Author | @archetypevc
Compiled by | Odaily Planet Daily (@OdailyChina)
Translator | DingDang (@XiaMiPP)
Editor's Note: As we step into 2026, the focus of the industry narrative is quietly shifting. Structural adjustments are underway in capital, infrastructure, user demands, and content distribution methods. Archetype has gathered insights from multiple industry researchers to outline several trends that are currently forming and may evolve into key investment windows in the coming year.
This article is compiled by Odaily Planet Daily, aiming to present readers with Archetype's perspective on the underlying driving forces of the next phase of crypto applications and the structural inflection points most worthy of attention in 2026.
When "Chains Built for Apps" Finally Make Sense: The Best Time for Appchains Has Arrived
by Aadharsh Pannirselvam
In a nutshell: Blockchains that are deliberately designed, finely crafted, and built and optimized based on underlying primitives for specific application scenarios will truly explode in the next year or two.
The developers, users, institutions, and capital recently flooding onto the chain are very different from those in previous cycles. They have their own distinct cultures and preferences (i.e., their definition of "user experience"), which are often more important than abstract concepts like "decentralization" or "censorship resistance." Sometimes, these needs align with existing infrastructure; other times, they require completely different chain structures.
User experience is particularly crucial for crypto applications like Blackbird and Farcaster that target non-professional users and are abstracted. Designs that would have been considered heretical three years ago, such as node deployment, a single sequencer, or customized databases, are now the most reasonable solutions. The same applies to stablecoin chains or trading scenarios (like Hyperliquid, GTE) that are highly dependent on latency and price precision.
But not all new applications are like this.
Today, an important counterforce is growing: the preference for "privacy" from both institutions and the general user base. The experiential demands of different applications vary greatly, so their underlying chain architectures should not be uniform.
The good news is that assembling a chain from scratch that meets application needs is far simpler than it was two years ago. It's now more like "building a custom PC."
You can pick every hard drive, fan, and power cable yourself—but most of the time, this isn't necessary. You can simply choose from a range of highly compatible, customizable preset configurations, like selecting a Digital Storm or Framework. If you need a degree of personalization, you can add your own components on top of their proven base. This ensures stability while maintaining high flexibility.
Similarly, when applications can freely assemble and adjust primitives like consensus mechanisms, execution layers, data storage, and liquidity structures, they can build chain forms with different cultural characteristics, allowing the application's own "experience definition" to be natively supported, thus forming a differentiated competitive advantage. This difference is like the distinct yet shared underlying commonalities between a ToughBook, ThinkPad, desktop tower, and MacBook.
More importantly, each component becomes a tunable "knob," without worrying about affecting the whole system or being constrained by the upgrade pace of the parent protocol.
With Circle's acquisition of Informal Systems' Malachite, the importance of independently mastering "customized block space" has become an industry trend. In the coming year, I look forward to seeing more applications quickly build their own chains and block space using "default templates" and underlying modules provided by Commonware, Delta, etc., much like using HashiCorp or Stripe Atlas.
Ultimately, this will allow applications to truly control their own cash flow and build moats through chain structures that better fit the user experience.
Prediction Markets Will Continue to Innovate (But Only Some Players Will Break Through)
By Tommy Hang
In this cycle, prediction markets have undoubtedly been one of the best-performing application categories. The total trading volume of fully on-chain prediction markets has repeatedly hit new highs, with weekly trading volumes consistently breaking $2 billion, clearly indicating that this category has taken a key step towards becoming a mass-consumer product.
Amid the hype, a large number of new projects attempting to challenge Polymarket and Kalshi have emerged. But identifying "true innovation" amidst the noise is key to judging which projects are worth focusing on in 2026.
From a market structure perspective, I am most interested in solutions that reduce spreads and enhance open interest. Although market creation is still permissioned and curated, liquidity in prediction markets remains thin for both market makers and traders. The best paths include optimizing routing system, introducing different liquidity models, and improving collateral efficiency based on products like lending.
Categorized trading volume is a key factor determining platform success. For example, over 90% of Kalshi's trading volume in November came from sports markets, indicating that some platforms are naturally better suited to compete for specific liquidity. In contrast, Polymarket's trading volume in crypto and political markets is 5–10 times higher than Kalshi's.
Of course, on-chain prediction markets still have a long way to go before achieving true "mass adoption." For instance, the 2025 Super Bowl alone generated $23 billion in off-the-books betting volume in a single day, more than 10 times the current daily trading volume of all on-chain markets combined.
Closing this gap requires teams that can truly solve the underlying challenges of prediction markets. I will continue to watch these players next year.
Agentic Curators: Expanding DeFi's Next Layer with "Intelligent Agents"
By Eskender Abebe
Today's DeFi "asset screening and risk configuration layer" (curation layer) tends towards two extremes: either fully algorithmic (fixed rate curves, preset rebalancing rules) or fully reliant on humans (risk committees, active managers).
Agentic curators represent a third path: AI agents (LLM + tools + recurrent scheduling) managing the risk and strategies of vaults, lending markets, and structured products, not by executing fixed rules, but through "reasoning"—logical deduction about risk, yield, and position strategy.
Take Morpho's market as an example: designing an attractive yield product requires defining collateral asset rules, LTV limits, and various risk parameters. This is currently a human bottleneck, and intelligent agents can scale this process. Soon, you will see intelligent agent curators competing directly with algorithmic models and human managers.
So, when will DeFi's "Move 37" arrive?
Many fund managers have extreme attitudes towards AI: either believing LLMs will automate all trading desks, or that they will immediately collapse in real markets. But both overlook the real structural change: AI agents combine emotionless execution, systematic strategy, consistent policy constraints, and reasoning ability, whereas humans are noisy and pure algorithms are too fragile. In the future, LLMs will act like "architects," designing risk frameworks, strategy constraints, and portfolio structures, while truly high-frequency and sensitive computations are still executed by deterministic code.
When the cost of deep reasoning drops to "pennies," the strongest vaults will no longer be managed by the smartest people, but driven by the strongest computing power.
Short-Form Video Will Become the New "Traffic Entry Point"
By Katie Chiou
Short-form video is becoming the default interface for "global users to discover and purchase content." TikTok Shop's GMV exceeded $20 billion in the first half of 2025, nearly doubling year-on-year, acclimating global users to "watch and buy," where entertainment is the storefront.
Instagram is shifting Reels from a defensive product to a core revenue engine. This format brings more exposure and is accounting for an increasingly larger share of Meta's projected ad revenue for 2025. Whatnot has proven that real-time, personalized sales methods can achieve conversions at speeds unattainable by traditional e-commerce.
The logic is simple: real-time viewing leads users to make decisions faster. Every swipe is a potential purchase point. Therefore, platforms are quickly blurring the line between "recommendation feed" and "payment process"; the information feed itself is the new point of sale, and every content creator is a distribution channel.
AI will accelerate this trend: reducing video production costs, increasing content quantity, allowing creators and brands to test ideas in real-time. More content, more touchpoints, the more incentive platforms have to optimize conversion efficiency every second.
And crypto fits this trend perfectly: faster content requires faster, cheaper, programmable payment channels. As shopping becomes frictionless and directly embedded into the content itself, we need a system that can settle micro-payments, programmatically distribute and allocate revenue, and track the contributions of various parties in complex influence chains. Cryptocurrency is built for such processes; it's hard to imagine how the era of hyper-scale streaming-native commerce would develop without it.
Blockchain Will Drive New AI Scaling Laws
By Danny Sursock
Over the past few years, the AI narrative has been dominated almost entirely by giants and unicorns, with decentralized innovators long overlooked. However, away from the spotlight, several crypto-native teams have made astonishing progress in "decentralized training and inference," and have moved from whiteboard demos to real testing and production environments.
Today, teams like Ritual, Pluralis, Exo, Odyn, Ambient, Bagel are ready to "take the main stage" and welcome their golden age. This new generation of competitors is expected to have an explosive, orthogonal impact on the fundamental development trajectory of artificial intelligence.
Distributed training environments are breaking through existing scaling limits; asynchronous communication and parallelization schemes are being proven feasible in training tasks of real scale.
New consensus mechanisms and privacy technologies are making "verifiable" and "confidential" inference a reality.
A new generation of chain architectures combines "truly intelligent contract systems" with more general computing models, enabling AI agents to use crypto assets as a unified medium of exchange, forming a complete autonomous computing closed loop.
The foundational work is done.
The next challenge is to bring these underlying facilities to large-scale production and prove that blockchain can drive fundamental innovation in AI, not just serve as a conceptual slogan or fundraising story.
RWA: Real World Assets Will Truly Land in the Real World
By Dmitriy Berenzon
RWA (Real World Assets) has been discussed for years, and now it's finally迎来规模化 adoption—mass adoption of stablecoins,完备顺畅的 on/off-ramps, and clearer regulatory frameworks have made it happen. According to RWA.xyz data, over $18 billion in various assets have been issued on-chain, compared to just $3.7 billion a year ago. This trend will continue to accelerate in 2026.
It's important to note that "Tokenization" and "Vaults" are two different RWA design patterns: the former digitizes real-world assets, while the latter channels on-chain capital into offline yield scenarios.
In the future, I look forward to seeing tokenization cover a wider range of fields: from gold and rare earth metals, to short-term credit for business operations, to public and private equity, and more global fiat currencies. We can even be "bolder"—eggs, GPUs, energy derivatives, wage advances, Brazilian government bonds, Japanese yen... all can be put on-chain.
In essence, RWA is not about "putting more things on the chain," but about upgrading the way global capital is allocated. The high barriers to entry, low transparency, and fragmentation of traditional markets can be redefined on public chains and combined with DeFi primitives to achieve composability.
Of course, many assets will still face challenges such as transfer restrictions, lack of transparency, poor liquidity, risk management, and distribution efficiency, so infrastructure that solves these problems will be equally critical.
The Agent-Driven Product Renaissance is Coming
By Ash Egan
The core of the next generation internet will no longer be the apps we scroll through, but the "intelligent agents" we converse with.
We all know that bots and agents are rapidly increasing their share of all network activity. Rough estimates suggest this proportion has now reached about 50%, including both on-chain and off-chain activities. In the cryptocurrency space, bots are increasingly参与 in trading, management, assistance, contract scanning, and performing various operations on our behalf, from token swaps and fund management to smart contract audits and game development.
This is precisely the beginning of a "programmable, agentified internet." And 2026 will be the first year crypto product design is truly "agent-centric" (in a positive, non-dystopian way).
The future form is still emerging, but for me at least, I hope to spend less time clicking between pages and more time managing on-chain agents in chat-like interfaces: like Telegram, but the conversation partner is a "specific application/specific task agent." They will be able to formulate and execute complex strategies, automatically gather information most relevant to me across the network, and present it as trading results, risks and opportunities to watch, and curated information. I give them tasks, and they track opportunities, filter out all irrelevant information, and execute at the optimal time.
The infrastructure needed for this vision is actually already prepared on-chain. Combining the default open data graph, programmable micro-payments, on-chain social graphs, and cross-chain liquidity rails, we have everything needed to support a dynamic agent ecosystem. The "plug-and-play" nature of the crypto world means the繁琐 processes and obstacles agents face will be significantly reduced. Compared to Web2 infrastructure, blockchain's readiness for this agent revolution cannot be overstated.
This might be the most crucial point here: This is not just automation, but a liberation from Web2's data silos, various frictions, and unnecessary waiting. We are seeing this transformation firsthand in search: about 20% of Google searches now directly provide AI overviews, and data shows that once users see this overview, the probability of them clicking on traditional search result links drops significantly. Manually flipping through pages of information is becoming obsolete. A programmable, agent-driven network will extend this experience further to all the applications we use daily, which I believe is a tremendous boon.
In this new era, we will: reduce mindless scrolling, reduce emotional panic trading, and time zone differences will be彻底抹平 (no more saying "wait for Asian markets to wake up"). Interaction with the on-chain world will become simpler and more expressive for both developers and ordinary users.
As more assets, systems, and users move on-chain, this cycle will continuously自我强化 and accelerate. The more opportunities on-chain, the more intelligent agents deployed, the more value unlocked. Rinse and repeat. But what we build now, and how we build it, will determine whether this "agent-driven network" ultimately becomes a thin layer of noise and automation, or truly ignites a user-empowering, vibrant, and innovative product renaissance.
