怎样在 CEX 挑选潜力代币?

链捕手Опубліковано о 2025-01-06Востаннє оновлено о 2025-01-06

作者:Haotian

 

相信很多人打开 DEX 都有随地捡宝的感觉,毕竟到处都是 10-300M 技术品相还不差的热门项目,然而打开 CEX 却又到处是 300M-3B 的老叙事项目,内心各种五味杂陈 Plus 一言难尽。但事实上,CEX 也有一套顺着趋势捡宝的逻辑,听我简单唠唠:

1)这一轮 AI Agent 叙事潮正在快速演进,从 AI MEME——>单体 AI 应用——>AI Agent Launchpad——>AI Agent 框架标准——>AI Agent 链化。 

纯 MEME 不多说一直都是速通造势逻辑;单体 AI 由于要拼体验,短期 AIXBT 可以独霸一方; Launchpad 平台一定会被 Virtual 吃尽先发优势;框架和标准由于背靠复杂的技术逻辑,想象空间大,但当前还处于大乱斗阶段;唯独 AI Agent 链化方向相对「有迹可循」。

2)「链化」」的逻辑一定不是某个公链跳出来说,我可以支持一切 AI Agent 跑在链上。这样的公链一定有,比如, 

@NEARProtocol、ICP、BNBChain,都有这样的潜力。但问题是,现阶段 AI Agent 还处于 MEME 化「发行资产」的初级阶段,只是能跑 AI Agent,却不能调动起开发者的积极性,无济于事。

因此第一波的「链化」叙事一定来自 ai16z 生态系列。比如  @focEliza。

3)而且「链化」的核心点在于为大语言模型 LLM 更好的兼容和服务 AI Agent 构建基础设施,比如:

1、AI Agent 要实现自主化生成私钥和管理资产,就势必要依赖 TEE 大基建,那么,能提供 TEE 成熟解决方案的项目就是 CEX 中老叙事中的潜在 Alpha。因为 TEE 在传统 ZK、MPC、FHE 等一众密码学算法竞争压力下一直不起眼,而突然间就成为了新一波 AI 大叙事的核心基建。

试想一些小市值、低存在感、技术扎实、又踩上新风口的项目,一定会是众多 CEX 标的中潜在的宝,( $PHA 的表演秀逻辑正是如此,如果你认同 TEE 大基建对 AI Agent 的价值意义,TEE 赛道的后劲就一定还在)。

2、AI Agent 要实现 Memory 的分层优化和匹配,目标就得构建适用于 AI Agent 的 Data Avaliability(DA)能力,传统 EVM 公链在 DA 板块延伸出了 Blob 外挂空间、第三方 DA War 等诸多精彩的叙事。

专属于 AI Agent 的 DA 能力如何打造也一定会是话题焦点:包括 DA 区块空间如何精准记录 LLM 上下文有效语义,Character 角色设定插件如何实时和 DA 区块进行交互,多区块空间中的数据如何有效匹配多模态交互,多区块空间分层存储数据成本问题(短期、长期、工作记忆等)等等,都是新型 DA Builder 亟待攻克的难题。这也是 focEliza 以 DA 能力作为核心破局点的原因。

3、AI Agent 要实现单模态的可验证性和多模态的可信交互,都得依赖公链级的可信验证处理。在 TEE 基建在解决大模型生成和应用私钥上的环节,唯有 TEE 显然不能解决单点 TEE 硬件被物理爆破问题,以及 TEE 执行程序的共识可验证问题。

这些往长远说都得依赖区块链天然去中心化的节点验证共识和智能合约协作调用环境来实现。

也就是说,一个共识强大的区块链分布式系统会让 AI Agent 的智能变得「靠谱」有保障。所以顺着这个逻辑,目前正在致力于为 AI Agent 提供 zkVM 底层框架、ZK Oracle 预言机方案、ZK Bridge 跨链解决方案、「链抽象」公链级应用解决方案等项目,也都在这一轮「链化」的叙事趋势当中。

以上

当 TEE+DA+Oracle+zkVM+ 链抽象等等一系列为 AI Agent 提供「链化」能力的解决方案成熟后,这个时候,AI Agent 需要的去中心化分布式算力、去中心化微调推理环境、去中心化数据来源、去中心化 IP 通信和激励等等一系列 AI Platform 等平台做的事情才会真正成为「刚需」。

某种程度上,基于这一步 AI Agent 链化方向的努力,io、Aethir、Vana、SaharaAI 等大 AI 平台项目也才会有可用武之地。

厘清楚这个逻辑,DEX 之上如何在乱纪元之中取到真经,CEX 之中又如何在沉寂之中捡到宝,不多说,自然会有清晰的判断。

Пов'язані матеріали

Strategy's Bitcoin Sales Cap Far Exceeds $1.25 Billion: A Detail the Market Overlooked

The article discusses how MicroStrategy's potential Bitcoin sales go far beyond the announced $1.25 billion "reserve-building capacity." It clarifies a key distinction in the company's "BTC Monetization Program": selling Bitcoin to *build* a new dollar reserve (the $1.25B cap) versus selling to *replenish* the existing USD Reserve after it's used for expenses like preferred share dividends. The recent $216M BTC sale for dividend payments was a "replenishment," leaving the headline $1.25B building quota untouched. The plan actually outlines three potential funding pools from BTC sales: 1) Building the reserve ($1.25B cap), 2) Covering preferred share/ debt costs (no specified cap), and 3) Funding buyback programs (up to $20B). This means the structured sales potential exceeds $30 billion, not including uncapped replenishment sales. The piece argues this marks MicroStrategy's shift from a passive "buy-and-hold" Bitcoin proxy to an actively managed entity using BTC as a balance-sheet tool to manage its complex capital structure (common stock, preferred shares, debt, reserve). This creates new dynamics and potential conflicts, as actions benefiting one part (e.g., selling BTC to pay dividends) may pressure another (e.g., undermining the "never sell" narrative). Investors must now parse the company's specific terminology ("build" vs. "replenish") to understand the true scope of future BTC sales, which is significantly larger than the market initially perceived.

marsbit4 хв тому

Strategy's Bitcoin Sales Cap Far Exceeds $1.25 Billion: A Detail the Market Overlooked

marsbit4 хв тому

Goldman Sachs Report Deconstructs the Competitive Landscape of China's AI Large Models: Who Will Be the Long-Term Winner?

Goldman Sachs analyzes China's AI large language model (LLM) landscape, identifying key players and a strategic shift towards efficiency and global expansion. The report highlights that Chinese open-source/open-weight models are closing the performance gap with top global proprietary models at significantly lower cost, driven by architectural innovations like MoE. This enables a "two-tier" market: a high-end segment (e.g., GLM5.2, Qwen3.7 Max) with pricing at ~$1 per million tokens, and a low-end, price-sensitive global segment. Open-source strategies aid adoption but limit monetization, as deployments via third-party platforms (e.g., AWS Bedrock, Alibaba Cloud) may not generate direct revenue for model creators. The industry is thus moving towards "open-weight + community license" models with revenue-sharing to improve unit economics. Internationally, the focus is shifting from "token maximization" to ROI-driven enterprise adoption, particularly in non-U.S. markets. Major cloud platforms are integrating Chinese models (e.g., DeepSeek, MiniMax). Using a competitive framework based on pricing power, cost advantage, and financial strength, Goldman Sachs identifies **Zhipu AI** and **DeepSeek** as leaders in foundational text models, and **ByteDance** (with Seedance) leading in multimodal/video generation. **MiniMax** and **Kuaishou** are also rated favorably. The firm forecasts China's AI model API/subscription revenue growing from ~RMB 35bn (2026E) to RMB 879bn by 2030.

marsbit4 хв тому

Goldman Sachs Report Deconstructs the Competitive Landscape of China's AI Large Models: Who Will Be the Long-Term Winner?

marsbit4 хв тому

$8 Trillion: The Second-Largest IPO in History Has Arrived

SK Hynix Makes History with World's Second-Largest IPO. The global memory chip leader SK Hynix debuted on Nasdaq, raising $26.5 billion and achieving a market cap exceeding $1.2 trillion. This marks the largest U.S. IPO by a foreign company and the second-biggest globally. The company's journey is a remarkable turnaround. Founded in 1983, its predecessor, Hyundai Electronics, faced near-bankruptcy during industry downturns before being acquired by SK Group in 2011. A pivotal early bet on HBM (High Bandwidth Memory) technology, initially with AMD in 2013, ultimately paid off with the AI boom. SK Hynix now supplies HBM3 to NVIDIA and commands 58% of the global HBM market. Driven by soaring AI demand, SK Hynix reported staggering Q1 2026 profits with a 72% operating margin. Its surging stock made it South Korea's second trillion-dollar company. Profits are shared widely with employees through a new bonus system tied to 10% of annual operating profit. The article highlights an ongoing "super memory cycle" fueled by AI, with market forecasts predicting massive growth. This presents a historic opportunity for Chinese memory chip makers. ChangXin Memory Technology (CXMT) is set for a domestic IPO, potentially reaching a ~$420 billion valuation as China's top DRAM producer. Yangtze Memory is also preparing to go public. While these "domestic storage leaders" are gaining ground, the article notes they still face technology and margin gaps compared to established giants like Samsung and SK Hynix.

marsbit3 год тому

$8 Trillion: The Second-Largest IPO in History Has Arrived

marsbit3 год тому

Торгівля

Спот
活动图片