, but fundamentally changes the cost structure for malicious actors to mount attacks. Therefore, the "uniqueness property" - or Sybil resistance - is an indispensable property for any persona certification system.
—Eddy Lazzarin
From prediction markets... towards comprehensive information efficient aggregation
The 2024 US election brought prediction markets into the mainstream, but as an economist who studies market design, I believe that 2025 will see a revolution not in prediction markets themselves, but in the broader set of technology-enabled distributed information aggregation mechanisms that prediction markets have laid the groundwork for, applicable to community governance, sensor networks, finance, and more.
The past year has already demonstrated the concept, but it's worth noting that prediction markets themselves are not always the best method for information aggregation: even for global "macro" events, they may be unreliable; for more "micro" questions, the prediction pools may be too small to yield meaningful signals. However, researchers and technologists have decades of design frameworks for incentivizing people to truthfully share what they know in different information environments - from data pricing and purchase mechanisms, to "Bayesian truth serum" for eliciting subjective assessments, many of which have already been applied in crypto projects.
Blockchains have been a natural choice for implementing these mechanisms, not only because they are decentralized, but because they facilitate open, auditable incentive structures. Crucially, blockchains also make the output results public, so everyone can interpret the results in real-time.
—Scott Duke Kominers
More and more companies will accept stablecoin payments
Stablecoins have found product-market fit over the past year - unsurprisingly, as they are the cheapest way to send US dollars, enabling fast global payments. Stablecoins also provide a more accessible platform for entrepreneurs to build new payment products: no "gatekeepers", minimum balances, or proprietary software development kits (SDKs) required. However, large enterprises have yet to realize the massive cost savings and new revenue opportunities they can unlock by shifting to these payment rails.
While we've seen some enterprises express interest in stablecoins (and early adoption in peer-to-peer payments), I expect to see a bigger wave of experimentation in 2025. Small and medium-sized businesses with strong brands, captive audiences, and high payment costs (like restaurants, coffee shops, corner stores) will be the first to ditch credit cards. Since they can't benefit from credit card fraud protection (given in-person transactions), and the transaction fees hurt them the most (30 cents per $5 coffee - that's a huge profit margin hit!)
We should also expect to see large enterprises adopt stablecoins. If stablecoins do accelerate the history of banking, enterprises will try to bypass payment providers and directly add 2% to their bottom line. Enterprises will also start seeking new solutions to the problems currently solved by credit card companies, like fraud protection and identity verification.
—Sam Broner
Countries Explore Blockchain-Based Government Bonds
Blockchain-based government bonds will create a government-backed, interest-bearing digital asset without the surveillance concerns of Central Bank Digital Currencies (CBDCs). These products could open up new sources of demand for collateral use in DeFi lending and derivatives protocols, adding more integrity and robustness to these ecosystems.
As such, as innovation-friendly governments further explore the benefits and efficiencies of public, permissionless and immutable blockchains this year, some countries may attempt to issue blockchain-based government bonds. For example, the UK has already explored digital securities through the Financial Conduct Authority's (FCA) sandbox mechanism; its Treasury has also expressed interest in issuing digital gifts.
In the US - given the Securities and Exchange Commission's (SEC) plans to require the clearing of Treasuries through traditional, cumbersome and expensive infrastructure next year - there are expected to be more discussions around how blockchain can improve bond trading transparency, efficiency and participation.
——Brian Quintenz
"DUNA" to See Wider Adoption in the US Blockchain Ecosystem
In 2024, Wyoming passed a new law recognizing Decentralized Autonomous Organizations (DAOs) as legal entities. "DUNA" or the "Decentralized Unincorporated Nonprofit Association" is designed specifically to enable decentralized governance of blockchain networks, and is the only viable structure for native US projects. By incorporating DUNA into decentralized legal entity structures, crypto projects and other decentralized communities can imbue their DAOs with legal legitimacy, enabling more economic activity while shielding token holders from liability and managing tax and compliance requirements.
DAOs - the communities that govern open blockchain networks - are a necessary tool to ensure networks remain open, non-discriminatory, and do not extract value unfairly. DUNA can unlock the potential of DAOs, and multiple projects are already working to implement it. As the US prepares to nurture and accelerate its crypto ecosystem in 2025, I expect DUNA to become the standard for US projects. We also anticipate other states adopting similar structures (with Wyoming leading the way), especially as decentralized applications beyond cryptocurrencies (such as physical infrastructure/energy grids) flourish.
—Miles Jennings
Online Liquid Democracy Moves Offline
As frustration with current governance and voting systems grows, there is now an opportunity to experiment with new, technology-driven modes of governance - not just online, but in the physical world as well. I've previously written about how DAOs and other decentralized communities can allow us to study political institutions, behaviors, and rapidly evolving governance experiments at scale. But what if we could apply those learnings to governance in the real world through blockchain?
We could ultimately leverage blockchain for secure, private electoral voting, starting with low-risk pilots to alleviate cybersecurity and auditing concerns. But crucially, blockchain will also enable us to experiment with "liquid democracy" at the local level - a way for people to directly vote on issues or delegate their votes. This idea was initially proposed by Lewis Carroll (the author of Alice in Wonderland, and a prolific researcher of voting systems); however, large-scale application has long been impractical. Advances in computation, connectivity, and blockchain have now made new forms of representative democracy possible. Crypto projects have already applied this concept, generating a wealth of data on how these systems operate. Local governments and communities can draw on our recent research findings.
——Andrew Hall
Builders Will Reuse Infrastructure, Not Just Reinvent It
Over the past year, many teams have continued to reinvent the wheel within the blockchain stack - another custom set of validators, consensus protocol implementations, execution engines, programming languages, and RPC APIs. These efforts sometimes marginally improve on specific functionalities, but often lack broader or baseline capabilities. Take the case of domain-specific programming languages for SNARKs: while an ideal implementation may allow developers to generate higher-performance SNARKs, in practice it may lag behind general-purpose languages in compiler optimizations, developer tooling, online learning materials, AI programming support, and even SNARK performance itself.
As such, I expect 2025 to see more teams leveraging others' contributions, reusing more off-the-shelf blockchain infrastructure components - from consensus protocols and existing staked capital, to proof systems. This approach will not only save builders immense time and effort, but also allow them to relentlessly focus on differentiating the value of their products/services.
Ultimately, the necessary infrastructure to build quality Web3 products and services is already in place. Like other industries, these products and services will be built by teams that can successfully navigate complex supply chains, not those who scoff at "not invented here".
——Joachim Neu
Crypto Companies Will Start from the Desired User Experience, Not Let Infrastructure Dictate It
While the blockchain technology infrastructure is fascinating and diverse, many crypto companies have not been choosing their own infrastructure - in some ways, the infrastructure is choosing the UX (user experience) for them and their users. This is because specific technology choices at the infrastructure layer directly impact the final UX of a blockchain product or service.
But I believe the industry will overcome this implicit ideological barrier: that technology should dictate the final UX. By 2025, more crypto product designers will start from the desired end-user experience they want, and then select the appropriate infrastructure from that. Crypto startups no longer need to over-focus on specific infrastructure decisions before finding product-market fit - they can focus on actually finding product-market fit.
We don't have to get bogged down in specific EIPs, wallet providers, intent architectures, etc. - we can abstract these choices into a holistic, full-stack, plug-and-play approach. The industry is ready for this: a rich programmable blockchain landscape, increasingly mature developer tools, and chain abstraction is starting to democratize crypto design. Most end-users don't care what language their daily-used products are written in. The same will start to happen in crypto.
——Mason Hall
"Hidden Rails" Will Help Usher in Killer Web3 Apps
What makes blockchain unique is its technical superpowers, but this has also hindered mainstream adoption so far. For creators and fans, blockchain unlocks interoperability, ownership, and monetization... yet industry jargon (like Non-Fungible Tokens, zkRollups, etc.) and complex designs have erected barriers for those who could benefit most from these technologies. I've experienced this firsthand in countless conversations with media, music, and fashion executives interested in Web3.
Many breakout consumer technologies have followed a similar path: starting with the technology itself; then having iconic companies or designers abstract the complexity; this paves the way for breakthrough applications. Think of email's evolution - the SMTP protocol is hidden behind a "Send" button; or credit cards, where most users don't think about the payment rails. Similarly, Spotify didn't revolutionize the music industry by flaunting file formats, but by pushing playlists to our fingertips. As Nassim Taleb said, "Overdesign breeds fragility, simplicity scales."
So I believe our industry will adopt the principle of "hidden rails" by 2025. The best decentralized apps are already focusing on more intuitive interface design, to make interactions as simple as tapping a screen or swiping a card. By 2025, we'll see more companies prioritize clean design and clear communication; successful products won't require explanation - they'll just solve problems directly.
——Chris Lyons
The Crypto Industry Finally Has Its Own App Store and Discovery
When crypto apps are blocked by centralized platforms like the Apple App Store or Google Play, it limits their ability to reach users at the top of the funnel. But now, we're seeing some new app stores and marketplaces that provide this distribution and discovery function without gatekeepers. For example, the World App marketplace from Worldcoin not only stores personal identity proofs but also allows access to "mini-apps," attracting hundreds of thousands of users to multiple apps in just a few days. Another example is the free decentralized app (dApp) store for Solana phone users. These examples suggest that hardware (like phones, orbs) may be a key advantage for crypto app stores, just as Apple devices played a role in the early app ecosystem.
Meanwhile, other stores host thousands of decentralized apps and Web3 development tools (like Alchemy) across popular blockchain ecosystems. Blockchains are also playing dual roles as publishers and distributors in gaming (e.g. Ronin). But it's not just entertainment and games: if a product already has existing distribution channels, like through messaging apps, migrating it on-chain can be challenging (exception: Telegram/TON network). The same is true for apps with large distribution channels in Web2. But we may see more of these migrations happen by 2025.
—Maggie Hsu
Crypto Owners Becoming Crypto Users
In 2024, the crypto space saw major political developments, with key policymakers and politicians making positive statements about it. We also continue to see it evolve as a financial movement (e.g. how Bitcoin and Ethereum ETPs are expanding investor access). In 2025, crypto should further develop as a computing movement. But where will these new users come from?
I believe it's time to re-engage the currently "passive" crypto holders and convert them to more active users, as only 5-10% of crypto holders are actively using crypto. We can convert the 617 million people who already hold on-chain crypto into more active users - especially as blockchain infrastructure continues to improve, reducing transaction costs for users. This means new applications will emerge for both existing and new users. And the early applications we've seen - like stablecoins, DeFi, Non-Fungible Tokens, gaming, social, DePIN, DAOs, and prediction markets - are also starting to become more suitable for mainstream users as the community focuses more on user experience and other improvements.
—Daren Matsuoka
Industries May Start Tokenizing "Unconventional" Assets
As the crypto industry and other emerging technologies mature, and infrastructure costs decrease, the practice of asset tokenization will spread across industries. This will allow assets that were previously inaccessible due to high costs or lack of recognition as valuable to not only achieve liquidity, but more importantly, participate in the global economy. AI engines can also consume this information as unique data sets.
Just as fracking technology unlocked previously unrecoverable oil reserves, tokenizing unconventional assets may redefine revenue generation in the digital age. So seemingly science fiction scenarios become more plausible: for example, individuals could tokenize their biometric data and then lease that information to companies via smart contracts. We've already seen some early examples through decentralized science (DeSci) companies, which are using blockchain technology to bring more ownership, transparency and consent to medical data collection. We don't know how this will unfold, but these types of developments will enable people to leverage previously untapped assets in a decentralized way, rather than relying on governments and centralized intermediaries to provide them.
—Aaron Schnider
