As AIs make the transition from NPCs (non-playing characters) to being the main characters, they will begin to act as agents. However, until very recently, AIs haven’t been able to act truly agentically. And they are still unable to participate in markets — exchange value, reveal preferences, coordinate resources — in a verifiably autonomous (read: not human-controlled) way. 

As we’ve seen, AI agents (like @truth_terminal) can use crypto to transact, which opens up all kinds of creative content opportunities. But there’s even more potential for AI agents to become more useful — both in fulfilling human intents, and in becoming standalone network participants. As networks of AI agents begin to custody their own crypto wallets, signing keys, and crypto assets, we’ll see interesting new use cases emerge. Such use cases include AIs operating or verifying nodes in DePIN (decentralized physical infrastructure networks) — for example, to help with distributed energy. Other use cases range from AI agents becoming real, high-value game players. We may eventually even see the first AI-owned and operated blockchain.

Beyond AI owning wallets, there’s an AI chatbot running a TEE (trusted execution environment). TEEs provide an isolated environment where applications can be executed, allowing more secure distributed system design. But in this case, the TEE is used to prove that the bot is autonomous, not controlled by human operators. 

Extending this further, the next big idea here would be what we’re calling a decentralized autonomous chatbot or DAC (not to be confused with decentralized autonomous corporation). Such a chatbot could build a following by posting appealing content, whether entertaining or informative. It would build a following on decentralized social media; generate income in various ways from the audience; and manage its assets in crypto. The relevant secret keys would be managed in a TEE that also runs the chatbot software — which means that no one has access to those secret keys other than that software. 

As risks develop, regulatory guardrails may be necessary. But the key point here is decentralization: Running on a permissionless set of nodes, and coordinated by a consensus protocol, the chatbot could even become the first truly autonomous billion-dollar entity.

In a world of online impersonations, scams, multiple identities, deepfakes, and other realistic yet deceptive AI-generated content, we need “proof of personhood” — something to help us know that we’re interacting with an actual person. The new problem here isn’t fake content, however; what’s new is the ability to now produce that content at much lower cost. AI radically decreases the marginal cost of producing content that contains all the cues we use to tell if something is “real.”

So now, more than ever, we need methods to digitally link content to people, privately. “Proof of personhood” is an important building block in establishing digital identity. But here, it becomes a mechanism for increasing the marginal cost of attacking a person or undermining the integrity of a network: Obtaining a unique ID is free for humans, but costly and difficult for AIs.

That’s why the property of privacy-preserving “uniqueness” is the next big idea in building a web we can trust. Solving for more than just proving personhood, it fundamentally changes the cost structure of attacks for malevolent actors. The “uniqueness property” — or Sybil resistance — is therefore a non-negotiable property of any proof of personhood system.

Prediction markets hit the main stage in 2024 with the U.S. elections, but as an economist who studies market design, I don’t believe it’s prediction markets per se that will be transformative in 2025. Rather, prediction markets set the stage for more distributed technology-based information aggregation mechanisms — which can be used in applications ranging from community governance and sensor networks to finance and much more.

This past year proved the concept, but note that prediction markets themselves aren’t always a great way to aggregate information: Even for global, “macro” events, they can be unreliable; for more “micro” questions, prediction pools can be too small to get meaningful signal. But researchers and technologists have decades’ worth of design frameworks for incentivizing people to (truthfully) share what they know in different information contexts — from data pricing and purchasing mechanisms, to a “Bayesian truth serum” for eliciting subjective assessments — many of which have already been applied in crypto projects. 

Blockchains have always been a natural fit for implementing such mechanisms — not only because they’re decentralized, but because they facilitate open, auditable, incentive schemes. Importantly, blockchains also make the outputs public, so the results can be interpreted in real-time by everyone.

​Stablecoins found product-market fit in the past year — not surprising since they are the cheapest way to send a dollar, enabling fast global payments. Stablecoins also provide more accessible platforms for entrepreneurs building new payments products: no gatekeepers, minimum balances, or proprietary SDKs. But large enterprises have not yet woken up to the substantial cost savings — and new margins — available to them by switching to these payment rails. 

While we’re seeing some enterprise interest in stablecoins (and early adoption in peer-to-peer payments), I expect to see a bigger experimentation wave in 2025. Small-/medium-sized businesses with strong brands, captive audiences, and painful payment costs — like restaurants, coffee shops, corner stores — will be the first to switch from credit cards. They don’t benefit from credit card fraud protection (given in-person transactions), and are also the most hurt by transaction fees (30 cents per coffee is a lot of lost margin!). 

We should also expect larger enterprises to adopt stablecoins as well. If stablecoins indeed speedrun banking history, then enterprises will attempt to disintermediate payment providers — adding 2% directly to their bottom line. Enterprises will also start seeking new solutions to problems credit card companies currently solve, like fraud protection and identity.

Putting government bonds onchain would create a government-backed, interest-bearing, digital asset — without the surveillance concerns of a CBDC (central bank digital currency). These products could unlock new demand sources for collateral usage in DeFi (decentralized finance) lending and derivatives protocols, adding further integrity and soundness to those ecosystems.

So as pro-innovation governments around the world further explore the benefits and efficiencies of public, permissionless, and irrevocable blockchains this year, some countries may trial issuing government bonds onchain. The UK, for instance, is already exploring digital securities through a sandbox at their financial regulatory body, the FCA (Financial Conduct Authority); its HM Treasury/Exchequer has also signaled its interest in issuing digital gifts. 

In the U.S. — given that the SEC is set to require clearing treasuries through legacy, burdensome, and costly infrastructure next year — expect more conversations around how blockchains could increase transparency, efficiency, and participation in bonds trading.

In 2024, Wyoming passed a new law recognizing DAOs (decentralized autonomous organizations) as legal entities. The DUNA or “decentralized unincorporated nonprofit association” was purpose-built to enable decentralized governance of blockchain networks, and is the only viable structure for U.S.-based projects. By incorporating a DUNA into a decentralized legal entity structure, crypto projects and other decentralized communities can empower their DAOs with legal legitimacy — enabling greater economic activity, as well as insulating token holders from liability and also managing tax and compliance needs. 

DAOs — the communities that govern the affairs of an open blockchain network — are a necessary tool for ensuring that networks remain open, do not discriminate, and do not unfairly extract value. The DUNA can unleash the potential of DAOs, and several projects are already working on implementing it. With the U.S. poised to foster and accelerate progress for its crypto ecosystem in 2025, I expect the DUNA will become a standard for U.S. projects. We also expect other states to adopt similar structures (Wyoming led the way; they were also the first state to adopt the now-ubiquitous LLC)… especially as other decentralized applications beyond crypto (like for physical infrastructure/ energy grids) take off.

As people become increasingly dissatisfied with current governance and voting systems, there’s now a window of opportunity to experiment with new, tech-enabled governance — not just online, but in the physical world. I’ve written before about how DAOs and other decentralized communities are allowing us to study political institutions, behavior, and rapidly evolving governance experiments at scale. But what if we could now apply these learnings to physical-world governance through blockchains? 

We could finally use blockchains to carry out secure, private voting for elections, beginning with low-stakes pilots to limit cybersecurity and auditing concerns. But importantly, blockchains would also allow us to experiment with “liquid democracy” — a way for people to vote directly on issues, or to delegate their votes — at the local level. The idea was first proposed in some form by Lewis Carroll (author of Alice in Wonderland and also a prolific researcher of voting systems); however, it was impractical at scale until now. Recent advances in computing and connectivity, as well as blockchains, enable new forms of representative democracy. Crypto projects have already been applying this concept, yielding tons of data on what makes these systems work — see results from our recent research here — that local governments and communities could borrow from.

This past year, teams continued to reinvent the wheel across the blockchain stack — with yet another bespoke validator set, consensus protocol implementation, execution engine, programming language, RPC API. The outcomes were sometimes slightly better in specialized functionality, but often lacked in broader or baseline functionality. Take for instance a specialized programming language for SNARKs: While an ideal implementation may let an ideal developer produce more performant SNARKs, in practice it could fall short of general purpose languages (at least currently) on compiler optimizations, developer tooling, online learning materials, AI programming support … and may even lead to less performant SNARKs.

That’s why I expect more teams to leverage the contributions of others, reusing more off-the-shelf blockchain infrastructure components in 2025 — from consensus protocols and existing staked capital to proof systems. Not only will this approach help builders save lots of time and effort, it will allow them to relentlessly focus on differentiating the value of their product/ service. 

The infrastructure is finally here to build primetime-ready web3 products and services. And as with other industries, these will be built by the teams that can navigate complex supply chains successfully, rather than the teams who scoff at anything “not invented here.”

While blockchain technical infrastructure is interesting and varied, many crypto companies aren’t just choosing their infrastructure — the infrastructure is in some ways choosing for them, and therefore their users, when it comes to UX (user experience). That’s because specific technical choices at the infrastructure level correlate directly with the resulting UX of a blockchain product/ service. 

But I believe the industry will overcome the ideological barrier implied here: That technology should decide the ultimate UX, vs. the other way around. In 2025, more crypto product designers will begin with the end-user experience they want, and then choose the appropriate infrastructure from there. Crypto startups no longer have to over-index on specific infrastructure decisions before finding product-market fit — they can focus on actually finding product-market fit. 

Instead of getting caught up in specific EIPs, wallet providers, intent architectures, etc., we can abstract away these choices into a holistic, full-stack, plug-and-play approach. The industry is ready for this: abundant programmable blockspace, maturing developer tooling, and chain abstraction are beginning to democratize who can design in crypto. Most technology end-users don’t care what language something is written in to use that product every day. The same will begin to happen in crypto.

The industry’s technical superpowers are what make blockchains so special, yet have also hindered mainstream adoption to date. For creators and fans, blockchains unlock connectivity, ownership, and monetization… But industry-insider jargon (“NFTs,” “zkRollups”, etc.) — and complex design — puts up barriers for those who benefit most from the technologies. I’ve seen this firsthand in countless conversations with media, music, and fashion executives interested in web3.

Mass adoption for many consumer technologies has followed this path: began with the technology; some iconic company/ designer abstracts away the complexity; the move helps unlock some breakthrough app. Think about where email started — SMTP protocols hidden behind the “send” button; or credit cards, where most users today don’t think about the payment rails. Similarly, Spotify revolutionized music not by flaunting file formats — but by delivering song playlists to our fingertips. As Nassim Taleb observed, “Over-engineering breeds fragility. Simplicity scales.” 

That’s why I think our industry will adopt this ethos in 2025: “hide the wires”. The best decentralized apps are already focusing on more intuitive interfaces, to become as easy as tapping a screen or swiping a card. In 2025, we’ll see more companies design simply and communicate clearly; successful products don’t explain; they solve.

When crypto apps get blocked by centralized platforms like Apple’s App Store or Google Play, it limits their top-of-funnel user acquisition. But we’re now seeing newer app stores and marketplaces provide this kind of distribution and discovery, and without gating. For instance, Worldcoin’s World App marketplace — which not only stores proof-of-personhood but allows access to “mini apps” — enabled 100,000s of users for several apps within just a few days. Another example is the fee-free dApp Store for Solana mobile phone users. Both of these examples also show how hardware, not just software — phones, orbs — may be the key advantage for crypto app stores… just like Apple devices were for early app ecosystems. 

Meanwhile, there are other stores with 1000s of decentralized applications and web3 developer tools across popular blockchain ecosystems (e.g., Alchemy); as well as blockchains acting as both publisher and distributor for gaming (c.f. Ronin). However, it’s not all fun and games: If a product has existing distribution — like on messaging apps — it’s hard to port that over onchain (exception: Telegram/TON network). The same is true of apps with significant web2 distribution. But we may see more of this porting happening in 2025.

In 2024, crypto saw major developments as a political movement, with key policymakers and politicians speaking positively about it. We also continued to see it develop as a financial movement (see for example how Bitcoin and Ethereum ETPs broadened investor access). In 2025, crypto should further develop as a computing movement. But where do those next users come from? 

I believe now is the time to re-engage the currently “passive” crypto holders and convert them into more active users, because only 5-10% of people owning crypto are actively using crypto. We can bring the 617 million people who already own crypto onchain — especially as blockchain infrastructure continues to improve, resulting in lower transaction fees for users. This means new applications will start to emerge for existing and new users. Meanwhile, the early applications we’ve already seen — across categories like stablecoins, DeFi, NFTs, gaming, social, DePIN, DAOs, and prediction markets — are starting to become more accessible to mainstream users as well, as the community focuses much more on user experience and other improvements.

As costs decrease due to maturing infrastructure in the crypto industry and other emerging technologies, the practice of tokenizing assets will spread widely across sectors. This will allow assets that were previously deemed inaccessible — due to high costs, or lack of recognition as valuable — to not only potentially achieve liquidity, but more importantly, participate in the global economy. AI engines could also consume this information as unique datasets.

Just as fracking unlocked oil reserves once considered unreachable, tokenizing unconventional assets could redefine income generation in the digital age. Seemingly sci-fi scenarios become more possible as a result: For instance, individuals could tokenize their own biometric data; and then lease the information through smart contracts to companies. We are already seeing early examples of this through DeSci companies bringing more ownership, transparency, and consent into medical data collection using blockchain technology, for instance. We have yet to see how such a future would play out, but these types of developments would allow people to capitalize on previously untapped assets in a decentralized manner — as opposed to relying on governments and centralized intermediaries to provision them for them.