Author: Haseeb, Partner at Dragonfly; Translated by 0xjs@Golden Finance
I either look like a prophet or I look like an idiot, but one thing is for sure: I will annoy a lot of people with bags.
I will divide my predictions for 2025 into six parts: L1/L2, Token Issuance, Stablecoins, Regulation, 'AI Agents,' and Crypto x AI.
1. L1/L2
The distinction between L1 and L2 is fading. Users can no longer perceive differences between L1 and L2 (did they ever?). The blockchain space (combined L1 and L2) has become too crowded and is due for a major reshuffle. Integration will no longer be about technological advantage — but about having a unique niche market and building stickiness through GTM.
Despite the strong capabilities of SVM and Move, EVM market share is expected to grow significantly by 2025. This growth will be driven by Base, Monad, and Berachain. It's no longer about compatibility, but rather because EVM/Solidity has more training data, and LLMs will write most of the application code in 2025. Having a battle-tested library of deep cryptographic contracts will also become a watershed, as LLMs struggle to write low-level code. In the era of LLM development, the importance of DevEx and footguns is dwarfed by training data and reliable libraries.
Solana will force more blockchains to optimize for low latency. We will shift from a TPS war to a latency war — just like the infrastructure of doublezero and ultra-low latency L2s like MegaETH will drive user expectations for web2 responsiveness. Looking forward to more people embracing optimistic user interfaces, pre-confirmations, intents, email onboarding, in-browser wallets, and progressive security. The emergence of SwapPrivy.
Hyperliquid has shown that when specialized chains focus on specific applications and prioritize user experience and easy bridging, they can work. More projects will follow this model. The old dream of one chain to rule them all has been shattered.
2. Token Issuance
The era of massive airdrops through loyalty programs is over. We are moving towards a dual-track world.
Track One: If a project has a clear North Star metric, like an exchange or lending protocol, they will purely distribute tokens based on points. They won't care if they are being farmed or played — they are essentially distributing tokens as a rebate/discount for the core KPI of the protocol, while the airdrop farmers are still your actual users.
Track Two: Projects without a clear North Star metric (like L1 and L2) will turn to crowdfunding. They may conduct small-scale airdrops to reward social contributions, but most tokens will be distributed through crowdfunding. Airdrops done for vanity metrics are already dead. Those do not truly flow to users but to industrialized airdrop farmers.
Memecoins will continue to compete for market share with 'AI agent' coins. I think this is a shift from financial nihilism to financial over-optimism. (Yes, I coined this term.)
3. Stablecoins
The use of stablecoins will surge, especially among small and medium enterprises. It’s not just about trading and speculation — real businesses will start using on-chain dollars for instant settlements.
Banks have noticed this: announcements for banks to issue stablecoins are expected by the end of 2025. They don't want to fall behind. But especially in the case of Lutnick serving as the Secretary of Commerce, Tether will still maintain its first position.
Ethena is expected to absorb more capital, especially in the coming year as Treasury yields continue to decline. When the opportunity cost of capital decreases, it makes the underlying trading yield more attractive.
4. Regulation
The U.S. will pass stablecoin legislation while broader market infrastructure reforms (FIT21) will be postponed. The adoption of stablecoins will accelerate, while Wall Street adoption, asset tokenization, and other TradFi integrations will lag behind.
Under Trump's leadership, Fortune 100 companies will be more willing to offer cryptocurrency to consumers, with tech companies and startups showing a higher risk appetite. Trump's inauguration will create a clear regulatory celebration until explicit rules and enforcement priorities are established. During this time, active expansion of cryptocurrency integration with Web2 platforms is expected.
5. AI Agents (this is the longest section as my thoughts here may be controversial — please read to the end!)
The 'AI agent' craze may last until 2025. But it will eventually fade. This isn't the long-term destruction to be wary of with AI, but it will become a focal point in CT because it is the most social.
Current AI agents are not genuinely intelligent agents. These are chatbots attached to memecoins; they have almost no intelligence other than posting on Twitter. The current 'AI agents' are mostly 'Wizard of Oz' agents — a human behind the scenes ensuring the AI doesn't go off track. This won't change quickly, as current agents are too poor (even Fortune 100 companies have yet to use agents in production). Current agents are easily manipulated to say crazy things that damage their brands or can be jailbroken to steal all their resources. See Freysa for a glimpse of what true autonomous AI looks like — if your favorite AI hasn't been jailbroken, it’s because it's a Wizard of Oz AI.
That said, I believe this trend will accelerate. Chatbots can indeed replace many KOLs, as chatbots never sleep, they are always passing information, and they are not as greedy as human KOLs. Furthermore, most KOLs are not very creative. Even today, real-time information aggregation/amplification can easily be replaced by algorithms (see @aixbt_agent).
These chatbots are very appealing to us right now because they are very novel. It's like seeing an elephant paint. The first time you see it, you don't care whether the painting is good — it looks spectacular. But by the 1000th time, the novelty fades. I believe as these chatbots reach a stable state, this will start to happen.
You can see this today through aixbt — it has become quite adept at aggregating data about different projects. By next year and the next generation of agents, perhaps aixbt's illusions will be a bit less, more in-depth, and have smarter insights. But how much will you notice? For most, it might feel the same.
I believe this novelty and market desire will last until 2025. It takes time for cryptocurrency to get tired of shiny things. But by 2026, I think a sudden reversal will occur. Chatbots will become ubiquitous, and people will lose interest in them. The sentiment will reverse. Stories of their favorite human KOLs losing their livelihoods will spark a class consciousness. Users will start to discriminate against human KOLs, even if their content is less consistent.
In response to this pro-human bias, chatbots will start to hide that they are AI, trying to pass as human to attract more attention from the market. Future chatbots won’t make money through memecoins like today, but through sponsorships, affiliate links, and tokens they own, just like human KOLs. KOLs will frequently be accused of being chatbots, and you will see scandals exposing AI. All of this will get very strange.
But there is a darker side. Keep in mind that LLMs are currently excellent wordcel (note: people with superb writing skills and linguistic intelligence), but not good enough in other respects. What’s the best way to make money as a wordcel in cryptocurrency? First, by becoming an influencer, of course, but closely followed by becoming a scammer. You’ll start to see a surge in autonomous scam bots. These bots will explode in number, comparable to ransomware and crypto hijacking after 2017. This is expected to become a real social issue.
However, while chatbots are likely to remain a focus in 2025, the long-term disruption of AI will not appear at the social level.
No, it won’t appear in the trading realm either. AI will not provide everyone with their own 'trading agent' or micro hedge funds. Yes, AI will scale everyone, but they will scale personnel proportionally based on capital, data, and infrastructure. Therefore, you should expect AI to supercharge existing trading firms that have capital scale and data scale. In other words, trading firms will become better at making money. It will also break down hierarchies between trading firms (most trading firms will become quite good, as everyone will have access to cloud-based IQ 150 quantitative analysts).
Over time, AI will make the market extremely efficient, even in smaller niche markets, leaving ordinary traders with little advantage, even if they have their homemade assistant AI. The value of original research will decline significantly. Nonetheless, the increase in competition and liquidity should be a boon for those of us injecting noise into the market. (This also means liquidity for Polymarket!)
So, if the big news isn’t chatbots or trading bots, what is it? This is my core argument, for some reason, almost no one talks about it: truly influential AI agents will be software engineering agents.
Why is this a big deal? Ask yourself: what are the main inputs of our industry? What expensive inputs are hindering more applications, more wallets, better infrastructure, better everything? The answer is software. If AI agents lead to a collapse in software prices, that will change everything.
In the post-AI era, you won’t need to raise millions of dollars for seed rounds but will be able to launch applications using $10,000 of AI cloud computing. Self-funding projects like Hyperliquid and Jupiter will go from exception to norm. The number of on-chain applications and experiments will absolutely surge. For a software-driven industry, this deflationary shock will lead to an on-chain renaissance.
The implications for security are profound. AI-driven static analysis and monitoring will become ubiquitous, making it easier for everyone to achieve security. These AIs will be fine-tuned on EVM/Solidity or Rust codebases and trained on large security audit and attack medium databases. They will undergo RL training in simulated adversarial blockchain environments. I am increasingly convinced that, in terms of security, AI tools will ultimately favor defenders over attackers. You will have AIs constantly red teaming contracts while other AIs reinforce them, formally verify their properties, and hone their skills in incident response and remediation.
Meanwhile, of course, AI-style memecoins can be traded. But truly intelligent agents will have a greater impact than tweeting and promoting their own tokens.
6. True Crypto x AI
Above, I detailed the impact of AI on cryptocurrency (which is the main direction of impact), but cryptocurrency will also impact AI.
Truly autonomous agents will use cryptocurrency to pay each other. This will be especially true once there are loose regulations on stablecoins — you might even see large companies running AI agents using stablecoins for inter-agent payments, as they are easier to set up than bank accounts.
We will also see more and larger-scale decentralized training and inference experiments. Promising new projects like Exo Labs, Nous Research, and Prime Intellect will pave the way for a true alternative to centralized training and corporate self-operating models. The NEAR protocol is also making a strong effort to create a full-stack, trust-neutral, and permissionless AI stack.
Another place where crypto and AI intersect is user experience. Post-AI wallets will be transformed — AI wallets should be able to handle bridging, optimize transaction routes, minimize fees for you, mask interoperability issues or frontend errors, and help you avoid obvious scams or rug pulls. You won’t have to switch between multiple different wallets, nor will you need to change RPC or rebalance your stablecoins. AI will handle all of this for you. This might not become reliable enough to change the user experience of cryptocurrency until 2026. But when all of this arrives, what impact will it have on blockchain network effects? What happens when users no longer care — or even experience — which chain applications exist on?
This field is still young, but I hope we will see it take off soon. In the long run (say mid-2026), I expect most of the market cap for 'AI x Crypto' to be located here. -
That's all my predictions. I hope that by this time next year, everyone can stop working!
