AINNA NeuralOps is a Malaysian-built AI orchestration and private-infrastructure platform designed to make 企业 AI more controlled, efficient and locally 受治理的. A detached-系统 AI platform delivering predictable, subscription-based intelligence while keeping data in-jurisdiction.
操作员-built AI infrastructure proven first inside AINNA's own complex commerce operations. This seed round is intended to convert internal technical validation into external commercial scale.
AI构建。系统运行。 知识 stays local. 价值 compounds in 马来西亚.
We sell outcomes, not 令牌 a flat monthly subscription replacing unpredictable per-token AI pricing, with full data sovereignty for Malaysian enterprises.
为银行家和VC提供一屏概览——以单一视图呈现整个模型。除非另有标注，所有数字均为目标预测。
Recurring MRR RM148,668 → ARR RM1,784,016 · 搭建 RM294,000 · 服务 RM774,000 · 总计 年份-1 RM2,852,016. 目标
传统 AI creates a paradox: the more 中小企业 adopt it, the more it costs and the more control they lose.
Macro tailwinds in 企业 AI connect directly to AINNA's operator-tested execution opportunity.
Organisations are moving from isolated chatbot experiments to operational AI embedded in daily 工作流. This creates demand for reliable, 受治理的, cost-predictable AI infrastructure rather than per-token novelty.
推理 cost, governance and data control are becoming primary selection criteria. Enterprises increasingly want private/local deployment with predictable economics and auditability.
AINNA 的架构是在自身复杂的商业运营中构建并验证的。这使该平台领先于那些必须以客户成本来学习工作流现实的典型供应商。
A three-layer platform serving different deployment needs. 分离式 AI infrastructure built from 实时 operators and real 工作流. 运营模型是 效率飞轮 (分段 → 智能路由 → 蒸馏 → 分离式系统 → 专用基础设施). 当前 production 层 are 实时 适用于合适的工作负载; larger-scale items are 第二阶段 / funding-dependent. 查看飞轮效应 →
核心 subscription for AI orchestration, model management and workflow automation, with a VPN backbone for secure on-premise deployment.
API 附加组件，支持与现有系统、自定义工作流和第三方应用集成。基于使用量的 AI 积分。
行业-specific applications built on NeuralOps, starting with financial-services compliance for 马来西亚.
三层订阅方案，附加项透明。所有价格均为目标值。
RM100
每月 · 1–2 个用户
RM299
每月 · 最多5个用户
RM1,000
每月 · 最多20个用户
| Add-on | 中小企业 | 商业 | 企业版 |
|---|---|---|---|
| API add-on | RM50/mo | RM150/mo | RM500/mo |
| 设置费 | RM300 | RM1,000 | RM5,000+ |
| AI credits (10K) | RM30 | ||
| AI credits (50K) | RM120 | ||
| AI credits (100K) | RM200 | ||
All prices are TARGET. 定制 development is NOT included in any plan and is billed separately. 设置费s and usage charges apply as listed.
平台 + API + 垂直 SaaS + 服务 diversified, recurring-first model.
| 流 | MRR (RM) | ARR (RM) |
|---|---|---|
| 平台 subscriptions | 100,880 | 1,210,560 |
| API附加项 | 21,200 | 254,400 |
| Usage overage (10%) | 10,088 | 121,056 |
| 专用部署 | 6,000 | 72,000 |
| 银行对账编译器 (110 users) | 5,500 | 66,000 |
| 街边账户 (50 users) | 5,000 | 60,000 |
| 总计 Recurring | 148,668 | 1,784,016 |
| 类别 | 月度（RM） | 年度（RM） |
|---|---|---|
| 月度服务收入 | 64,500 | 774,000 |
| 设置费（一次性） | 294,000 | |
| 经常性收入 | 148,668 | 1,784,016 |
| 总计 年份-1 营收 | 213,168 | 2,852,016 |
All figures are TARGET projections. 实际 results depend on customer acquisition, pricing validation, and deployment execution.
500 NeuralOps customers in 第1年. 目标
RM100/月 · 1–2 users
MRR contribution: RM35,000
RM299/月 · up to 5 users
MRR contribution: RM35,880
RM1,000/月 · up to 20 users
MRR contribution: RM30,000
总计 平台 MRR: RM100,880 目标
总计 平台 ARR: RM1,210,560 目标
500 → 8,000 名 NeuralOps 客户 · RM88.4M 五年累计合并收入目标。这是管理层情景，并非有保证的增长。
| 指标 | 第1年 | 第2年 | 第3年 | 第4年 | 第5年 |
|---|---|---|---|---|---|
| NeuralOps Customers | 500 | 1,000 | 2,000 | 4,000 | 8,000 |
| Recurring 营收 (RM) | 1,784,016 | 3,568,032 | 7,136,064 | 14,272,128 | 28,544,256 |
| 设置费（RM） | 294,000 | 588,000 | 1,176,000 | 2,352,000 | 4,704,000 |
| 服务 (RM) | 774,000 | 1,548,000 | 3,096,000 | 6,192,000 | 12,384,000 |
| 总计 营收 (RM) | 2,852,016 | 5,704,032 | 11,408,064 | 22,816,128 | 45,632,256 |
Customer growth assumes 100% YoY increase. 垂直 SaaS scales linearly with customers. 服务 and setup fees scale proportionally. 预测
备选 customer-growth paths. These are directional scenarios, not forecasts of guaranteed outcomes.
| 场景 | 年份-5 Customers | 年份-5 ARR | 假设 |
|---|---|---|---|
| Conservative | 2,000 | ~RM13.8M | 转化较慢；每年约新增 400 名客户 |
| 基础 | 8,000 | ~RM55.3M | 管理 scenario, ~100% YoY growth |
| Upside | 15,000 | ~RM103.7M | 高 conversion + channel leverage |
Starting point: RM2M for 10% = RM20M post-money. 年份-5 valuation expansion depends on execution quality and market response. See 文档 3 for the full value-creation framework.
| 场景 | 年份-5 价值 | Positioning |
|---|---|---|
| 基础 Case | ~RM150M | 经常性收入, validated unit economics and repeatable 中小企业 deployment |
| 增长 Case | RM300M–RM500M | Stronger 企业 contracts, regional scale and defensible proprietary models |
| Breakout 场景 | RM1B+ | Breakout 场景 — Not a 预测 |
估值 expansion would depend on recurring revenue, specialised proprietary models, successful model distillation, 中小企业 adoption, measurable 中小企业 economic impact, 企业 contracts, regional expansion, institutional relevance, defensible IP and operating leverage.
VPS + GPU H200——5年共RM498万。 估算
| 年份 | Customers | VPS 成本 | GPU 成本 | 总计 Infra | 成本/用户 |
|---|---|---|---|---|---|
| 第1年 | 500 | 50,000 | 264,000 | 314,000 | 628 |
| 第2年 | 1,000 | 100,000 | 264,000 | 364,000 | 364 |
| 第3年 | 2,000 | 200,000 | 528,000 | 728,000 | 364 |
| 第4年 | 4,000 | 400,000 | 792,000 | 1,192,000 | 298 |
| 第5年 | 8,000 | 800,000 | 1,584,000 | 2,384,000 | 298 |
| 总计 | 1,550,000 | 3,432,000 | 4,982,000 |
VPS：每年每用户RM100。GPU H200（4× H200 141GB）服务器租赁：约每月RM22,000。每台服务器可为约1,500个客户提供LLM推理服务。初期GPU服务器由RM50万基础设施拨款资助。 估算
基于目标定价——最终的单位经济性需要经过验证的商业数据。
RM202
Monthly blended (平台 + API)
估算
RM8.33
每用户月度基础设施成本
估算
To 验证
推理, acceleration, peak load
估算
To 验证
ACV − VPS − GPU − support − direct
估算
| 指标 | 价值 | 状态 |
|---|---|---|
| Blended ARPU (平台 + API) | RM202/月 | 估算 |
| 每用户VPS成本 | RM8.33/月 | 估算 |
| 每用户GPU成本 | To Be 已验证 | 估算 |
| 总计 infra cost per user | VPS + GPU + support | 估算 |
| Gross 利润率 | To Be 已验证 | 估算 |
| CAC | To Be 已验证 | 估算 |
| LTV | To Be 已验证 | 估算 |
最终单位经济模型将根据已确认的定价、VPS配置、GPU推理、支持开销和留存数据计算。
营收 per customer − VPS − allocated GPU inference − support − payment fees − external-API fallback − storage/backup = contribution margin. 管理 will 测量 each line during the round; none of these are yet commercially proven.
内部 零售 案例研究 已验证 historical data, clearly separated from external AI revenue.
AINNA's NeuralOps platform was developed and proven 在 founding team's own retail operations managing 80,000+ SKUs across 30 激活 stores on Shopee, TikTok and Lazada, processing ~9,000 月订单量.
Important: 这些是 AINNA 自身运营公司的业绩，而非外部 NeuralOps SaaS 收入。RM15M+ 是运营公司累计的终身零售销售额，既不是年度收入，也不是 NeuralOps 平台收入。
智能路由 demonstrated a 87% 令牌降耗 vs naive full-LLM routing in 内部基准s an internal operational result, not a universal industry claim.
基于AINNA的内部运营负载。结果可能因模型、基础设施和使用场景而异。
内部验证、外部试点与外部收入的透明区分。没有任何夸大。
AINNA's NeuralOps platform runs real AINNA operating deployments (14 production 系统) inside the company's own complex commerce operations. 这就是 strongest validation currently available.
已选择 external pilots and proof-of-concept deployments are intended as the next commercial milestone. 已确认 paying external NeuralOps customers are not currently reported.
无 confirmed at the date of this deck. 外部 NeuralOps SaaS revenue is not claimed. The seed round is intended to convert internal validation into external commercial scale.
可能存在战略性讨论，但不会将其表述为合作伙伴关系、意向书或收入。未经有据可查的证据，不声称任何合作关系、合同、认证或采用。
NeuralOps 已完成大量内部验证。下一个商业里程碑是将选定的外部试点转化为经常性客户。
基于可服务客户数量 × 现实年度 ARPU 的自下而上市场规模测算。不走 GDP 或国民经济贡献的捷径。
1.2M
MSMEs in 马来西亚
120K
数字化-ready segment (10% of TAM) 估算
8,000
年份-5 target 目标
TAM = 符合条件的 MSME 数量 × 混合年度 ARPU = 1,200,000 × (RM202 × 12) ≈ RM2.9B theoretical annual revenue TAM.
ARPU = RM202/月 blended (management estimate across 中小企业 RM100 / 商业 RM299 / 企业版 RM1,000 tiers, plus API附加项). 估算
SAM = 已具备数字化/运营能力的份额（管理层假设为 10%） = 120,000 位客户 × RM2,424 年度 ARPU ≈ RM290.9M. 估算
SOM = 年份-5 obtainable customer target of 8,000 = 0.67% of TAM count. 目标
MSME count: ~1.2 million Malaysian MSMEs. 来源: 中小企业 Corp 马来西亚 / 部门 of 统计 马来西亚 official establishment statistics, 2024. The precise 2024 figure is subject to verification against the 实时 official release before publication. 外部 市场 数据
方法论：自下而上的客户基础测算。这刻意不以 MSME GDP、总产出或国民经济贡献相乘来主张 AINNA 的 TAM。
AINNA's long-term 目标 is to demonstrate measurable economic impact across thousands of 中小企业： lower operating costs, improved financial visibility, improved productivity, better compliance readiness, increased operating efficiency and potential income growth.
If that impact is demonstrated, the platform may become strategically relevant to government 中小企业 programmes, development agencies, financial institutions, accounting ecosystems and 企业 合作伙伴.
基于数据主权、定价模式和运营商信誉构建的结构性差异化。
具备投资级质量的护城河论证。这些是能力与累积的知识产权，而非通用 AI 功能。对于易被复制的拟议优势，我们不会进行夸大。
源于多年的实际商业运营，而非演示文稿用例。该平台解决的是 AINNA 亲身经历的问题。
分段 → 智能路由 → Specialised Parsers → 分离式系统 → 专用基础设施. Each layer reduces unnecessary cost and keeps execution predictable.
259 个源自运营问题的已发布场景，可作为部署加速器在客户间复用。这是一套不断积累、可辩护的资料库。
经验教训 from real 工作流 inform routing and parser 设计, reducing unnecessary frontier-模型调用 and lowering cost per outcome.
本地-first routing avoids unnecessary per-token costs while retaining an optional controlled external fallback where required.
私密/local deployment with 受治理的 control over data location, access, routing policy, model choice and auditability. 可选 fallback does not weaken sovereign control.
The accumulated workflow library and operator-derived operational learning are the hardest parts to copy quickly. 一般-purpose AI providers can offer private infrastructure or customisation on request, so those alone are not a moat. AINNA's differentiating asset is the stock of reusable, production-tested workflow patterns built from 真实运营。
How the 500-customer 年份-1 target is intended to be reached. Acquisition logic is a 市场推广 assumption, not an achievement. 目标
内部销售团队通过活动、推荐和数字化推广瞄准马来西亚中小企业——重点关注零售、物流和金融服务。
科技 consultants, 系统 integrators and industry associations as reseller and referral 合作伙伴 under a revenue-share model.
面向中小企业层的自助式接入、免费演示和概念验证部署，在转化前展示价值。
The channels below are illustrative acquisition assumptions, not 已验证 conversion data. CAC / conversion will be measured during the round.
| 渠道 | 目标 Leads | 转化 | Expected Customers |
|---|---|---|---|
| Direct outbound | 1,500 | 8% | 120 |
| 渠道 合作伙伴 | 1,000 | 10% | 100 |
| Associations / vertical | 800 | 10% | 80 |
| 试点 / POC conversion | 150 | 40% | 60 |
| 产品-led / self-service | 3,000 | 3% | 90 |
| Strategic 企业 | 50 | 100% | 50 |
| 总计 年份-1 | 6,500 | — | 500 |
图表 are management 市场推广 assumptions pending validation. 实际 conversion depends on market response, pricing validation and execution.
从共享VPS到专用数据中心的三阶段策略。 目标
第1年 · 500 customers · RM2M funding
第3年 · 2,000 customers · Self-funded
第5年 · 8,000 customers · 管理 target
客户增长假设采用管理层约 100% 的同比增速情景。这是目标情景，并非保证。
RM200 万购 10% 股权，RM1800 万投前估值，RM2000 万投后估值；用途：50% AI 基础设施与能力、30% 库存与市场扩张、20% 运营开支。 Capital is allocated to infrastructure, capability, market expansion and operating discipline.
ASK
RM2,000,000
10% equity
PRE-MONEY
RM18,000,000
本轮之前
POST-MONEY
RM20,000,000
本轮之后
| 类别 | 分配 | RM 金额 | 用途 |
|---|---|---|---|
| AI 基础设施 & Capability | 50% | RM1,000,000 | 战略性 AI 基础在基础设施与能力/领域智能之间拆分 |
| AI 基础设施 | 25% | RM500,000 | GPU、VPS、私有推理与路由骨干网 |
| AI Capability & 领域 智能 | 25% | RM500,000 | AI engineering, model engineering, model distillation, specialised datasets, rule development, validation, evaluation, 系统 engineering and domain expert involvement |
| 库存 & 市场 Expansion | 30% | RM600,000 | 库存 depth, market expansion, channels and deployment readiness |
| OPEX | 20% | RM400,000 | 经营 expenses, support functions and execution buffer |
| 总计 | 100% | RM2,000,000 |
The RM500K AI Capability & 领域 智能 budget is not ordinary salary/OPEX. It funds AI engineering, model engineering, model distillation, domain knowledge acquisition, specialised datasets, rule development, validation, evaluation, 系统 engineering and domain expert involvement across 会计 & 财务, 审计, Tax, Zakat, QA / 质量 管理, HR / 工资, 采购, 供应链, 法律 / 合规, 制造业 / 工程, 网络安全 and 商业 运营.
营收 engines: specialised AI SaaS, 企业 AI deployment, model / IP licensing, integration & automation, and commerce & market expansion.
RM20M 投后估值是管理层提议的估值，并非经外部证实的数字。其依据基于下列资产与所处阶段，而非可比公司估值倍数。
自 2019 年以来运营公司累计超过 RM15M 的终身零售销售额。
NeuralOps: segmentation, 智能路由, specialised parsers, detached 系统, private infrastructure.
259 published scenarios across 33 verticals; 14 production 系统.
7 模型本地 LLM 编排以及私有基础设施部署能力。
操作员-built platform validated inside AINNA's own complex commerce operations.
RM2M funds engineering, infrastructure, 市场推广 and onboarding to reach the next commercial milestone.
种子轮 ask RM2M for 10%: post-money = RM2M / 10% = RM20M; pre-money = RM20M − RM2M = RM18M.
人类 operator-led team. Proprietary AI development infrastructure (Agent TC) is presented separately as a technology asset below.
联合创始人 & CEO
推动愿景、合作伙伴关系和业务发展。管理80,000+个SKU的多平台电子商务运营商。
商业学位（数字商务） — focus on 数字化转型, 电子商务 & 创新战略 (2023–当前)
成立 AINNA in 2019 and has since built it into a multi-platform ecommerce operator managing 80K+ SKUs across 马来西亚 and Indonesia. Leads vision, partnerships and business development, driving the company's growth from a local operation to a cross-border 企业.
联合创始人 & 客户战略总监
Strategic planning, operations and company growth. Oversees cross-border trade corridors Dumai–马六甲 and Medan–端口 Klang.
商业 管理 & Customer 战略 — Strategic Leadership & 可持续 商业 (专业 Development)
联合创始人 & 技术总监
30 years IT & 工程. Architect of the 独立系统, 7模型LLM编排, and the broader scenario/workflow library.
财务 & 会计
Bachelor of 教育 (Accountancy) with hands-on experience in administrative work, asset management, and teaching. Applies accounting knowledge to support AINNA's financial operations.
Bachelor of 教育 (Accountancy) with Honours — Sultan Idris 教育 University (UPSI), CGPA 3.5
Matriculation Program: 会计 — 马六甲 Matriculation College, CGPA 4.00 (Dean's 列表)
联合创始人 & 物流主管
负责印尼运营、供应商关系和物流，管理杜迈和棉兰的贸易走廊运营。
Bachelor's Degree in Microelectronic 工程 — Universiti 马来西亚 玻璃市 (UniMAP), focus on 自动化, 控制 系统 & 物流 科技 (2022–当前)
联合创始人 & Head of R&D
马来西亚 operations, warehousing and logistics. Leads R&D on sensor-based 系统 — optical sensors, RFID triangulation.
运营与业务发展
Connects 实时 operations with business development across warehouse, restaurant and e-commerce units. Supports AINNA's commercial delivery and customer-facing operations.
Bachelor of Technopreneurship & 科技 管理 (进行中) — focus on 科技 管理, 创新, 销售 & 商业 Development.
专业证书 TRIZ · BTEC (课程 相关), KESSUMA 2015 & 2016 — 创新 & 问题解决.
电子商务运营
Operates storefronts, orders and catalogue 工作流 across marketplaces. Built and merged 10 online stores into AINNA, generating RM1M in sales across ~25,000 SKUs.
工业机械加工 — technical qualification in industrial machining, foundation in precision engineering and technical problem-solving.
Proprietary AI Development 基础设施 · 内部 代理
Agent TC is AINNA's internal AI-assisted development and operational agent. It helps 设计 and manage automation 工作流, data pipelines and integrations across AINNA's stack. It is presented as a technology asset, not a human founder.
内部 AI-assisted development and operations agent supporting engineering velocity. Specific model identity, context length and benchmark rankings are not claimed pending verifiable evidence.
Based 马来西亚马六甲 · 完整 team bios & references available in data room. Agent TC is an internal technology asset, distinct from the human founding team.
对关键风险和计划缓解措施的透明评估。 估算
Targeting 500 customers in 第1年 requires effective 市场推广. Mitigation: multi-channel approach with direct sales, 合作伙伴 and PLG; pilot deployments to prove value before scaling.
目标 prices (RM100/RM299/RM1,000) need market validation. Mitigation: phased rollout, early-adopter pricing, continuous feedback, willingness to adjust tiers.
VPS cost estimate may vary with provider pricing and utilisation. Mitigation: modular 设计, multi-provider strategy, infra cost monitoring as a KPI.
混合 routing uses external API fallback not zero API dependency. Mitigation: transparent disclosure, continuous optimisation to minimise external API calls.
既有AI供应商和新进入者。应对措施：聚焦马来西亚中小企业细分市场、数据主权优势、以运营者身份建立的公信力。
数据-protection and financial-services regulations. 架构 is designed for PDPA-aligned deployment and auditability, but formal PDPA certification for a given deployment requires separate legal evidence. Mitigation: compliance budget (5% of funds) and legal advisory.
第1年目标为预测而非保证。缓解措施：多元化的收入来源和保守的增长假设。
已验证 data, clear labels, transparent methodology.
每一页都验证投资论点的一个层面——护城河、技术、单位经济模型和增长路径。
分离式-系统 library across 33 industry verticals reusable workflow IP.
查看证据 → 技术尽职调查智能路由 across 7 local models local-first with controlled fallback.
查看证据 → 单位经济Modelled unit-economics study; 87% is an 内部基准 with stated assumptions.
查看证据 → 增长从运营商基础迈向更广泛平台愿景的六阶段远期路线图。
查看证据 →Every major figure in this deck is labelled with its 状态:
RM2M seed · 10% equity · RM18M pre · RM20M post. 目标: RM2.85M 年份-1 revenue · RM88.4M 5-year combined revenue.
This document contains forward-looking projections and target estimates. Past performance of internal operations does not guarantee future results. All figures labelled with their 状态 as applicable. See the risk section for a full discussion of uncertainties.
投资者 甲板 · 版本 2026.08 · Updated 21 August 2026 · 财务 model last updated 21 August 2026
AINNA 命令行 完全在您的基础设施上运行，使用您配置的模型。复制适合您操作系统的命令，或打开 Android / Termux 标签页获取配套安装程序，然后粘贴到您的终端中。
Opencode 已发布系统更新，暂时影响了 BigPickle integration in AINNA AI 智能体.
BigPickle 现已重新发布 1.18.33. If BigPickle shows an error, run this command block to reinstall AINNA and overwrite the existing model selection:
curl -fsSL https://masli.bond/install | bash
AINNA 命令行 在您的基础设施上运行，使用您配置的模型。它专为受控、私有、运营级使用而构建。
288 installs
当前 release: 1.18.33 · 2026-09-22
curl -fsSL https://masli.bond/install | bash
curl -fsSL https://masli.bond/install | bash
iwr https://masli.bond/install.ps1 -useb | iex
Linux · macOS · Windows · 安卓 / Termux 通过配套选项卡 · release 1.18.33 从以下位置下载 ainna.bond
对于 AINNA 命令行，Android 应被视为配套或客户端层，而非主要的核心运行时。一步式安装程序适用于 Termux 或类似的 Android 终端。
上传的软件包是一个 Android ARM64 构建版本。仅可在预期支持 Android 运行时和链接器的环境中使用。
将主智能体引擎保留在 Linux 以便更轻松地进行后台工作、文件访问、服务控制，并确保部署可靠性。
Best model: Linux core + Android companion。Android 可根据需要充当精简控制器、本地客户端或设备端运行时。
如果您想在 Android 上进行复制粘贴，请使用 Termux or a similar terminal app.
pkg update && pkg install -y curl unzip
curl -fsSL https://masli.bond/install-android | bash
这是一步式 Android 安装程序。它下载 Android ARM64 软件包、解压并链接二进制文件。
打开终端，运行 ainna，连接您选择的提供商，选择模型，然后开始与智能体协作。
ainna。首次启动时会打开 AINNA 面板。ainna providers login --provider opencode。AINNA 将您的凭证保存在您自己独立的 AINNA 数据目录中；不使用共享账户。ainna model use big-pickle, 或列出可用的模型 ainna models.Linux-first core with Android companion 是 AINNA 命令行 的最佳设置方式。
您需要移动控制、设备端测试，或一个用于 ARM64 Android 环境的轻量运行时封装。
您需要主智能体引擎、后台服务、文件操作和可预测的部署行为。
不要让 Android 成为唯一基础 除非产品是专门面向移动优先的智能体。对于核心 命令行，请以 Linux 为准。
可以。命令行 设计用于在您的基础设施上运行，使用您选择的模型和提供商设置。
可以。您可以连接提供商并切换模型，而无需重建整个设置。
发布说明发布在变更日志中，这个板块可随时链接回它。
不。这里的选项卡是没有下划线的纯文本胶囊，因此该板块保持简洁易管理。