AI Agent泡沫破灭之后:Web3 AI的真正价值在哪?

Odaily星球日报Published on 2025-04-14Last updated on 2025-04-14

Abstract

下⼀波Web3 AI的浪潮正在酝酿。

Coinspire 是一个长期关注 Web3 与 Ai 板块的研究社区,提供项目研报 / 一级投研 / 二级市场分析等 Alpha 机会。

AI Agent泡沫破灭之后:Web3 AI的真正价值在哪?

引言

去年第四季度,AI 代理⼈赛道像⽕箭⼀样飙升,从⼏乎为零迅速涨到 200 亿美元以上的市值。各种“智能体”⼀时间爆⽕——搞笑的、甚⾄搞怪的,接着⼤家⼜开始憧憬那些能⾃动炒币、帮你赚钱的“⾦融代理⼈”,还有各种号称要投资其他代理⼈的 InvestmentDAO,甚⾄还有“⼈类与代理⼈共治”的 DAO 组织,玩法花样百出,仿佛⼀夜暴富不是梦。

但⻛⼝来得快,去得也快。泡沫⼀破,很多项⽬⼀个接⼀个凉了。不过,还是有⼀些基础设施/有实⽤价值的 AI 项⽬正在接管市场。真正的价值开始浮现,下⼀波 Web3 AI 的浪潮正在酝酿,这⼀次可能不只是炒作——值得⼤家认真关注。

我们都知道,⼀旦有新赛道、或有新热点(⽐如 Web2 AI、特朗普⽀持加密货币和 AI 这种组合),市场根本不在乎基本⾯。只要看起来热闹、有噱头、Demo 漂亮,不管实际⽤处⼤不⼤,市值轻松冲到上亿美元。

@virtuals_io 就是这波浪潮⾥最会讲故事的项⽬。他们精准打出市场、占据⽤户⼼智,叙事做得顶好。于是开发者想在它上⾯发项⽬,散户也跟着炒 hype。

后来 @elizaOS 出现,⾛了⼀条完全不同的路线——它开源 AI,让任何⼀个开发者都能轻松上⼿,⾃⼰动⼿挖⾦⼦。这个理念⼀下⼦引发了⼴泛共鸣,社区快速壮⼤,GitHub 上的 star 和 fork 数像坐⽕箭⼀样涨(到现在还在涨)。

Virtuals ⽣态的总估值曾⼀度冲到 50 亿美元以上,Eliza ⾼峰期也有接近⼀半的体量。其他⼀些有趣的 AI 代理⼈,⽐如 AIXBT,也⼀度冲上 10 亿市值。

当然,现在的画⻛已经完全不同了。新发的、表现不错的代理⼈项⽬,市值⼤多在 300 万到 1000 万之间;⽼项⽬的话,市值也压缩到 1000 万到 5000 万区间。整个板块的估值天花板被压缩了,总市值也从巅峰的 200 亿美元掉到了现在的 40 〜 60 亿区间。

Part.1 基础设施崛起,Web2 AI 加速来袭

现在的市场,已经不再迷信那些“看起来⽜逼”的泡沫项⽬,⽽是开始关注真正的基本⾯。特别是在 Web2 的 AI 模型发展越来越快的背景下,⼤家更关注基础设施和去中⼼化 AI 的⻓期价值

Meta 的 Llama、OpenAI 的 GPT、X 上的 Grok、DeepSeek,还有阿⾥出的 Qwen——⼏乎每个⽉都在卷更新,模型更强、更快、更智能。⽐如最近 ChatGPT 的图像⽣成功能⼀上线,⻢上就带⽕了“吉⼘⼒⻛”图⽚,直接刷屏。

Web2 的消费者产品端也在⻜速进化。因为底层 AI 能⼒变强了,很多以前做不到的产品体验,现在都能实现了。像 Lovable、Bolt、Cursor、Windsurf 这些新⼯具,让开发者的开发效率提升⼏个档次,功能更新⼜快⼜多。AI 代理⼈和智能⼯作流现在已经渗透到各个⾓落,⽽且⼊门门槛越来越低。对⽤户来说,换⼯具⼏乎没有成本——哪个不好⽤、价格贵,⽴⻢就能找到 UI 更好、体验更顺的替代品。整个市场卷得越来越猛,但也推动了真正有价值的东⻄加速落地

Part.2 数据主权觉醒:谁才是真正的数据主⼈?

在这⼀切⾼速发展的同时,越来越多的⼈开始意识到⼀个问题:现在到处都是各种 AI 代理应⽤,但它们⼤多数⽤的都是中⼼化技术——那我的数据到底归谁所有?我的聊天记录会去哪?如果我跟 AI 聊了些私密的内容,它真的会保密吗?还是会被上传、分析、拿去训练别⼈家的模型?

这个问题在 @OpenAI 最近的更新之后显得更关键了——ChatGPT 的“记忆功能”现在可以引⽤你过去所有的对话,⽣成更个性化的回应。这功能确实很酷,想象⼀下,未来每个⼈都有⾃⼰的 AI 私⼈助理、聊天伴侣、情感⽀持……但这也意味着,你的数据会被某个平台“⻓期持有”,你不再是数据的主⼈

⼀旦别⼈控制了你的对话、偏好、情绪甚⾄⽣活习惯,那后果可能就不只是“体验更好”那么简单了。

这也是为什么“数据主权”这个话题,正在成为 AI + Web3 的下⼀个重点。真正属于⽤户的数据,才是最有价值的未来

Part.3 去中⼼化 AI(DeAI)的崛起

我在去年就做过⼏个预测,其中⼀个是到 2025 年第⼆季度,去中⼼化 AI 会真正⾛⼊⼤众视野。尤其是在⼤家对隐私安全、数据所有权越来越在意的背景下,那些能提供保密性、可验证性、⽤户数据归属透明性的底层基础设施会获得更多关注和使⽤。

⽬前我们看到三个主要⽅向的趋势正在浮现:

一、Web2 AI 的 VC 趋势

▶ Y Combinator ⽀持的⼀些初创公司正在推出垂类 AI 代理⼈(专门解决某个领域问题);

▶ a16z 开始布局下⼀个阶段的消费者 AI 产品趋势,提出⾃⼰的投资逻辑;

▶ Perplexity 启动了专门投 AI 的基⾦。

二、Web3 AI 的 VC 趋势

▶ 开始下注去中⼼化 AI 基础设施;

▶ 分布式训练⽹络、算⼒共享等赛道逐渐升温。

三、Web3 AI 的散户趋势

▶ AI Agent ⽣态仍然是热点之⼀;

▶ 消费级 AI 应⽤逐渐丰富,从⽣产⼒⼯具到情感陪伴都在尝试落地;

▶ ⽤户对“我⽤的 AI 产品是不是真正为我服务,⽽不是收割我数据”越来越在意"这些趋势交织在⼀起,共同推动 DeAI 从概念⾛向实⽤阶段。2025 年将是验证去中⼼化 AI 价值的关键时刻。

Part.4 Web2 vs Web3 AI:完全不⼀样的节奏和玩法

Web2 这边,市场体量远⼤于 Web3。因为有太多传统企业正在想办法⽤ AI 来转型,优化⾃⼰的业务流程——⽐如:获取更多客户、提升转化率、增加销售额等。这类企业通常都有很明确的需求,很多还集中在某个细分领域⾥,所以他们希望找到能精准解决⾃⼰“特定痛点”的 AI ⼯具。这也吸引了很多年轻创业者,瞄准这些细分需求去做垂类 AI 代理⼈。

相⽐传统 SaaS,AI 代理⼈能带来的效益更直接——要么省⼀⼤笔钱,要么直接吸引更多客户赚钱。所以,这类 AI ⼯具的订阅价格也能卖得更贵,很多初创公司上线⼏个⽉就冲到年营收百万、千万美⾦,不是没有道理的。

但 Web3 的玩法完全不同。区块链本⾝就是为去中⼼化 AI(DeAI)量⾝定做的基础层。所有⾏为可以链上验证,不可篡改;天然提供去信任的环境;⽀持去中⼼化的计算;⽤户可以真正拥有⾃⼰的数据、模型和使⽤场景。说⼈话就是——Web3 AI 未来要让⽤户知道⾃⼰的数据被怎么⽤、理解 AI 的决策过程、⾃⼰掌控模模型和⽤例、还能从中赚钱。

「Web3 的 VC 们已经在为这种未来下注了」

Part.5 为什么散户爱 AI agent

对 Web3 散户来说,DeAI(去中⼼化 AI)确实挺难懂的:⼀堆新词新概念,听起来跟外星⽂差不多。所以最开始他们最容易被吸引的,是那些看得懂,有趣好玩的 AI 代理⼈——⽐如会说话的聊天机器⼈,能讲段⼦会整活的那种。这种“娱乐型 AI 代理⼈”确实挺吸粉,但时间⼀⻓,散户也开始发现这些东⻄好像没啥真正⽤处?再加上这段时间市场⾏情⼜差,⼀⼤堆没⽤的项⽬就慢慢被淘汰掉了,⽽那些有⽤的、能提供功能的代理⼈虽然估值也下来了,但还活着。

这波“清洗”让越来越多⼈意识到:只有具备实际⽤例和核⼼产品能⼒的 AI 项⽬才有未来。于是,项⽬⽅开始转向两个⽅向。要么⾃⼰开发真正的 AI 产品,解决实际问题;要么跟真正有技术、有价值的 DeAI 项⽬合作,⽐如 @AlloraNetwork、@opentensor(Bittensor)

这种转变有两个积极意义:让⼤家开始关注原本“看不懂”的底层基础设施;让 AI 代理⼈不再只是表演⼯具,⽽是能做实事的产品。像 @AskBillyBets 和 @thedkingdao 这样的项⽬,已经成了典型案例——不仅功能强,还顺带把像 Bittensor 这种酷炫的 DeAI 技术带到了⼤众视野中。这说明⼀个趋势:散户虽然不懂技术,但他们会慢慢被“真有⽤”的产品教育出来。

刚刚提到的项⽬Bittensor 我觉得最有意思的⼀点就是这是⼀个普通⼈也能参与投资的去中⼼化 AI ⽣态系统。现在⼤多数 DeAI 项⽬还太早期,只有 VC 或“圈内战略合作伙伴”才能投,连代币都没发的⽐⽐皆是。但 Bittensor 不⼀样。⽤户可以直接⽤ $TAO 投票⽀持你看好的⼦⽹(subnet),相当于“换仓”进这些 DeAI 项⽬的⼦代币,提前上⻋。

虽然我之前也吐槽过:跨链桥、交易体验有点⿇烦……但他们的底层技术、产品逻辑,还有整体氛围真的很强。尤其是 @rayon_labs 的存在,让整个⽣态的 UX/UI 设计更朝着“⽤户友好”的⽅向发展。因为在 Bittensor 的机制⾥,每个⼦⽹要想获得更多奖励(挖矿激励),就得靠市场认可——谁有⽤、谁⽜,谁才能拿更多分发

所以,对这些⼦⽹来说,“让⽤户理解你在⼲嘛”变得超级重要。Rayon Labs 正在做这件事。他们的产品⽅向⾮常明确:为普通⽤户优化 UI/UX。他们不仅有多个实⽤的⼦⽹ (⽐如 Gradients:⼀个超⽅便的 AutoML 平台,⽤户可以直接在上⾯训练模型,点点按钮就能跑起来。以及他们最新推出的旗舰产品也⾮常酷:Squad AI Agent 平台:你可以像拖图块⼀样拖拽模块来创建 AI Agent,真正做到“零代码搭建 AI 代理⼈”。这体验有点像 Web3 版的“傻⽠式 AI ⼯⼚”,很适合不懂技术的⽤户⼊门)。

「总的来说」,Bittensor ⽣态现在不仅是技术最先进之⼀的 DeAI 项⽬,⽽且在普通⽤户参与友好度上也⾛在最前⾯。这种做产品逻辑清晰、⽤户友好的团队,正是让这个⽣态变得有热度的关键⾓⾊。

我们正处在⼀个由 Web3 AI 主导的⼤变⾰时代。过去那种靠炒作拉⾼市值的泡沫已被实际的基础设施、去中⼼化 AI 和真实的应⽤场景所取代。⽆论是企业在 Web2⾥利⽤ AI 优化业务,还是散户在 Web3 ⾥体验新型代理⼈的便捷,未来的数据主权和⽤户参与都将成为关键。Web3 AI 还远远没到⾼潮。真正的戏,才刚刚开始。

*Coinspire 平台上的所有内容仅供参考,不构成任何投资策略的要约或建议,基于本文内容所做的任何个人决策均由投资者自行承担责任,Coinspire 不对由此产生的任何收益或损失负责。投资有风险,决策需谨慎

Trending Cryptos

Related Reads

How Far Are We from the End of the Crypto Bear Market?

How Much Longer Until the Crypto Bear Market Ends? A persistent negative Coinbase Bitcoin Premium Index, reaching a record 46 consecutive days of negative values, underscores the ongoing crypto bear market. The downturn accelerated in late May/early June when MicroStrategy (referred to as "Strategy" in the text), a major BTC holder, sold a small amount of Bitcoin (32 BTC), shattering market confidence and triggering sharp declines in BTC, ETH, and SOL. This was compounded by sustained net outflows from US spot Bitcoin ETFs. June saw Bitcoin briefly fall below $60,000, its worst weekly performance since 2022, breaking below the critical 200-week moving average—a signal some analysts consider confirmation of a bear market. While some institutional analyses in mid-June suggested the bottom might be near (e.g., around $53,600 based on realized price, or a potential end during the 2026 World Cup summer), further price drops quickly disproved these optimistic forecasts. Key bear market indicators include: the severe de-pegging of MicroStrategy's preferred shares (STRC) from their $100 NAV, which later recovered partly due to company stabilization plans; a record number of BTC (over 10.83 million) and a significant portion of long-term ETH holders now in unrealized loss; and Bitcoin's price trading below its 200-week moving average. Predictions for the bear market's end vary. Some, like investor Yilihua, suggest July-August 2026 could present a final buying opportunity. Others, like miner Jiang Zhu'o'er, predict a bottom between $42,000-$44,000 in October-December 2026, based on the historical lag between MicroStrategy's mNAV ratio bottom and Bitcoin's price bottom. Technical indicators like the BTC 4-year average price index dipped briefly below 1, and the Coinbase Premium Index remains negative, suggesting a need for a price rebound to around $77,000 for a return to positive territory. In the absence of major external catalysts, the consensus leans toward the bear market persisting for another 2-3 months, with late September to early October 2026 being a potential turning point window.

Odaily星球日报1h ago

How Far Are We from the End of the Crypto Bear Market?

Odaily星球日报1h ago

Solana Ecosystem Shows Signs of Recovery: On-Chain Governance Upgrade, Tokenized Stocks and Meme Coins Heat Up

Solana Ecosystem Shows Signs of Revival: Governance Upgrades, Tokenized Stocks, and Memecoins Heat Up While the broader crypto market faces a downturn, Solana has shown relative strength with its price rising nearly 15% recently. This resilience is attributed to warming ecosystem activity, particularly in tokenized stocks and memecoins, coupled with a major upgrade to its on-chain governance. In the Real World Assets (RWA) sector, Solana now leads all public chains in both the number of unique holder wallets and the quantity of RWA assets. Its tokenized stock weekly trading volume has surged to a record $1.42 billion, capturing about 96% of the market share, largely driven by the Backpack exchange. Simultaneously, the Solana memecoin sector has reignited, fueled by the rapid ascent of the ANSEM token following endorsements from a prominent crypto influencer. This has boosted activity across platforms like Pump.fun and increased fee revenue for several Solana-based exchanges. A key development is the launch of Solana Governance Proposals (SGP), a new on-chain mechanism. It allows validators with at least 100,000 SOL delegated to submit proposals for community vote, enhancing decentralized decision-making. SGP will operate alongside the existing technical proposal process (SIMD), focusing on broader ecosystem governance. Despite these positive signals in specific areas, overall on-chain activity and transaction volumes still lag behind previous bull market peaks, indicating the recovery remains partial rather than ecosystem-wide.

marsbit1h ago

Solana Ecosystem Shows Signs of Recovery: On-Chain Governance Upgrade, Tokenized Stocks and Meme Coins Heat Up

marsbit1h ago

While Semiconductor Stocks Plunge, Anthropic Plans to Develop a 2nm Chip

Anthropic, the AI company behind Claude, is exploring the development of its own custom AI chip, according to a report from The Information. The company is in early discussions with Samsung Electronics to manufacture the chip using Samsung's most advanced 2-nanometer process and packaging technology. While the project is still in preliminary stages, including defining chip specifications, and could be abandoned, it marks a strategic step for Anthropic. The move comes as the company seeks greater control over its computing costs and hardware optimization, particularly for inference tasks to run its models more efficiently and cheaply. Samsung's potential involvement follows its participation as a strategic investor in Anthropic's recent $65 billion funding round. For Samsung, partnering with a major AI lab represents a significant opportunity for its foundry business to compete with market leader TSMC in advanced semiconductor manufacturing. Anthropic's CEO, Dario Amodei, has previously highlighted the immense financial challenge of securing enough computing power for anticipated growth, making cost-effective inference a critical focus. The company would join other tech giants like Google, Amazon, Microsoft, Meta, and OpenAI in pursuing custom AI silicon. However, analysts note this trend creates deeper interdependencies rather than independence, as US AI labs become more tightly woven into Asian semiconductor supply chains. Despite this move, Anthropic remains heavily reliant on a multi-cloud, multi-vendor strategy for its immediate computing needs. It has secured massive, long-term commitments for capacity from Amazon Web Services (Trainium chips), Google (TPUs), and even leased a large GPU cluster from xAI. For now, Nvidia continues to dominate the AI chip market, with its share reportedly growing to 74%.

链捕手1h ago

While Semiconductor Stocks Plunge, Anthropic Plans to Develop a 2nm Chip

链捕手1h ago

Trading

Spot

Hot Articles

Discussions

Welcome to the HTX Community. Here, you can stay informed about the latest platform developments and gain access to professional market insights. Users' opinions on the price of AI (AI) are presented below.

活动图片