Breakthrough and Rebuilding: A Comprehensive Outlook on the 2025 Crypto World
TechFlow Selected TechFlow Selected
Breakthrough and Rebuilding: A Comprehensive Outlook on the 2025 Crypto World
Asset issuance remains the main theme.
Navigating Web3 tides with focused insightsAuthor: Zeke, Researcher at YBB Capital
Introduction
From the inscriptions boom to the election of the first crypto-friendly president, 2024 is drawing to a close. This year saw an unusually peculiar "bull market" in crypto—altcoins underperformed, memes dominated attention, and ultimately everything converged back into BTC. While there were setbacks and frustrations, crypto has indeed been moving toward a more positive trajectory. As we approach 2025, several key trends are worth watching. In this article, we offer a brief outlook for next year based on recent developments.
1. On AI
At present, chain abstraction projects often prioritize conceptual perfection so heavily that technical implementation becomes excessively complex, ultimately harming user experience. Projects incorporating intent-centric architectures face similar challenges. Whether built on centralized models (e.g., Telegram bots), structured designs combining on-chain and off-chain preprocessing, or distributed frameworks like Solver + Executor, these intent-based systems share common flaws. Users still need substantial DeFi knowledge; intents must be expressed clearly, simply, and precisely. When faced with complex or ambiguous requests, current intent systems struggle significantly, offering only limited functionality. Since Paradigm introduced the concept in mid-2023, intent-centric projects have largely produced more hype than tangible results, doing little to onboard new users or lower operational barriers. That said, given the clear market demand illustrated by Ethereum's Layer2 evolution, both abstraction and intent layers remain urgently needed.
Looking back at recent Layer2 developments, leading projects such as OP Superchain have formed growing L2 alliances. Zksync’s Elastic Chain and Arbitrum Orbit are also expected to follow suit, creating their own ecosystems. These alliances could eventually achieve seamless interoperability via solutions like interoperability clusters, addressing liquidity fragmentation and poor cross-chain compatibility across Ethereum’s L2 landscape. The competition among dozens of chains may thus evolve into a battle between major blocs. From a broader perspective, as the crypto market improves, new-architecture L2s like Movement and Fuel are racing to launch mainnets to capture scarce altcoin liquidity. For projects outside the top tier, fragmentation and lack of interoperability continue to worsen—some even use VMs with incompatible wallet plugins. Beyond onboarding new users, even regular blockchain users find the L2 ecosystem overwhelmingly complex, severely hindering development of non-financial applications.
Ethereum cannot achieve mass adoption without ecosystem alignment. An ecosystem requiring users to be semi-experts will never reach “Mass Adoption.” Solana and Ton’s逆势 growth this year highlights the importance of lowering entry barriers and delivering consistent, Web2-like user experiences. More directly, beyond marketing, both ecosystems mainly succeeded by simplifying asset issuance and making chain interactions nearly imperceptible. Thus, Ethereum needs a comprehensive, experience-first solution—but due to the core developers’ open ethos, enforcing ecosystem-wide standardization isn’t feasible.
I believe the first viable solution lies in AI browser agents. Ever since ChatGPT emerged, many envisioned AI revolutionizing app interaction by operating across multiple apps to form a super-app. Take travel planning: upon receiving a request, an AI could automatically book tickets, design itineraries, arrange meals, and schedule time. If equipped with long-term memory, it could further refine future plans based on past behavior.
Now, Google is preparing to launch Project Mariner—an AI browser agent powered by Gemini. As demonstrated by Google Labs director Jaclyn Konzelmann, once installed as a Chrome extension, the agent opens a chat window on the right side of the browser. Users can assign tasks like “create a shopping cart from this grocery list.” The AI then navigates to a grocery platform, adds items to the cart, and proceeds to checkout. After verification, the user completes payment (the agent itself lacks payment privileges). OpenAI is set to release a similar product next month.
Notably, while Google’s Project Mariner remains limited to select testers, I’ve already tested analogous agents developed for general users within crypto. After several hours of trial, these agents achieved about 60–70% accuracy in executing complex or vague intents (though cursor movements were relatively slow), autonomously completing actions such as token swaps across DEXs on various blockchains or bridging assets from Ethereum to L2s. My role was merely stating the intent and entering my wallet password.
Of course, these prototypes still rely on centralized model APIs. What synergies can emerge with crypto? Beyond serving as a superior intent-layer solution, AI browser agents will likely catalyze breakthroughs in AI wallets, decentralized compute, and decentralized data projects in 2025.
Consider a simple question: why has it taken until now to realize the vision of AI agents despite years of rapid AI advancement? Revisiting OpenAI’s progress reveals that language models typically advance faster than image-generation models because the internet serves as a vast corpus of text for training. The primary constraints for language models are compute power and energy. Agents, however, require extensive human labeling and feedback, along with costly inference processes. Crypto naturally enables labor mobilization through incentives. Within this economic model, upper-tier participants can provide labeled data and feedback in exchange for tokens via decentralized mechanisms, while lower tiers integrate decentralized compute and data networks. Once trained, models can be embedded into wallets and DeFi protocols via SDKs, forming truly intelligent AI wallets—and closing the loop. Other AI agent concepts can similarly benefit, as any Web3-compatible agent will require compute, labeling, and feedback to grow.
2. Stablecoins
Stablecoins remain a fiercely contested battleground and one of the highest-barrier sectors in crypto. Their application value is now widely recognized even beyond the industry. This year, traditional finance giants entered the space: PayPal launched PYUSD, BlackRock partnered with Ethena on USDb, and VanEck introduced AUSD targeting Argentina, Southeast Asia, and other regions.
As Tether and Circle strengthen their dominance, new stablecoin issuers are diverging into two paths. First, fiat-collateralized stablecoin providers are focusing on emerging markets—particularly Latin America—and niche use cases. Second, algorithmic stablecoins are increasingly adopting low-risk financial products as underlying assets, such as those discussed in our previous piece on Ethena and Usual. Looking ahead, more delta-neutral stablecoins will compete for short-side liquidity on CEXs, expanding hedging assets from BTC and ETH to higher-risk, less liquid public chain tokens to serve underserved markets. As for Usual-style stablecoins backed by short-to-medium-term U.S. Treasuries, innovation will likely center on protocol token mechanics and yield distribution rather than diversifying RWA collateral types—given Treasuries remain optimal. However, compared to constrained CEX liquidity, these instruments face less competition and offer greater scalability.
Overall, stablecoin development is trending toward more stable underlying assets and decentralized governance. Still, I hope to see fully decentralized, non-overcollateralized stablecoin protocols emerge in 2025.
3. Payments
With increasing regulatory acceptance and adoption of national stablecoins, the downstream payments sector is becoming a new frontier. High-throughput, low-gas heterogeneous chains like Solana and Move will serve as foundational infrastructure for payment applications. Traditional payments represent a mature, hyper-competitive red ocean—what transformative advantages can blockchain bring? Two commonly cited benefits stand out: first, optimizing cross-border payments by eliminating pre-funding requirements, enabling faster, cheaper, and easier remittances, thereby resolving trillions of dollars tied up in legacy systems. Second, serving emerging markets—a point I've previously emphasized. In regions across Africa, Asia, and Latin America, stablecoins have proven valuable by providing strong financial inclusion, helping citizens in developing nations combat hyperinflation caused by unstable governments and allowing participation in global financial activities and subscriptions to cutting-edge digital services.
Lily Liu, manager at the Solana Foundation, introduced the concept of “PayFi” at the 7th EthCC conference, expanding imagination around blockchain-powered payments. PayFi revolves around two core ideas. First is instant settlement—T+0 or even multiple daily settlements—eliminating delays and complexities inherent in traditional finance and dramatically accelerating capital turnover. Second is “Buy Now, Pay Never” (BNPL): for example, a user deposits $50 into a lending protocol to buy a $5 coffee. Once accrued interest reaches $5, it covers the cost, unlocking and returning the principal to the user’s account.
This concept opens numerous possibilities: startups could fundraise securely and transparently via PayFi; travelers could exchange currencies without relying on physical financial institutions; users could freely control timing of payments (delaying receipts to earn interest or paying early for discounts). Revenue models would also diversify. Beyond earning yield by depositing stablecoins into lending protocols, I believe users should be able to freely switch between different stablecoin types. As new stablecoins proliferate, individuals could dynamically choose the most suitable option based on risk tolerance, simultaneously capturing protocol token rewards and higher yields. For DeFi, if such a payment system achieves mainstream adoption, its growth potential would be unimaginably vast.
4. DEXs
In Section 1, we discussed L2 fragmentation and lack of interoperability. This developmental path introduces another issue: oversupply of blockspace, where infrastructure outpaces dApp development. This imbalance will lead to natural elimination of many long-tail chains over the coming years—an acute challenge for Ethereum if DA pricing fails to generate positive feedback loops for L2s.
Reviewing this cycle’s outperforming blockchains, most relied on strong communities, ecosystems, and promotional advantages funneled into asset issuance platforms, driving explosive TVL growth. Not every L2 can replicate this attention economy. Lack of killer applications remains a real challenge heading into next year. Following the trend, potential solutions include meeting future demands for AI agents—as mentioned earlier—or near-term opportunities in on-chain orderbook DEXs, privacy tools, payment stacks, and decision-making utilities.
I personally believe on-chain orderbook DEXs will become the next generation standard. AMM evolution shows ever-increasing technical complexity with diminishing returns—a point we’ve made in prior analyses of Uniswap. However, for L2s, performance and gas limitations remain significant hurdles. Advancements in matching algorithms and gas-efficient designs will be critical to overcoming these challenges.
5. Asset Issuance Remains the Main Theme
From 2023 to today—from inscriptions to AI meme platforms—the ability to issue assets has defined the past year’s hotspots. Extending this timeline further, asset issuance has arguably been the *only* enduring theme in crypto since the ICO era. Only the packaging and entry barriers have changed. Positively, user-driven speculation has accelerated infra and DeFi innovation. As these technologies gain recognition, blockchain inches closer to mainstream integration. Negatively, the game has grown purer and more absurd—lowered issuance thresholds make this dark forest even more dangerous. Today, anyone can launch a full-scale zero-sum game with just a click, an image, and a few words. Can we redirect this force toward more constructive ends—driving industry progress through speculation?
For instance, some AI meme projects are evolving into practical agents, moving beyond the nonsensical AI assistants of early versions. Recently popular DeSci (“Decentralized Science”) might be called “ICO for research.” Though currently meme-driven, in the long run, leveraging blockchain’s strengths could make traditional scientific research more transparent, accessible, easily funded, and collaborative. Still, actual implementation and evolution remain uncertain.
Similar to DeSci, I’ve previously explored ideas in GameFi articles—such as how blockchain could support indie games struggling with funding and talent shortages. A core problem with blockchain fundraising is the extremely low barrier to asset issuance, minimal restrictions, and excessive fundraising power (also attributable to ultra-low onboarding friction). How can rules enforce responsible fund usage and compel teams to continuously build genuinely valuable products? This is a crucial question we must address.
Let speculators speculate, let builders build—that is the prerequisite for blockchain’s continuous advancement. In 2025, we may witness more iterations of “ICO,” but I hope this speculative feast gives rise to the next “DeFi Summer.”
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
@YBBCapital
