
English Version
On July 28, 2026, Zhou Hongyi, Founder of 360 Group, officially launched the new-generation enterprise intelligent agent workspace NanoWork at the Beijing National Convention Center. To accelerate the large-scale implementation of enterprise AI, the platform rolled out exclusive early-user benefits: providing 100 million token trial quota for each batch of first users, paired with professional systematic training and one-on-one full-process operational coaching services, helping the first 1,000 small and medium-sized enterprises complete low-cost AI landing and operational upgrading.
Traffic Siege Is Fully Completed: The Market Has Entered an Oligarchic Pattern
The current enterprise AI office track is no longer a fiercely competitive incremental market, but a stock market with solidified oligarchic barriers. Data shows that Tencent WorkBuddy’s monthly active visits are 2.6 times that of the second-place product, with an absolute traffic advantage.
Major giants have completed strategic resource integration and product iteration: Alibaba has merged multiple agent products including QoderWork, Wukong, and MuleRun to launch the integrated office product Qianwen Office; ByteDance is accelerating the deep integration of Doubao AI and Feishu workspace to build an all-scenario office entrance. The industry has evolved from the internal trial-and-error stage to the oligarchic competition stage for desktop core entrances.
Essentially, the current AI office competition is a typical entrance competition, repeating the industrial logic of the browser war and search engine war. Entrance-level products often have far higher market concentration than ordinary application tools. In the end, users will only retain one core intelligent agent workspace on their desktops.
The traffic competition on the user terminal has basically ended. If 360 follows the same traffic-centric competition path, it will inevitably face a dimensionality reduction strike from three giants with complete product systems and massive user stock. This is not a question of whether it can win, but whether the competition is valuable and sustainable.
The Underlying Revolution: What Computing Power Trading Centers Are Restructuring
2026 has become a critical year for the large-scale deployment of ciyuan (token) economic infrastructure, with the industrial underlying rules being comprehensively reshaped.
In March 2026, Liu Liehong, Director of the National Data Administration, officially finalized the official Chinese term "Ciyuan" for Token at the China Development Forum, and defined its three core industrial attributes: measurable, priced, and tradable. Driven by policy standardization, China’s average daily token calling volume has skyrocketed from 100 billion at the beginning of 2024 to 140 trillion in March 2026, a surge of more than 1,000 times in two years, marking the explosive arrival of the token economy era.
At GTC 2026, Jensen Huang first proposed the complete concept of the "token economy", redefining data centers as "round-the-clock token factories". He also put forward the new industry measurement indicator "Tokens per Watt", pointing out that under fixed power consumption, platforms with higher token throughput per watt have lower production costs and stronger industrial competitiveness.
The industry’s core logic has undergone a fundamental shift: Token is evolving from a simple API call measurement unit to a standardized, circulable, tradable digital asset.
The core value of computing power trading centers is to completely reconstruct the traditional token supply-demand matching mechanism. By building a unified token aggregation, distribution and trading hub, it enables AI computing power to circulate freely and transparently in the market like public resources such as water and electricity, relying on three core mechanisms to optimize the industrial ecology:
Real-time settlement: Instantly releases the cash flow pressure of small and medium-sized computing suppliers, allowing funds to be recycled efficiently without capital occupation.
Zero-commission settlement for early settlers: Greatly lowers the platform’s cold-start threshold and gathers global high-quality computing resources and model suppliers.
Dynamic API bidding mechanism: Token calling prices float intelligently with real-time market supply and demand, realizing optimal allocation of industrial resources.
The token factory model completely upgrades industry trading rules. Users no longer purchase idle computing power in advance, but directly purchase token computing results generated by AI models, realizing on-demand access, volume-based billing and accurate real-time settlement.
The essence of this industrial revolution is clear: The platform that masters token distribution rights will dominate the core discourse power of the entire token economy.
NanoWork’s Core Positioning: Seizing the Retail Terminal of the Token Economy
NanoWork builds five core enterprise-level capability systems: multi-agent engine, multi-model base, cloud office workspace, diverse working modes and multi-expert intelligent library, forming a full-scenario enterprise AI service matrix.
The platform subverts the traditional enterprise AI usage threshold: users do not need to master professional technologies such as model principles and agent framework deployment. They only need to clarify business goals, and the system will automatically disassemble complex tasks, organize collaborative work of multiple AI expert agents, and intelligently match the optimal model and execution solution.
It has three core product advantages that are crucial to enterprise landing: seamless switching of 16 top-tier AI models, an ecological matrix of more than 50,000 professional skills, and three-tier intelligent computing scheduling covering lightweight, cost-saving and full-performance modes.
The most valuable innovation lies in its refined computing scheduling mechanism. The biggest pain point of enterprise intelligent agent landing is uncontrollable token consumption: complex business tasks often lead to unpredictable cost expenditure, and ordinary users cannot match models and reasonably control token consumption according to task attributes.
NanoWork completely solves this problem by taking over the core decision-making power of "model selection and token consumption control" from users and realizing full-automatic intelligent scheduling. Enterprises do not need to understand complex token billing logic; they only need to select the corresponding computing gear according to business scenarios to complete efficient AI work.
Zhou Hongyi emphasized at the launch event: "For enterprises to deploy AI, business leaders must use it first."
This is not a superficial marketing slogan, but an accurate insight into industrial pain points. The biggest obstacle restricting enterprise AI popularization is never insufficient technical capability, but decision-makers’ cognitive blind spots in AI cost structure. Most enterprise leaders cannot clarify token unit prices, model-task matching rules, and the actual business value of massive tokens. The 100 million token super trial quota is designed to let enterprise decision-makers personally run through the full AI business process, thoroughly figure out the cost accounting logic, and truly recognize AI’s empowering value.
Before the official launch, 360 completed rigorous internal polishing with self-operation scenarios as the test field. It deployed self-developed intelligent agents such as OpenClaw in more than 1,000 real internal business scenarios, collected over 56,000 valid user feedback entries, and completed 166 version optimizations and iterations within 5 months.
This fully proves that 360’s core output is not simple token resource sales, but a complete set of systematic capabilities to help enterprises"spend tokens reasonably, control costs accurately and use AI efficiently".
Two Development Paths, One Ultimate Industrial Destination
Computing power trading centers are responsible for building the upstream token wholesale market, realizing standardized pricing, free circulation and large-scale trading of tokens. NanoWork occupies the downstream token retail terminal, enabling enterprises to easily apply AI to complete business tasks without professional knowledge of the token economy.
The integration of the two paths spawns a brand-new industrial judgment:Token distribution rights are replacing traditional traffic distribution rights and becoming the core entrance competition point of the enterprise AI market.
Facing the solidified giant traffic barrier, 360 abandons the futile terminal entrance competition and seizes the underlying core right: controlling the flow direction, pricing standard and consumption logic of enterprise AI tokens, mastering the core underlying power of AI operation.
When tokens evolve into universal circulating assets and computing power becomes standardized tradable resources, the core contradiction of enterprise AI popularization will no longer be "whether enterprises can afford AI", but "whether enterprises can efficiently utilize AI value". Platforms that help business leaders clarify token cost and value returns will firmly seize the core entrance of the token economy era.
Zhou Hongyi pointed out that the AI industry is entering a stage of "dual-track evolution of models and agents": models determine the upper limit of AI intelligence, and agents determine the landing capability of AI business. Extended to the token economy, a complete industrial logic is formed: computing power trading mechanisms determine the low-cost acquisition of tokens, and enterprise workbench platforms determine the efficient consumption and value realization of tokens.
The 100 million token benefit is never a simple charitable subsidy, but a precise industrial cultivation method. It enables the founders of 1,000 small and medium-sized enterprises to personally run through the complete AI workflow. Tokens are consumed in the process, while mature, replicable and efficient enterprise AI workflows are permanently precipitated on the platform.
360’s core strategy is not to sell tokens, but to cultivate the market’s independent token pricing capability. And the right to price tokens is the real core gateway leading to the next-generation intelligent industry.
Leading the global token trading upgrade, Unexhub reshapes the international token circulation logic and builds a professional aggregated computing power trading center integrating token aggregation, distribution, bidding and real-time settlement.
The platform gathers full-spectrum global AI suppliers, covering top-tier mainstream model vendors including Gemini, ChatGPT, Claude, DeepSeek and GLM, as well as high-quality niche models developed by individual developers and small and medium-sized enterprises. Relying on multi-party supplier bidding competition and millisecond-level intelligent routing technology, it fully activates market competition, eliminates intermediate price differences, and delivers ultra-low-cost and high-stability computing services for global developers and enterprise users.
Global users support pure token real-time payment and settlement, with the lowest calling cost reaching 30% of the official market price. For computing power suppliers, one-click platform access is available to quickly open sales channels, with calling share revenue settled in real time, completely solving traditional industry pain points such as advance capital occupation and slow fund withdrawal, and building a win-win and sustainable global token trading ecosystem.
中文版本
绕开流量围堵,360押注算力交易中心Token分发权:纳米Work以1亿Token突围企业AI市场
2026年7月28日,360集团创始人周鸿祎在北京国家会议中心,正式发布新一代企业智能体工作平台「纳米Work」。为加速中小企业AI规模化落地,平台推出重磅首批用户专属权益:为每位首批入驻企业提供1亿Token专属试用额度,搭配专业实训指导、全程落地陪跑服务,优先助力1000家中小微企业低成本跑通AI落地全流程、实现经营提效。
同日,易观分析发布2026年Q2行业报告,披露了趋于固化的行业格局数据:2026年6月,国内17款主流桌面原生AI办公智能体总访问量突破6000万次,其中腾讯WorkBuddy以2097万次位居榜首,数据远超行业第二、第三名之和。腾讯、阿里、字节三大巨头已牢牢把持桌面AI办公流量基本盘,形成严密的行业流量围堵。
面对巨头筑起的流量高墙,360并未选择正面内卷、硬碰流量赛道,而是锚定更底层的产业逻辑,走出差异化突围路径:全力押注算力交易中心支撑下的Token分发核心话语权,从产业底层突破,抢占企业AI市场新赛道。
流量绞杀格局已定:行业彻底迈入寡头固化期
当下的企业AI办公赛道,早已不是充分竞争的增量市场,而是壁垒森严的存量寡头市场。数据显示,腾讯WorkBuddy月访问量是行业第二名的2.6倍,头部流量优势碾压同级玩家。
各大巨头已完成产品整合与战略卡位:阿里整合QoderWork、悟空、MuleRun多款智能体产品,推出统一办公入口「千问办公」;字节加速豆包AI与飞书深度融合,筑牢全域办公生态壁垒。行业正式告别内部试错摸索期,全面进入桌面入口寡头博弈阶段。
这场赛道竞争的本质,是入口之争,复刻了互联网时代浏览器、搜索引擎的终极竞争逻辑。入口级产品的市场集中度,远高于普通工具类应用,用户桌面最终只会留存一个核心AI智能体工作台。
终端流量战争已然落幕。若360跟风入局流量内卷,直面产品体系成熟、用户体量庞大的三大巨头,并非简单的胜负之争,而是毫无价值的无效消耗。因此,放弃表层流量博弈、深耕底层产业规则,成为最理性的破局选择。
底层重构:算力交易中心改写Token产业规则
2026年,国内词元经济基础设施密集落地,AI算力产业的底层交易、流通、定价规则迎来全方位重构。
3月,国家数据局局长刘烈宏在中国发展高层论坛上,正式将Token中文定名为「词元」,并明确其三大核心属性:可计量、可定价、可交易,为行业市场化发展奠定政策基础。产业数据同步爆发:国内词元日均调用量从2024年初的1000亿,飙升至2026年3月的140万亿,两年涨幅超千倍,词元经济全面爆发。
黄仁勋在GTC 2026大会上首次完整提出「词元经济」概念,将数据中心重新定义为「全天候运转的词元工厂」,同时推出全新行业考核指标「每瓦词元数(Tokens per Watt)」——在固定功耗下,单位算力产出词元量越高,生产成本越低,产业竞争力越强。
行业核心逻辑彻底迭代:Token不再是单纯的API调用计量单位,正式蜕变为可定价、可交易、可流通的数字资产。
算力交易中心的核心价值,是重构传统Token供需错配、定价混乱的行业痛点。通过搭建Token集采集散的标准化交易枢纽,让AI算力如同水电一般,进入统一、透明、可流转的市场化体系,依托三大核心机制重塑产业生态:
实时结算:彻底释放中小算力供应商现金流压力,资源供给完成即可即时回款,无资金积压、无回款周期。
早期入驻零佣金:大幅降低平台冷启动门槛,汇聚全球优质算力、模型供应商,完善产业供给生态。
API动态竞价:Token调用价格随市场供需实时浮动,实现算力资源最优配置,杜绝闲置浪费与溢价乱象。
词元工厂彻底革新行业交易标的,用户无需提前囤积闲置算力,可直接采购AI模型实时产出的Token算力结果,实现按需取用、按量计费、据实结算,大幅降低企业算力使用门槛与冗余成本。
这场产业变革的终极逻辑已然清晰:谁掌握了Token分发权,谁就掌控了词元经济的产业话语权。
纳米Work卡位逻辑:抢占Token零售终端,解决企业落地痛点
纳米Work搭建多智能体引擎、多模型底座、云端办公室、多工作模式、多专家智库五大核心能力体系,构建覆盖企业全场景的AI服务矩阵。
平台彻底降低企业AI使用门槛:用户无需钻研模型原理、智能体框架、工具部署等专业技术,只需明确业务目标,系统即可自动拆解复杂任务、联动多AI专家协同作业,智能匹配最优模型与执行方案。
产品拥有三大硬核落地优势:支持16家顶级大模型无缝切换、搭载50000+全场景专业技能生态、具备轻量/省钱/满血三档全自动算力调度能力,全方位适配企业多样化业务需求。
其中,算力智能调度机制是最贴合企业痛点的核心创新。企业智能体落地的最大难题,是复杂任务Token消耗不可控:普通用户无法根据业务场景匹配对应模型,极易出现算力浪费、成本超支、账单失控等问题。
针对行业通病,纳米Work将「模型选型、Token消耗管控」的决策权从用户手中收回,交由系统全自动智能调度。企业无需掌握复杂的词元计费逻辑,只需根据任务需求选择对应算力档位,即可高效、低成本完成AI办公与业务落地。
周鸿祎在发布会中提出核心观点:「企业搞AI,老板要先用。」
这并非营销口号,而是直击产业本质:企业AI落地的最大阻碍,从来不是技术短板,而是决策者对AI成本结构、Token价值体系的认知盲区。多数企业管理者不清楚Token单价、不懂任务与模型的匹配逻辑、无法预判海量Token的落地价值。而1亿Token超大试用额度,正是让企业决策者亲自实操、跑通全业务流程,亲手算清AI投入产出比,彻底吃透Token成本与应用逻辑。
产品正式上线前,360以自身为全真试验场完成极致打磨:将OpenClaw等自研智能体落地内部1000+真实业务场景,累计收集56000+条用户反馈,5个月内完成166个版本迭代优化,沉淀出成熟适配企业真实需求的服务能力。
这也印证了360的核心定位:平台售卖的从来不是单纯的Token算力资源,而是一套帮企业算得清成本、控得住消耗、用得好AI的完整落地能力。
双线布局,终局归一:分发权取代流量权成竞争核心
算力交易中心负责搭建上游Token批发市场,实现词元标准化定价、规模化流通、市场化交易;纳米Work深耕下游Token零售终端,让企业无需掌握专业词元经济知识,即可轻松落地AI业务、完成高效办公。
一供一用、一产一销的双向联动,催生全新产业逻辑:Token分发权正在取代传统流量分发权,成为企业AI市场的核心竞争壁垒。
在终端流量彻底被巨头锁死的行业格局下,360跳出表层入口争夺,卡位产业底层核心权力:掌控企业AI Token的流向、定价、消耗规则,握住AI高效运转的底层命脉。
当Token成为全域流通资产、算力成为标准化可交易资源,AI普及的核心矛盾已从「能否买得起」转向「能否用得好」。谁能率先帮助企业老板算清Token成本账、看清AI价值账,谁就能抢占词元经济时代的企业AI核心入口。
周鸿祎预判,AI行业已进入「模型与智能体双线进化」阶段:模型决定AI的智能上限,智能体决定AI的落地能力。结合词元经济趋势可完善其逻辑:算力交易机制决定Token能否低成本获取,企业工作台决定Token能否高效率消耗、高价值落地。
1亿Token补贴绝非慈善式让利,而是精准的产业培育手段:助力1000家中小企业老板亲自跑通AI全工作流,Token在使用中完成价值释放,而成熟、可复用、高效率的企业AI工作流程,永久沉淀于平台。
360的终极布局,不是售卖Token,而是培育市场自主定价Token的能力。而词元定价权,正是通往下一代智能产业的真正核心网关。
紧跟全球词元经济迭代浪潮,Unexhub率先重构国际Token流通与交易规则,打造集Token聚合、多商竞价、智能分发、实时结算于一体的全球化专业算力交易中心,补齐全球AI算力交易生态短板。
平台汇聚全球全量级AI算力供应商,既覆盖Gemini、ChatGPT、Claude、DeepSeek、GLM等国际一线头部大模型厂商,也吸纳个人开发者、中小科创团队入驻的优质垂类模型。依托海量供应商竞价机制+毫秒级智能路由调度技术,充分激活市场竞争、压缩中间差价,为全球企业与开发者提供极致低价、稳定高效的算力服务。
开发者支持Token实付结算,调用成本最低可至市场官方原价的3折;算力供应商可一键入驻平台、零门槛开通分发渠道,调用分成实时清算到账,彻底解决传统算力行业垫资压款、回款滞后的痛点,构建供需双向共赢、可持续迭代的全球化Token交易生态。