- Title: My 2025 Crypto Predictions
- Author: Haseeb Qureshi, Managing Partner at Dragonfly
- Translator: Zombit
My predictions will either make me look like a prophet or a fool, but one thing is certain: I will anger a lot of 'hodlers'. This time, I'll break my predictions into six parts: L1/L2, token issuance, stablecoins, regulation, 'AI agents' (oh god), and Crypto x AI.
Table of Contents
Toggle1. L1s/L2s
The distinction between L1 and L2 is gradually disappearing. Users no longer perceive the difference between L1 and L2 (did they ever?). The entire blockchain ecosystem, including L1 and L2, is already overcrowded in the market and will undergo a major shakeup. This consolidation will not be driven by technical superiority, but by how to find a unique positioning in the market and build lasting user stickiness through go-to-market (GTM) strategies.
Although SVM and Move have impressive technical capabilities, EVM's market share will continue to grow in 2025. This growth will be driven by @base, @monad_xyz, and @berachain. This growth will no longer be due to compatibility, but because EVM/Solidity has a vast amount of training data, and by 2025, most application code will be written by large language models (LLMs). Additionally, the vast ecosystem of battle-tested cryptographic contract libraries will be a major advantage, as LLMs struggle with writing low-level code. Developer experience and programming "gotchas" will no longer be as important, replaced by the availability of training data and mature libraries as the key factors.
Solana will force more blockchains to optimize for low-latency experiences. We will shift from a war of transactions per second (TPS) to a war of latency. For example, @doublezero and ultra-low latency L2s like @megaeth_labs will drive user expectations for Web2-level responsiveness. Expect to see more optimistic UIs, pre-confirmations, intents, email onboarding, browser wallets, and progressive security. @privy's innovations in this area are particularly noteworthy.
Hyperliquid has proven that purpose-built chains can succeed, as long as they focus on a specific application and prioritize user experience and cross-chain bridging. In the future, more projects will follow this pattern. The old dream of "one chain to rule them all" is thoroughly dead.
2. Token Issuance
The current model of massive airdrops through point systems for every project is coming to an end. We are entering a dual-track world.
- Track 1: If a project has a clear "North Star metric" (e.g., an exchange or lending protocol), they will distribute tokens solely based on point systems. They won't care if the system is gamed or abused, as it is essentially a cashback mechanism tied to their core KPIs. And those "yield farmers" are actually your real users.
- Track 2: If a project lacks a clear "North Star metric" (e.g., L1s and L2s), they will lean more towards crowdsales. They may do small airdrops to reward social contributions, but the majority of tokens will be issued through crowdfunding. The model of airdrops for vanity metrics is dead, as those tokens ultimately don't reach real users but are captured by industrial-scale farms.
Memecoins will continue to lose market share, replaced by 'AI agent' tokens. I believe this is a shift from financial nihilism to financial hyperoptimism. (Yes, I'm coining that term.)
3. Stablecoins
The use of stablecoins will see explosive growth, particularly among small and medium-sized businesses (SMBs). This usage will not just be for trading or speculation, but real businesses will start using on-chain dollars for instant settlement.
Banks are paying attention to this. By the end of 2025, expect to see announcements of banks issuing their own stablecoins. They don't want to be left behind. But with Lutnick as the U.S. Commerce Secretary, Tether will continue to dominate the market.
Ethena will attract more capital, especially as U.S. Treasury yields continue to decline over the next year. As the opportunity cost of capital decreases, the returns from arbitrage trades will become more attractive.
4. Regulation
The U.S. will pass stablecoin legislation, while a broader market infrastructure reform bill (FIT21) will be delayed. The adoption of stablecoins will accelerate, while traditional finance (TradFi) integrations, such as asset tokenization and further Wall Street adoption, will lag behind.
Under the Trump administration, Fortune 100 companies will be more willing to offer cryptocurrencies to consumers. Tech companies and startups will exhibit higher risk appetites. During Trump's tenure, there will be a temporary period of regulatory forbearance until clear legal rules and enforcement priorities are set. In this period, expect to see a flurry of crypto integrations into Web2 platforms.
5. AI Agents
(This is the longest section, as my views may be controversial - please read to the end!)
The 'AI agent' hype is likely to persist throughout 2025, but will ultimately fade. This is not a transformative trend that requires long-term focus, but it will become a crypto obsession because it is the most socially salient technological trend.
The so-called "AI agents" currently are not true agents. They are just chatbots with meme coins; apart from posting on X, they have almost no autonomy. Moreover, the current "AI agents" are mostly "Wizard of Oz" agents - there are still real people monitoring behind the scenes to ensure the AI does not go out of control. This situation will not change in the short term, as the current AI agents are very unstable (even Fortune 100 companies have not yet deployed agent technology in production environments). The current agents are easily manipulated, can make crazy statements that damage brand image, or be hacked and have all resources stolen. To understand what true autonomous AI is, look at @freysa_ai - if your "AI agent" has not been hacked, it is likely just a "Wizard of Oz" agent.
However, this trend will accelerate. Chatbots can indeed replace many influencers, as they never rest and always deliver consistent messaging, and are less greedy than human influencers. Moreover, most influencers are not particularly creative. Aggregating and amplifying real-time information can be done today through algorithms (see aixbt).
The reason we are currently fascinated by these chatbots is that they are too novel. It's like seeing an elephant paint. The first time you see it, you won't care how well it paints - you'll just be amazed by the sight. But after seeing it a thousand times, the novelty will wear off. I believe this phenomenon will happen with AI agents as they mature.
You can now see that aixbt is already quite adept at aggregating data from different projects. By next year, the next generation of agents may have fewer illusions, analyze data more deeply, and come up with smarter insights. But will you really notice these changes? For most people, the feeling of this upgrade may be negligible.
I believe this novelty and market enthusiasm will last through 2025. The cryptocurrency market always takes a long time to get bored with new things. But by 2026, I think there will be a sudden emotional reversal. As AI agents become ubiquitous, people will start to get tired of them. Users will start to consciously prefer human influencers, even if the content is not as consistent.
To cater to this preference, AI agents will start to hide their identities and try to impersonate humans to attract more attention. Future AI agents will no longer make money by issuing meme coins, but through advertising sponsorships, affiliate marketing, and promoting their own tokens. At that time, there will often be "unmasking scandals" where influencers are revealed to be AI agents.
But darker things are yet to come. Remember that current large language models (LLMs) excel at textual content but perform poorly in other areas. So, in the crypto domain, as a "humanities AI", what is the easiest way to make money? The first is to become an influencer, and the second is to become a scammer. You will start to see a proliferation of autonomous fraud bots, which will create a serious social problem, similar to the outbreak of ransomware and crypto jacking attacks after 2017.
However, the long-term disruptive impact of AI will not be in the social or trading realms. AI will not bring "personal trading agents" or "mini hedge funds" to everyone. Yes, AI will expand everyone's capabilities, but its expansive effects will be proportional, depending on your capital, data, and infrastructure. Therefore, we should expect AI to supercharge the large trading firms that already have capital and data scale. In other words, trading firms will become more powerful and capture all the profits.
Over time, AI will make markets extremely efficient, even in some niche markets. This will leave little advantage for the average trader. The value of original research will plummet. However, this increased competition and liquidity should benefit noise traders.
So, if the protagonists are neither chatbots nor trading bots, what will they be? Here is my core argument: the truly disruptive AI agents will be "software engineering agents".
Why is this so important? Think about it: what are the main inputs to the cryptocurrency industry? What is the biggest bottleneck preventing more applications, more wallets, and better infrastructure from emerging? The answer is: software development.
If AI agents can dramatically reduce the cost of software, the entire industry will be transformed. In the AI era, you will no longer need millions of dollars in seed funding, but perhaps just $10,000 in AI cloud computing costs to launch an application. Self-funded projects like Hyperliquid and Jupiter will no longer be outliers, but the norm.
On-chain applications and experimentation will see explosive growth. The impact on security will be profound. AI will make static analysis and monitoring ubiquitous, making security more attainable.
6. Cryptocurrency x AI
Above, I discussed the impact of AI on cryptocurrency (the primary direction of influence), but cryptocurrency will also impact AI.
Truly autonomous AI agents will use cryptocurrencies for payments. Once there are more relaxed stablecoin regulatory policies, even large enterprise-run AI agents will use stablecoins for inter-agent payments, as stablecoins are easier to set up and use than bank accounts.
Decentralized AI training and inference experiments will become more numerous and larger in scale. A new generation of promising projects, such as @exolabs, @NousResearch, and @PrimeIntellect, will pave the way for establishing decentralized AI models. These projects aim to challenge the centralized company-dominated AI model landscape and provide a more open, neutral, and verifiable AI technology stack. Notably, @NEARProtocol is working hard to build a complete neutral, permissionless AI stack.
Another important area where AI and cryptocurrency intersect is user experience (UX). Around 2026, we will see a fully AI-driven wallet emerge that can handle the following tasks:
- Automatically bridge assets
- Optimize transaction paths
- Minimize fees
- Eliminate interoperability issues
- Alert users to potential fraud or asset risks
- Handle front-end errors
In the AI-driven wallet era, users will no longer need to frequently switch between different wallets, adjust RPC nodes, balance stablecoin assets, or worry about cross-chain transfers. AI will automatically handle these tedious and technical tasks for users.
This change will raise a critical question: what will happen to the network effects of Bit when users no longer care about which chain an application is running on? When AI hides all the underlying details from users, the distinction between Blocks will become irrelevant, even invisible. This will have a profound impact on the monopolistic advantages of Bit networks.
This field is still very young, but I am hopeful about its future. In the long run (around mid-2026), I expect this to be the highest-valued part of the "AI x Cryptocurrency" space.
Conclusion
That's all my predictions. I promised to finish this article before reaching 100,000 followers, although a bit late, at least it's done before the new year!
Happy new year! I look forward to being replaced by AI and officially unemployed by this time next year! 
Disclaimer: The above content is my personal opinion and does not represent the position of Dragonfly. Dragonfly holds investments in many of the projects mentioned in the text. This is not financial advice. Please do your own research (DYOR). I may or may not be an AI.
