从算力交易中心批发1亿Token送给用户:纳米Work的小微企业AI抢占战正式打响

Paired with professional training and one-on-one full-process operation services, 360 takes the lead in helping 1,000 small and micro enterprises complete AI implementation and operational efficiency upgrading.


From Wholesale to Retail: Behind NanoWork’s 100 Million Tokens, Computing Power Trading Logic Is Being Redefined

English Version

On July 28, 2026, Zhou Hongyi, founder of 360 Group, officially launched NanoWork, a new-generation enterprise AI agent workspace. During the launch event, he announced a landmark support policy: providing each batch of early users with a 100 million-token trial quota. Paired with professional training and one-on-one full-process operation services, 360 takes the lead in helping 1,000 small and micro enterprises complete AI implementation and operational efficiency upgrading.

To visualize the value of 100 million tokens: converted at current mainstream AI computing pricing standards, this quota is equivalent to tens of thousands of US dollars in high-quality computing resources. In total, Zhou Hongyi released billions of tokens in one go.

However, the significance of this move goes far beyond a simple industrial subsidy. It marks a fundamental reshaping of the underlying logic of domestic AI computing power transactions and token value circulation.

NanoWork Computing Scheduling: Returning Cost Decision Power to the Platform

NanoWork is positioned as an all-scenario enterprise intelligent agent platform with comprehensive capabilities. It builds five core capability systems: multi-agent engine, multi-model base, cloud office, multi-work mode and multi-expert intelligent library. It supports seamless switching of 16 top-tier global AI models and integrates an ecological matrix of more than 50,000 professional AI skills.

Enterprises no longer need to invest heavily in researching model selection, agent frameworks and tool adaptation. Users only need to put forward business goals, and the platform will automatically disassemble tasks, organize collaborative work of different AI expert agents, and match the optimal model and execution solution intelligently.

Its most core advantage lies in the three-speed automatic computing scheduling mechanism, which divides computing resources into lightweight, cost-effective and full-performance modes to adapt to tasks of different complexities.

This design solves a universal industry pain point that most AI products avoid: tokens have real costs, while most enterprise users lack the ability to calculate and control token consumption accurately. Complex business tasks often lead to unpredictable token waste and out-of-control computing bills.

NanoWork fundamentally solves this problem by taking over the decision-making rights of "which model to use and how many tokens to consume" from users and realizing intelligent, refined and cost-controlled full-process scheduling via the platform system.

Zhou Hongyi shared a thought-provoking viewpoint at the launch event: "For enterprises to adopt AI, business leaders must use it first."

This is not a superficial marketing slogan, but a precise insight into the bottleneck of enterprise AI implementation. The biggest obstacle to enterprise AI landing is never insufficient technical capability, but decision-makers’ ignorance of the AI cost structure. Most enterprise leaders cannot clarify the unit price of tokens, the matching logic between tasks and models, and the actual business value of massive token resources.

The 100 million-token trial quota is precisely designed to solve this problem: letting enterprise decision-makers experience the whole process personally, clearly calculate the input-output ratio of AI, and truly understand what scenarios tokens are suitable for and how much value AI can create.

Tokens Evolving From Consumables to Circulating Currency: The Rise of National Computing Trading Centers

Combining NanoWork’s strategic layout with a series of key industrial events in July 2026, we can clearly perceive the upcoming comprehensive reshaping of the computing power industry.

In July 2026, computing power trading centers officially became the new core infrastructure of China’s AI industry.

On July 1, Nanjing Token Collection Store officially launched in Jiangning Qilin Science and Innovation Park, the city’s first one-stop service carrier for computing tokens. Taking tokens as the unified measurement standard for computing power, it fully connects the whole industrial chain of computing supply, model development, industrial application and financial supporting services.

On July 19, Tonglu County Government signed a strategic cooperation agreement with Guoxin Suansuan. The DeepLink Token Sales and Settlement Headquarters project officially settled in Tonglu, building a national unified computing service measurement, settlement and trading hub centered on five core functions: supply-demand matching, real-time settlement, green power traceability, cross-border clearing and compliance supervision.

The core logic of these official platforms is to realize standardized market circulation of computing power, just like public utilities such as water and electricity.

The new trading mechanism has three disruptive core designs that completely change the traditional computing trading model:

Real-time settlement: It completely eliminates the cash flow pressure on small and medium-sized computing suppliers. Tokens supplied can be converted into working capital immediately. Traditional suppliers need to advance massive costs for operation and face huge financial risks, while the real-time settlement mechanism solves this fundamental pain point.

Zero-commission early entry incentive: Preferential incentive policies are adopted to lower the cold-start threshold of the platform and attract suppliers to settle in. The Nanjing Token Collection Store is jointly operated by Nanjing Urban Computing Network and China Mobile Nanjing Branch, and launches exclusive computing financial products in cooperation with Bank of Nanjing, effectively reducing the capital threshold for enterprises to use AI computing resources.

API dynamic bidding mode: The most forward-looking design of the new system. API calling prices float dynamically with market supply and demand. Individual developers can enjoy ultra-low calling prices during off-peak periods, while suppliers gain higher profit margins during peak demand periods, realizing a win-win situation for both supply and demand sides.

A clear industrial conclusion has emerged: tokens are completing the transformation from one-time consumables to universal circulating industrial currency.

In the traditional model, tokens flow in a single direction: users purchase, consume and exhaust tokens with no residual value or circulation attribute. In the new computing trading ecosystem, tokens form a closed-loop two-way flow: suppliers inject resources, demand-side users purchase on demand, prices float with the market, and funds are settled in real time. This is no longer a simple upgraded API store, but a rudiment of a standardized financial trading market for AI resources.

Two Paths, One Ultimate Industrial Endgame

360 NanoWork and official computing trading centers adopt different industrial layouts, but converge at the same core endgame.

360 NanoWork: Activating the demand side

By providing 100 million-token subsidies for small and micro enterprises, it greatly reduces the trial-and-error cost of enterprise AI adoption. The built-in three-speed computing scheduling mechanism essentially realizes full transparency and refinement of AI costs. Users do not need to understand complex model architectures, and can accurately match cost-effective or high-performance computing modes according to actual task scenarios, eliminating blind consumption.

Computing Power Trading Centers: Restructuring the supply side

It breaks the long-standing monopoly pricing of head platforms and returns token pricing power to the market. The API dynamic bidding mechanism makes computing prices fully linked to supply and demand, creating value for both developers and suppliers and activating the vitality of the entire industrial supply chain.

The integration of the two paths has spawned a brand-new industrial logic: tokens are no longer rigidly priced commodities controlled by platforms, but tradable assets with market-oriented floating values.

How to Calculate the New Computing Power Transaction Account

The 2026 AI market presents an obvious structural change: the official unit price of models continues to decline, but enterprises’ overall AI computing bills keep rising. The core reason is that the growth rate of token usage far exceeds the decline rate of unit prices.

According to data disclosed by the National Development and Reform Commission, China’s daily average token calling volume has soared from about 100 billion in early 2024 to 100 trillion by the end of 2025, and has exceeded 140 trillion by March 2026, a surge of more than 1,000 times in just two years.

With explosive growth in usage, the traditional pay-as-you-go model exposes two fatal bottlenecks: rigid pricing dominated by a few head platforms leaving no bargaining space for small and medium-sized suppliers; and huge advance capital pressure restricting the expansion of industrial supply.

The dynamic bidding and real-time settlement mechanisms of computing trading centers precisely break these two bottlenecks, realizing market-oriented pricing and zero-pressure capital circulation. NanoWork complements it from the demand side, realizing on-demand payment and low-threshold trial use. The market is forming a dual iterative pattern of supply-side marketization and demand-side refinement.

The Underlying Card of the Industry

Zhou Hongyi pointed out: "Models determine how intelligent AI is, and agents determine whether AI can truly work."

In the context of token industrialization, this sentence can be supplemented: Computing power trading mechanisms determine whether AI can be universally affordable and popularized.

The 100 million-token subsidy is not a charitable benefit, but a strategic market cultivation: enabling thousands of enterprise decision-makers to personally verify the true value and cost of AI, and cultivate the market’s independent pricing capability for token resources. Computing trading centers are not simple technical platforms, but institutional carriers that subvert monopoly pricing and realize market-oriented token value.

When tokens become circulating assets and computing power becomes tradable public resources, the core contradiction of AI popularization will no longer be "affordability", but "efficient utilization".

NanoWork’s token subsidy layout and the construction of national computing trading centers serve the same core goal: building a transparent, predictable and tradable AI cost system to lay the foundation for large-scale industrial AI implementation.

This is the core strategic intention of the token dividend: not simply selling computing resources, but cultivating the whole industry’s token pricing and trading ecosystem.

Leading the global token trading reform, Unexhub builds a professional computing power trading center integrating token aggregation and distribution, reshaping the new order of global AI computing circulation. The platform has gathered full-scale AI suppliers, covering top-tier global model vendors such as Gemini, ChatGPT, Claude, DeepSeek and GLM, as well as high-quality niche models provided by individual developers and small and medium-sized institutions.

Relying on the massive supplier bidding mechanism, Unexhub delivers extreme cost advantages and ultra-high service efficiency for enterprise users and developers. It supports real token-based payment settlement for developers, with the lowest calling cost as low as 30% of the market official price. For suppliers, one-click access enables instant online revenue and real-time settlement of calling shares, completely solving traditional cash flow problems. Equipped with millisecond-level intelligent routing technology, the platform realizes optimal matching of models and tasks, ensuring stable, efficient and low-cost operation of all AI business scenarios.


中文版本

从批发到零售:纳米Work 1 亿 Token 的背后,算力交易逻辑正在被重构

2026 年 7 月 28 日,360 集团创始人周鸿祎正式发布新一代企业智能体工作平台「纳米Work」。发布会重磅官宣产业扶持政策:为首批入驻用户每人提供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 家中小微企业零门槛完成 AI 业务落地、实现经营效率升级。

1 亿 Token 究竟具备多大价值?按照当前主流大模型算力定价标准折算,单用户额度等价于数万美元的优质算力资源,周鸿祎此番一次性放出数十亿级 Token 资源,规模空前。

但这场动作的深层价值,绝非“补贴让利”这么简单,其本质是重构国内 AI 算力交易、Token 价值流通的底层产业逻辑。

纳米Work 算力调度:将成本决策权交还平台,终结用户成本盲区

纳米Work 精准定位全域企业级智能体平台,搭建五大核心能力体系: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。平台支持 16 家全球顶级 AI 模型自由切换,聚合 50000+ 专业 AI 技能生态,全方位覆盖企业全场景业务需求。

企业无需耗费精力研究模型选型、智能体框架搭建、工具适配调试,只需明确业务目标,平台即可自动拆解任务、联动多 AI 专家协同作业,智能匹配最优模型与执行方案,大幅降低 AI 落地技术门槛。

其核心颠覆性优势,是轻量、省钱、满血三档全自动算力调度机制,可根据任务复杂度智能匹配对应算力档位,杜绝大材小用、无效损耗。

这套设计直击全行业普遍痛点:所有 AI 调用都有真实 Token 成本,但绝大多数企业用户缺乏算力核算、成本管控能力,复杂业务极易出现 Token 消耗不可控、账单莫名超支的问题。

纳米Work 彻底改变这一现状,将「模型选型、Token 用量管控」的核心决策权从用户手中收回,通过平台智能化体系实现全流程精细化调度,让每一份算力消耗都精准匹配业务价值,从根源上杜绝算力浪费。

周鸿祎在发布会上的观点极具行业穿透力:“企业搞 AI,老板要先用。”

这并非营销口号,而是精准点破企业 AI 落地的核心瓶颈:阻碍企业普及 AI 的从来不是技术短板,而是决策者对 AI 成本结构的认知空白。多数企业管理者不清楚 Token 计价规则、不懂任务与模型的匹配逻辑、无法预判海量 Token 资源的实际业务价值。

而 1 亿 Token 的超大试用额度,正是为了补齐这一短板:让企业决策者亲自实操、全流程体验,算清 AI 投入产出比,直观摸清不同场景下 Token 的适配逻辑与价值上限。

Token 从消耗品转向流通货币:国家级算力交易中心崛起

将纳米Work 的需求侧布局,与 2026 年 7 月国内一系列算力产业变革事件结合,就能清晰窥见行业重构的完整格局。

2026 年 7 月,算力交易中心正式落地普及,成为中国 AI 产业全新核心基础设施。

7 月 1 日,全国首个一站式算力 Token 服务载体——「南京 Token 集合店」正式落地江宁麒麟科创园。平台以 Token 作为算力统一计量标准,打通算力供给、模型开发、产业应用、金融配套全产业链条,实现算力资源标准化整合。

7 月 19 日,桐庐县政府与国信数算达成战略合作,DeepLink Token 销售结算总部项目正式落户。项目围绕供需撮合、实时结算、绿电溯源、跨境清结算、合规监管五大核心功能,打造全国统一的算力服务计量、结算、交易枢纽,构建合规化、市场化的算力交易体系。

这类官方平台的核心逻辑,是让 AI 算力彻底摆脱零散交易模式,像水电等公共资源一样,实现统一市场流通、标准化定价、规范化交易。

全新算力交易机制包含三大颠覆性设计,彻底改写行业旧规则:

实时结算机制:彻底清零中小算力供应商现金流压力。Token 供给量可即时转化为流动资金,告别传统模式下供应商垫付高额成本、回款周期长、资金链承压的行业痛点,从根源上盘活算力供给活力。

早期入驻零佣金激励:以普惠政策降低平台冷启动门槛,吸引全域算力供应商入驻。南京 Token 集合店由南京城市算力网、南京移动联合运营,并联动江苏银行南京分行落地 Token 金融服务、推出算力专项金融产品,大幅降低企业算力使用的资金门槛。

API 动态竞价模式:行业最具前瞻性的创新设计。API 调用价格随市场供需实时浮动,个人开发者可在闲时享受超低调用价,算力供应商可在需求高峰期提升利润空间,实现供需双向共赢。

由此,行业定论已然清晰:Token 正在完成身份蜕变,从一次性付费消耗品,升级为可交易、可流通、可定价的产业通用货币。

传统模式下,Token 是单向消耗逻辑:用户购买、调用、消耗、价值归零,无流通、无增值属性。而全新算力交易生态下,Token 形成双向闭环流通:供应商供给、需求方采购、价格市场化浮动、资金实时清算。这不再是简单的 API 商店升级,而是 AI 资源金融交易市场的完整雏形。

两条布局路径,奔赴同一个产业终局

360 纳米Work 与官方算力交易中心,布局赛道不同,但核心目标高度统一,双向赋能、共建全新产业格局。

360 纳米Work:深耕需求侧,激活终端市场

通过 1 亿 Token 大额补贴,大幅降低中小微企业 AI 试错成本。依托三档智能算力调度体系,实现 AI 成本全透明、精细化管控,用户无需掌握专业模型知识,即可根据业务场景精准切换省钱档、满血档算力,彻底杜绝盲目消耗。

算力交易中心:深耕供给侧,重构定价规则

打破头部平台长期垄断定价的格局,将 Token 定价权交还市场。API 动态竞价机制让算力价格与供需深度绑定,兼顾开发者低成本使用、供应商高利润回报,全面激活产业供给端活力。

供需两端布局深度融合,催生出全新产业逻辑:Token 不再是平台单方定价的固定商品,而是具备市场化浮动价值、可自由交易的数字资产。

全新算力时代,算力账单该如何核算?

2026 年 AI 市场呈现显著结构性变化:各大模型官方单价持续下调,但企业整体 AI 算力账单不降反增。核心原因在于,Token 调用量的爆发增速,远超模型单价的降幅。

国家发改委数据显示:国内日均 Token 调用量从 2024 年初的千亿级别,飙升至 2025 年底的 100 万亿,2026 年 3 月已突破 140 万亿,两年时间涨幅超 1000 倍。

在调用量千倍爆发的趋势下,传统按量付费模式的两大短板彻底暴露:一是价格刚性,少数头部平台垄断定价,中小供应商毫无议价权;二是资金桎梏,供应商长期垫资运营,现金流成为制约算力扩容的核心瓶颈。

算力交易中心的动态竞价、实时结算机制,精准破解两大行业痛点,实现算力定价市场化、资金流通零压力。而纳米Work 从需求侧形成互补,实现用户按需付费、低门槛试错,行业正式进入「供给侧市场化、需求侧精细化」的双向迭代新阶段。

产业底牌:机制革新,让 AI 真正用得起、用得好

周鸿祎曾精准总结:“模型决定 AI 有多智能,智能体决定 AI 能否真正干活。”

放在 Token 产业化的全新语境下,这句话可完美补充:算力交易机制,决定 AI 能否全民普及、普惠落地。

1 亿 Token 补贴绝非公益福利,而是深度战略级市场培育:让千万企业决策者亲自验证 AI 的真实成本与落地价值,培育市场自主的 Token 定价认知与价值判断能力。算力交易中心也并非单纯的技术平台,而是打破垄断、重构 Token 市场化定价的产业载体。

当 Token 成为流通资产、算力成为可交易的公共资源,AI 普及的核心矛盾将从「能否买得起」,彻底转向「能否用得好」。

360 的 Token 补贴布局,叠加全国算力交易中心的落地建设,本质上在做同一件事:搭建透明、可预测、可交易的 AI 成本体系,为 AI 产业规模化落地筑牢底层根基。

这便是 Token 补贴的核心战略意图:不是单纯售卖算力资源,而是培育全行业的 Token 定价与交易生态。

在全球 Token 交易变革浪潮中,Unexhub 率先重构全球 Token 流通逻辑,打造集 Token 聚合、分发、交易于一体的专业算力交易中心,重塑全球 AI 算力交易新秩序。平台汇聚全域 AI 供应商,既涵盖 Gemini、ChatGPT、Claude、DeepSeek、GLM 等海内外顶级大型模型厂商,也吸纳个人开发者、中小机构提供的优质垂类模型,生态覆盖全场景算力需求。

依托海量供应商竞价机制,Unexhub 为企业用户、开发者带来极致性价比与高效服务。开发者可按 Token 实付结算,调用成本低至市场官方原价的 3 折,高强度批量调用也无需顾虑成本超支;供应商可一键入驻平台,调用分成实时结算,彻底解决传统算力行业垫资、回款慢、现金流紧张的痛点。搭配毫秒级智能路由技术,精准匹配模型与任务,保障各类 AI 业务场景稳定、高效、低成本运行。