国际数据公司(IDC)最新发布的全球人形机器人跟踪报告显示,2026上半年,全球人形机器人出货量接近2.5万台,同比增长432.1%,产业发展进一步从技术验证向商业化应用探索阶段迈进。应用场景持续扩展,加速向工业制造、零售、交通物流、休闲旅游及家庭消费延伸;与此同时,行业竞争也从本体性能逐步转向具身智能综合能力。IDC预计,2030年全球人形机器人出货量将超过75万台,市场将进入规模化应用加速发展阶段。

全球人形机器人市场延续爆发式增长,2030年有望超75万台

2026年以来,全球人形机器人市场延续高速增长态势,产业发展进一步从技术验证迈向商业化应用探索。IDC数据显示,2026年上半年,全球人形机器人出货量近2.5万台,同比增长432.1%;市场规模超过7.4亿美元,同比增长322.7%。

市场高速增长的同时,人形机器人产业的供需两端也正在发生积极变化:一方面,工业、物流等应用探索持续加速,头部厂商不断扩充生产能力;另一方面,中国市场开始出现面向个人消费者的小型人形机器人,家庭及个人市场逐步成为新的潜在增长空间。此外,具身智能大模型快速迭代,机器人对复杂环境的理解、决策和任务执行能力持续提升,为人形机器人拓展更多应用场景提供技术支撑。

基于供给能力持续提升、应用场景加速拓展以及消费级市场逐步打开,IDC上调了全球人形机器人长期市场预测,预计2030年全球人形机器人出货量将超75万台左右,较此前预测上调约50%。IDC认为,未来全球人形机器人市场将逐步形成工业应用、商业服务与消费市场多元驱动的增长格局,市场也将从“技术可行”进一步迈向“规模化应用与商业价值验证”。

中国成为全球人形机器人市场增长的重要引擎

IDC数据显示,2026年上半年,中国人形机器人市场保持快速增长,出货量超1.9万台,同比增长426.3%,占全球市场比重约77.9%,成为全球人形机器人市场增长的重要驱动力。

中国市场的快速发展,一方面得益于政策持续推动。2026年以来,地方政府、产业园区及国有企业持续加大对人形机器人产业的支持力度,并通过建设创新平台、发布应用场景和推进采购部署,加速人形机器人从技术研发向实际应用落地。

另一方面,中国具备较为完整的人形机器人产业链和活跃的厂商生态。整机厂商、零部件企业、AI模型企业及系统集成商加速协同,推动机器人核心零部件、整机制造和智能化能力持续迭代,产品成本和量产效率不断改善。

与此同时,中国厂商正在成为全球人形机器人市场增长的重要力量。2026年上半年,中国厂商出货量占全球市场超95%,多家中国厂商进入全球出货量领先阵营,头部厂商在产品迭代、量产交付和应用落地方面均保持较快推进。

从市场发展路径来看,中国正在形成“政策推动、产业协同与应用牵引”共同驱动的市场发展模式。随着汽车、3C、物流、能源等行业应用持续推进,中国有望继续保持全球人形机器人市场的重要增长地位,并成为推动全球产业规模化发展的关键市场。

应用场景持续扩展,人形机器人加速进入多行业与家庭消费市场

随着人形机器人产品成熟度提升和应用探索持续深入,市场应用边界正在不断拓展。2026上半年,科研教育、表演展示、政府(数采中心)等应用场景合计出货量占比已由2025年全年的83.8%下降至69%,应用结构进一步多元化。其中,科研教育场景持续下沉,从高校科研逐步向职业院校、中小学人工智能教育延伸;表演展示类应用占比有所下降;政府数采中心建设仍是当前重要出货方向。

与此同时,人形机器人开始加速向工业制造(汽车制造、高科技与电子、其他工业制造、航空航天等)、零售、交通物流、休闲旅游等更多行业延伸,厂商与行业用户持续推进场景验证和商业化应用探索。

工业制造是当前人形机器人应用探索的重要方向。汽车制造、高科技与电子及其他工业制造等行业具有较明确的智能化需求和半结构化作业环境,为人形机器人进入真实生产流程提供了较多落地机会。智元机器人AGIBOT、优必选Ubtech、银河通用Galbot等厂商重点布局工业领域。

与此同时,零售、交通物流、休闲旅游等服务行业开始出现更多应用探索,人形机器人正在从进一步向服务型工作场景延伸。

值得关注的是,家庭消费市场也开始出现实际出货。厂商推出小尺寸、低成本人形机器人,并通过电商等渠道触达个人消费者,提供儿童教育、个人陪伴等功能。随着2026下半年个人陪伴机器人进一步拉开序幕,消费级产品有望推动人形机器人的用户群体从企业和机构向个人消费者扩展。

整体来看,人形机器人正从少数行业的技术验证,逐步进入更多行业和终端消费场景,应用范围的持续扩张正在成为推动市场规模增长的重要因素,也进一步拓宽了人形机器人的市场边界。

具身智能技术加速演进,厂商竞争从本体向智能化能力延伸

人形机器人正在进入从“硬件能力竞争”向“具身智能综合能力竞争”转变的关键阶段。随着大模型、世界模型、视觉语言动作模型(VLA)以及机器人数据闭环技术快速发展,行业竞争重点正从机器人结构设计、运动性能和硬件参数,逐步转向机器人对物理世界的理解、任务规划、操作执行和持续学习等综合能力。

从技术发展来看,人形机器人的具身智能能力并非单一模型能力,而是由多个维度共同构成,主要包括:

  • 认知智能(Cognitive Intelligence): 物理世界感知、理解、推理及任务规划能力;
  • 运动智能(Motion Intelligence): 运动生成、全身协调、动态移动及复杂环境适应能力;
  • 操作智能(Manipulation Intelligence): 感知与动作闭环、灵巧操作及工具使用能力;
  • 交互智能(Interaction Intelligence): 人机交互、环境交互及多智能体协同能力;
  • 数据智能(Data Intelligence): 真实采集数据、仿真数据、数据治理及持续学习能力;
  • AI工程能力(AI Engineering): 模型部署、软硬件协同及AI系统工程化能力;
  • AI生态(AI Ecosystem): 模型、数据、开发工具、硬件及应用合作伙伴生态;
  • AI部署成熟度(AI Deployment Maturity): 从实验室验证到真实场景部署、持续运营及规模化应用的能力。

随着VLA、世界模型等技术持续演进,上述能力正在加速融合。厂商也开始从单一技术突破,逐步构建覆盖模型、数据、工程和应用的完整技术体系,推动机器人从执行预设程序向理解环境、自主规划和完成复杂任务演进。

当前,不同厂商在具身智能技术路线和能力建设积累上呈现差异化发展,部分厂商更加重视基础模型与认知能力,部分厂商强化运动控制智能,也有厂商围绕数据闭环、AI工程化和实际部署持续投入。但随着技术体系逐步成熟,单一维度的能力优势将难以形成长期竞争壁垒,厂商需要逐步补齐认知、运动、操作、交互、数据、AI工程、生态及部署等多维能力。未来,人形机器人厂商的竞争将从当前的差异化技术路线,逐步走向具身智能多维综合能力与规模化落地能力的全面竞争。

IDC中国研究经理李君兰认为,整体来看,2026 年上半年全球人形机器人市场延续高速增长态势,产业正从技术验证进一步迈向商业化应用探索。中国凭借政策推动、产业协同与应用牵引,成为全球市场增长的重要引擎;应用场景从科研教育、表演展示向工业制造、零售、交通物流、休闲旅游及家庭消费持续延伸;厂商竞争也从本体性能逐步转向具身智能综合能力。随着具身智能大模型快速迭代、供应链成熟与成本持续改善,人形机器人有望在更多真实场景中完成商业价值验证,并在 2030 年前后进入规模化应用加速发展阶段。未来,谁能更快打通“技术—数据—工程—场景”的闭环,谁就更有机会在下一阶段竞争中占据主动。

进一步交流

以上为Worldwide Annual Humanoid Robotics Tracker报告核心数据洞察,机器人产品及行业分类及定义请参考《IDC’s Worldwide Annual Humanoid Robotics Tracker Taxonomy, 2026》。如需了解细分市场完整数据、厂商格局等详情,欢迎与IDC团队联系。请点击此处与我们联系。

Lily Li

Lily Li - Research Manager

Lily is the Research Manager for China Robotics and Embodied Intelligence, specializing in market research on embodied intelligent robots. She has long focused on the development trends of China’s embodied intelligence robotics industry, systematically studying the evolution of robot hardware,…

A CTO explaining, for the third time this quarter, why the company runs five different models across four teams describes most enterprise AI programs heading into the new year. No one outside the AI group can name all five with confidence. Eighteen months ago, the question was simple: which frontier model to standardize on. Today the answer is a constellation of small, regional, and domain-specific models, each chosen for a defensible reason and none of them mapped against the others.

What worries the CTO now is accountability: who owns the five models already in production, and what happens to the roadmap when three of them get deprecated in the same product cycle. That’s a governance and org-design question, and right now it’s being handled like a procurement one.

Match the model to the mission

The instinct to default to the biggest, most capable frontier model for every task is the assumption worth retiring. IDC tracked 66 open language models released in 2024 alone, with 34 more in the second quarter of 2025, and the release pace hasn’t slowed. A meaningful share of that growth is smaller, domain-tuned, and regionally optimized models built to do one job well.

That matters directly for a model-selection decision happening this quarter. Model choice should follow the use case. A smaller or domain-specific model, sometimes a regional one, frequently wins on cost and fit for a given task, where a general-purpose frontier model pays for breadth the task never uses. The best model for the job is rarely the biggest one.

Who owns the growing model roster

That accountability question is getting harder to answer, because the roster it’s about keeps growing. Enterprises are entering what IDC calls a multimodel, multimodal, multiagent era, one that needs evaluation, architecture, and orchestration competencies that didn’t need to exist two years ago. A dedicated GenAI evaluation category has caught up fast enough that IDC’s 2025 MarketScape already assessed 13 vendors in it, a category that barely existed the year before.

That governance work doesn’t end once a model is approved. Approval isn’t the finish line. The evaluation category emerged specifically because the portfolio keeps changing after signature, and testing has to run alongside it, not just at procurement. Your engineering team can track any one model closely. The portfolio as a whole is the harder watch, and right now it’s nobody’s job specifically.

The organizational response to that growth is a named owner. IDC is tracking a new role, the Chief AI Officer, with a formal strategic mandate most org charts didn’t carry two years ago. The role exists specifically to own decisions that used to default to whichever engineering team shipped first. Enterprises without that named owner are accumulating sprawl and data debt, plus the integration risk that comes with both, none of it budgeted for.

Expect this to formalize rather than stay ad hoc: model-selection governance tied to use case, cost, and compliance, continued funding for evaluation tooling as its own budget line, and the Chief AI Officer, or an equivalent, picking up ecosystem-level decisions no single engineering team can make alone. None of that requires waiting for the title to exist on your org chart first.

Three questions worth asking this quarter

None of this sits apart from the governance questions already landing on the CFO’s and CISO’s desks. Choosing a model is no longer a one-time technical decision made at kickoff. It’s now tied to what it costs and who’s accountable for it, the same fight already playing out over AI budget ownership and agent audit trails elsewhere in the organization.

Whether or not the Chief AI Officer title exists yet in your org chart, three questions are worth asking now:

  • Who owns the model portfolio today, by name?
  • What’s the plan when a model your team depends on gets deprecated mid-cycle?
  • Is your evaluation process a one-time procurement gate, or an ongoing discipline?

You don’t choose a model anymore. You govern a constellation. That starts with naming an owner this quarter, before the next model gets added to the roster without one.

Ryan Smith - Content Marketing Director - IDC

Ryan Smith is the Director of Content Marketing at IDC, where he leads brand-level content and social media strategy, aligning research insights with compelling storytelling to engage technology decision-makers. With a background in both IT and marketing, Ryan brings a unique blend of technical understanding and creative strategy to his work. He’s also a seasoned storyteller, speaker, and podcast host who believes the right message, told the right way, can drive both trust and transformation.

Christina Cardoza - Content Marketing Manager - IDC

Christina Cardoza is a Content Marketing Manager at IDC, where she specializes in brand content and social media strategy. With a background in journalism and editorial leadership, she has a proven ability to transform complex technology topics into clear, actionable insights.

2026 年,随着生成式 AI 与企业核心系统的持续化融合,中国企业资源管理(Enterprise Resource Management,ERM)领域的 AI Agent 正从通用问答和单点辅助,走向核心业务流程的深度嵌入。市场竞争也从技术功能展示,转向可落地、可复用的完整业务闭环。对于财务、采购、人力等关键场景,企业关注的不再只是模型能否给出答案,而是智能体能否理解业务、调用服务、完成任务,并在权限与规则约束下持续创造经营价值。

近日,国际数据公司(IDC)发布《IDC MarketScape:中国企业资源管理市场 AI Agent 2026 年厂商评估》(Doc# CHC54441226,2026 年 9 月)研究报告。报告基于厂商调研与访谈、公开信息以及最终用户体验,从能力与战略两个维度展开综合评估,帮助技术提供者明确产品升级与交付重点,也为企业用户更新 AI Agent 选型标准、推进业务场景落地提供参考。 如需了解更多评估结果,可参考所发布的报告。

基于本次评估,IDC 梳理出中国ERM 市场AI Agent 发展的五大核心趋势。需要强调的是,当前市场仍处于从标杆共创向规模复制演进的阶段,产品发布、场景覆盖与真实业务闭环之间依然存在距离。企业需要通过实际任务链,而非单纯的功能演示,判断产品成熟度与适配程度。

趋势一:从 Copilot 辅助到 Agent 自主执行:业务闭环是能力分水岭

早期 ERM 智能化应用主要集中在智能问答、数据查询、文档撰写和单据审核等场景,以单点功能优化为主。如今,部分厂商的 AI Agent 已能够理解并拆解企业任务,调用内部数据和业务服务,推进多步骤任务执行。自动化看板搭建、动态预算调整、采购方案生成、招聘流程协同等场景,正在推动产品从“提供建议”走向“完成工作”。

判断产品是否具备真正的 Agent 能力,不能只看检索结果或分析建议,还应验证从自然语言输入到服务调用、数据回写和系统结果的完整任务链,以及数据缺失、权限不足、接口失败和低置信度情况下的处理机制。不过,自主执行不意味着所有流程都应无人化。高风险财务核算、资金调拨、敏感人事管理和复杂经营决策仍需人工复核或固定程序兜底。IDC 预计,未来两年,“低风险流程自动闭环、高风险决策人机协同”将成为主流治理模式。

趋势二:统一入口与多智能体协同演进 —— 由分散工具迈向跨系统业务执行

统一 AI 入口正在成为下一代 ERM 系统的重要竞争方向。分散的 SaaS 功能模块、专属智能体、技能应用市场与第三方工具,可以通过统一入口整合和调度,推动交互模式从“模块菜单”主导转向“意图+上下文+可执行服务”主导。企业级平台也需要支持任务拆解、工具编排、多 Agent 协同、数字员工独立身份,以及定时、事件和流程触发等能力。

统一入口的价值并不止于增加一个对话窗口,真正的基础是业务语义、指标口径、组织权限和流程能力的贯通。厂商需要将 ERM 能力标准化、服务化、可调用化,使 Agent 执行任务时能够继承既有权限与审计体系。原生业务底座的优势目前更多体现在存量客户和自有生态,跨第三方系统的协同深度仍需验证。企业在选型测试中应至少接入一个非厂商自有系统,核查任务拆分是否可解释、智能体权责是否清晰,以及任务失败补偿、跨 Agent 权限与成本管控是否有效。

趋势三:可信 AI 与幻觉管控成为刚需:以数据治理和确定性执行守住业务底线

财务、采购、人力等低容错率场景,决定了 AI Agent 不能仅以单轮问答准确率衡量。报告显示,相关厂商正在采用业务程序与大模型配合的执行方式:确定性计算和关键校验交由传统程序或规则引擎完成,大模型主要承担意图识别、任务规划、流程编排和结果解读。配合原始单据溯源、置信度阈值、规则抽检与回归测试,可以降低模型幻觉与随机性对核心业务的影响。对异常和高风险动作,应通过人工确认、阻断或回滚机制控制风险。

数据治理则是上述能力落地的前提。主数据编码、指标口径、历史字段、非结构化文档和老旧系统数据,需要在项目初期统筹规划,同时厘清数据域边界、权限继承和调用痕迹。模型调优无法替代业务流程治理,数据基础薄弱也不能靠增加模型能力弥补。企业应将可解释、可追溯、可回放、可阻断、可回滚作为准入门槛,并建立随版本更新持续回归的评测集,将可信治理落实为可验证的运行机制。

趋势四:私有化部署与信创适配加速渗透:部署能力和服务边界共同影响选型

面向数据不出域诉求较强的央国企和大型工业企业,本地、专属云、混合部署以及国产软硬件适配能力,正在成为重要的选型因素。报告中的厂商在部署能力上存在明显差异:部分已支持全本地私有化和多模型切换,部分仍以公有云为主,纯私有化能力尚未落地。对于国际厂商,全球产品已发布也不等于中国客户可以直接采购使用,本地合规进程与交付节奏同样需要核查。

企业不宜仅凭“支持私有化”的产品标签作出判断,还应验证目标部署环境中的功能可用性、模型接入与切换能力、既有系统集成条件,以及后续升级和运营安排。私有化交付也不意味着服务边界天然完整,部分厂商能够覆盖模型调优,部分则需要客户团队自行承担。明确实施、集成、数据治理、模型调优与运维的责任分工,才能评估方案的实际落地难度与长期使用成本。

趋势五:行业化场景深耕形成差异化:从项目共创走向可复制产品与持续价值

制造、钢铁、电力以及央国企财务共享、司库等领域,正在成为厂商发挥既有业务积累的重要切入点。行业规则、真实业务数据与长期实施经验,有助于智能体嵌入专业流程,但单一行业的成功不能直接等同于跨行业复制能力。当前不少产品仍处于试点或标杆共创阶段,实际成熟应用也更多集中在问数、审核、报告和月结等场景,场景数量并不能替代对使用深度与业务效果的验证。

ERM Agent 市场仍将长期保持“产品+服务”的形态。厂商需要通过共创、专项交付或前线部署工程(FDE)服务,联动业务、产品和研发团队,将定制场景沉淀为可复用的行业能力,同时合理控制定制范围。商业模式也从传统软件许可,转向席位订阅、算力消耗与项目服务相结合。企业应建立覆盖算力调用、并发峰值、故障重试和工具接入的总体拥有成本(TCO)体系,明确用量规则、计费标准与服务边界。能否把项目经验转化为可迭代产品,并持续兑现经营价值,将直接影响厂商的交付效率与长期竞争力。

IDC 给技术提供者的建议

建议一:以业务结果而非模型参数规划产品。围绕客户的目标场景、业务现状、量化指标和风险边界设计 Agent,明确助手、顾问与自主执行产品的能力差异。优先打通规则明确、风险可控的完整任务链,对高风险决策保留人工复核与程序兜底,并通过异常处理演示证明产品的执行可靠性。

建议二:将统一入口建立在原生业务能力之上。统一业务语义、指标口径、组织权限和审计机制,完善任务拆解、工具编排、多 Agent 协同与独立身份管理。把第三方系统接入后的语义一致性、权限传递和失败补偿纳入产品验证,避免统一交互界面与底层流程能力脱节。

建议三:把数据治理和可信执行前置到项目初期。协助客户统一主数据、厘清指标负责人和数据域边界,保障关键结果可追溯至原始单据。采用确定性程序与大模型协同执行,将置信度分流、人工审核、操作留痕、阻断回滚和回归测试纳入产品与交付流程,降低核心业务运行风险。

建议四:按客户约束完善部署方案和服务体系。明确公有云、专属云、混合部署与纯私有化的实际支持范围,针对重点客户需求推进国产软硬件适配和模型接入能力建设。对实施、系统集成、数据治理、模型调优及后续运维作出清晰约定,让客户能够判断上线条件、服务责任与持续运营投入。

建议五:以行业资产复用和长效运营支撑规模化。通过共创和专项交付沉淀可复用的业务规则、流程与行业能力,合理控制定制范围,将交付经验反馈至产品迭代。建立清晰的订阅、用量与服务计费体系,帮助客户测算完整 TCO、持续追踪量化业务成效,以可复制的经营价值而非智能体数量构建差异化竞争力。

IDC中国企业级应用软件市场高级研究经理徐文婷认为,2026 年 AI Agent 在 ERM 市场加速落地,产品形态与竞争格局呈现五大核心趋势,即从 Copilot 辅助向 Agent 自主执行升级,多智能体协同与跨系统打通成为标配,可信 AI 与幻觉管控成为企业采购刚需,私有化部署与信创适配加速渗透,行业化场景深耕成为差异化竞争关键。

进一步交流

以上为《IDC MarketScape:中国企业资源管理市场 AI Agent 2026年厂商评估》报告的核心趋势解读。如需了解完整评估结果、厂商排名详情、细分场景能力对比及选型建议,欢迎与IDC团队联系。请点击此处与我们联系。

Wenting Xu

Wenting Xu - Senior Research Manager

Wenting Xu, senior research manager with IDC China's Enterprise Application Research, focuses on research and analysis of the China enterprise system and software area. She provides intelligence and consulting services for local and multinational corporation IT vendors. Wenting has developed a deep…

For the past few years, Meta Connect has been at the center of ongoing debates: is the Metaverse dead? Are AI glasses the next big thing? Will Zuckerberg’s latest bet pay off? This year answered a bit of all three. But the biggest news wasn’t one flashy new product. It was how many products Meta is now selling, at how many prices, in how many countries.

The VR glasses gamble

The headline reveal was Meta VR Glasses, Meta’s first real pair of VR glasses instead of a bulky headset. They weigh about 100 grams, five times lighter than a Quest 3, because the compute is separated from the headset into a puck that’s carried in the pocket rather than sitting on your face. They have a 5K screen, full color see-through video, support for Dolby Vision and Atmos, and about three hours of battery life. They launch in spring 2027 and cost $1,299.

That price says it all. This thing is priced like a PC, and it’s really part PC, part game console, part movie theater, which is exactly how Meta is pitching it. But the real point isn’t the spec sheet. Its standalone nature and high-fidelity experience give Meta its own platform, one that doesn’t need Apple, Google, or Microsoft. This is a product for consumers who are already sold on the category or enterprises looking to pilot VR, not the mass market that Meta is chasing with everything else it announced this week.

The pricing move matters as much as the platform argument. Meta has spent three headset generations racing to the bottom on price, so a device costing roughly four times as much as a base Quest is a deliberate step up-market, not a discount play. The data shows why the timing works in Meta’s favor: shipments of this passthrough-VR category, the group Quest and Vision Pro both sit in, peaked in 2024, then fell 44% in 2025 and a further 27% this year.

That timing matters because Meta’s Quest-based headsets have taken the majority of this category’s unit sales every year since 2024, against under 7% for Apple’s visionOS, even though Apple’s average selling price runs above $3,000 against something that’s well below $1,000 for Android-based devices. Meta doesn’t need this device to win share it doesn’t have. It needs a flagship on shelves as the turn towards recovery arrives next year.

Filling out the price ladder

That mass-market push showed up everywhere else at the event as Meta launched multiple pairs of glasses and aims to have 100 designs available to consumers by the end of the year. First up is the Ray-Ban Meta Gen 3, starting at $449, with a 12MP camera, 3K video, nine hours of battery, and additional microphones compared to the previous gen that cut background noise by 90%. Below that is a new option: Ray-Ban Meta Audio, starting at $349. No camera, just an AI assistant, music, and calls, with 12 hours of battery. This is the pair of glasses that has a broader appeal due to its price point and sidesteps privacy concerns faced by Meta’s other glasses that are equipped with cameras. The reduction in price and the absence of the camera will likely help Meta find new users who are currently on the fence when it comes to buying smart glasses. However, the absence of the camera is also potentially cutting the product off at its knees, as many consumers buy smart glasses for the video/photo capabilities, and longer term, it also means missing out on multimodal AI capabilities that are enabled by an onboard camera.

These two products are sold in conjunction with Meta’s own glasses, no Ray-Ban name attached, which start at $249. Beyond these glasses, Meta also sells a special edition with Kylie Jenner and announced a new version with Blackpink’s Lisa. For the uncultured (like myself), these are big-name celebrity endorsements which help the glasses maker capture a fashion-forward and potentially younger audience. Meta now sells glasses from $249 all the way to $1,299. That’s a big change from having just one flagship product, unlike other brands. This puts Meta in a very strong position against the competition, letting it compare feature sets and prices across a wide audience.

Display glasses finally go global

The other big piece of news was regarding availability of the Ray-Ban Meta Display, the $799 pair with a built-in screen. These glasses are finally launching outside the US, in the UK, Canada, and parts of Europe this month. I posted about this on X when it broke: based on IDC’s findings, demand in the US has been so strong that Meta has struggled to keep enough in stock over the past year. Getting these to other countries matters for more than just people like me in Canada who want to try a pair. It shows Meta has finally fixed its supply chain enough to sell beyond its home market.

Muse expands its reach

Meta also gave more stage time to Muse, its AI assistant. Muse is getting a smarter model and new connections to other apps, and it’s moving onto the Ray-Ban Meta glasses themselves so it’s available all day, not just when you open an app. Meta also showed off Muse Charm, a small pendant gadget with a screen, speakers, mics, and a fingerprint sensor, meant to carry a “real-time voice and avatar” version of Muse with you. There’s not much detail yet, and more is coming later this year, but it’s clearly meant to tie the glasses lineup together with Meta’s bigger AI plans.

Putting Muse directly onto the glasses, rather than leaving it as something you open on a phone, is the more consequential move of the two AI announcements this week. A phone-based assistant only helps once you’ve stopped what you’re doing and pulled the phone out. An assistant built into glasses you’re already wearing can listen, watch, and help without any of that friction, translating a conversation or setting a reminder without a single tap. That shift, from an app you visit to an assistant that’s simply there, is what turns smart glasses from a nice camera into something people rely on daily, and gives them a reason to stay with one brand rather than treat glasses as a one-off gadget. Still, one of the challenges Muse faces is being confined to its glasses and keychain charm; Apple and Google have enabled their own AI assistants across multiple devices, including glasses, headsets, smartphones, PCs, and tablets. In those cases, Muse will likely be available as an application instead of part of the AI layer built into the operating system.

That’s the business worth watching: it’s an order of magnitude larger in near-term units than the VR headset category, and Muse is the feature meant to make people stay in it.

What the IDC data says

Meta owned 68.7% of the XR glasses/headset market in 2026Q2. Nobody else is close. The bigger story is in the mix underneath that number. Display glasses like the Ray-Ban Meta Display grew to 14.3% of shipments, up from 11%. Headsets captured 15.4% and thin/light glasses with (or without) cameras represented the largest share at 70.3%. The market is clearly moving toward light, cheap, everyday glasses, which is exactly what Meta spent this event building out. Pricier headsets such as Meta’s VR glasses represent a small but dedicated audience despite the technological marvel that they are.

The bottom line is that Meta’s dominant position and Zuckerberg’s willingness to continuously invest in this category is likely to keep them in the lead despite upcoming products from competitors. By filling out the price ladder, Meta is ensuring that consumers of all sorts, and more importantly developers, continue to create or port their best XR experiences on Meta’s devices. And Meta’s Muse AI is trying to do something similar by offering extremely low-cost (free to start) access to end users for a highly capable AI agent.

Call it a real opportunity, but a defensive one. Meta isn’t trying to take a headset market it doesn’t already have; it’s defending one it already dominates, before lighter AR glasses from Snap, Google, XREAL, and eventually Apple make the whole idea of a headset feel dated. The smart glasses side of the business, priced from $249 to $449 and growing on units far faster than on value, is where Meta is actually playing to win.

Jitesh Ubrani

Jitesh Ubrani - Director, Consumer Devices Research

Jitesh Ubrani is a Director at IDC leading a team of analysts within the Worldwide Consumer Device Trackers group, covering wearables, augmented reality (AR), virtual reality (VR), tablets, phones, PCs, gaming, and smart home devices, with a focus on market…

Imagine being responsible for making sure market intelligence reaches the right decision-maker before a call gets made without it, at a global technology company that manufactures, delivers services, and funds serious R&D all at once. Now imagine doing it with a team of two. For years, that’s held up. It just doesn’t scale.

That’s the team the company’s Head of Information Management runs. Her group doesn’t produce research. It makes sure the research that already exists, inside the company and from partners like IDC, gets used. “We try and make information available and findable as easily and quickly as possible,” she said, “so that if somebody’s in a meeting, they can get to it quickly, or they can come to us and we can get to it quickly.”

Two people can stay close to maybe the top few hundred decision-makers in a company of that size. But everyone else is on their own. That is exactly the gap the company is trying to close with AI and tools like IDC Quanta.

A decade of using IDC to pressure-test decisions, not just source data

Long before AI entered the picture, one of the company’s regional strategy teams built a standing cadence with IDC analysts: a call every six weeks, with her team submitting the questions two days ahead so IDC could weigh in before the conversation even started.

That extra step mattered more than it sounds. “A lot of times what would come back is, ‘No, that’s not the questions you should be asking,'” she said. Her team would come in wanting data on one topic; the analysts would redirect them to a different one entirely. On numbers the company couldn’t break out itself, IDC analysts walked through the reasoning behind a breakdown, stress-tested the logic with the team, and told them plainly when their assumptions didn’t hold up.

It’s a pattern that shows up across the company’s engagement with IDC. The relationship delivers a second opinion sharp enough to change the question being asked, well beyond data on demand.

The real cost of AI answers nobody can verify

As general AI tools spread across the organization, a new problem showed up alongside the old one: getting an answer got faster, but trusting it got slower. When a leader asks for information on a topic, her team has watched colleagues turn to whatever open AI tool is on hand, then come back with a case to make.

“I have to waste time going in and looking at who the analyst is, where they work, who they work for, is it a credible firm,” she said. “I can’t just say no, it’s not credible. I have to explain why.” In one recent instance, the firm behind numbers a colleague had pulled couldn’t even spell the company’s own name correctly.

For her, that verification tax, the time spent fact-checking an AI’s sources instead of using its answer, is the exact problem IDC Quanta removes. Because IDC Quanta’s answers are drawn from IDC’s own verified research and cite the report behind every response, her team no longer has to relitigate credibility before anyone can act on what it says.

“

That’s where firms like IDC will really excel, being able to leverage the credibility you already have, so people don’t have to question whether they can trust the number.

— Head of Information Management, Anonymous

The other detail that stood out to her: what IDC Quanta does when the answer isn’t fully available. Where some AI research tools simply decline to answer a question outside a client’s subscription, IDC Quanta tells her what it can share and names the specific reports that would close the gap. “That to me is so much more helpful,” she said. “If we’re working on a key strategic topic, well then let’s buy those two reports, and use IDC Quanta to keep pulling the data out from there.”

A partner, not just a provider

The company sees AI extending its analyst relationships further than two people could ever reach on their own. When the company rolled out AI internally, the number of questions landing on her personally dropped by half, because AI could get people most of the way to an answer without waiting on her team. She expects IDC Quanta to do the same for her own workflow: fewer questions she has to chase down herself, and more time for the harder problems that still need her judgment.

What she’s watching for next is smaller and more practical: a tool that can recognize an outdated chart she drops in and pull the updated version automatically, so she spends less time hunting for the right version and more time using what she already knows is true.

For her, that’s the real value of the relationship. “I think that’s what you need to be,” she said. “A partner, and not just a provider.”

Christina Cardoza - Content Marketing Manager - IDC

Christina Cardoza is a Content Marketing Manager at IDC, where she specializes in brand content and social media strategy. With a background in journalism and editorial leadership, she has a proven ability to transform complex technology topics into clear, actionable insights.

Qualcomm flew members of the IDC team to Maui for this year’s Snapdragon Summit, and the theme of the week has been agentic AI. CEO Cristiano Amon used the day-one keynote to argue that the smartphone is moving from an app-centric model, where the user starts every task, to an agent-centric one, where software acts on the user’s behalf and is instant, dependable, and frictionless. “In the agentic age, the experience begins with intent,” he said. On day two, Qualcomm took that argument well past the phone, to PCs, glasses, watches, earbuds, and emerging form factors such as pendants.

Day one: The smartphone remains central

Amon was clear that the phone isn’t going away. His thesis is that it becomes more important, because it’s the device that knows you best and is always with you. He also set a practical bar for adoption: people will hand tasks to an agent only if they trust it and if it’s easier than doing the task themselves.

Google’s Rick Osterloh joined Amon on stage with numbers that show how much of this currently runs in the cloud. Gemini has reached 1 billion monthly users, and token usage is up 7x year over year. So what’s the value of running AI locally?

Qualcomm’s answer is the personal layer. Amon said an on-device agent needs personal knowledge, secure permissions, the ability to act on your behalf, persistent context, real-time reasoning, and multimodal input. Most of that depends on what the phone knows about you.

Qualcomm also teased a co-processor called High Bandwidth Compute (HBC) for Snapdragon that moves memory and AI compute closer together, headed for smartphones, laptops, glasses, and cars. In its datacenter accelerator cards and other datacenter implementations, the term HBC is a play on High Bandwidth Memory (HBM) for servers and refers to a near-memory implementation for processing AI workloads. Integrating this into its device SoCs could make Qualcomm more competitive on on-device AI workloads, with faster performance and lower power consumption. More details are expected at Mobile World Congress in 2027.

Two new flagship smartphone chips

For the first time, Qualcomm split its top mobile platform in two: the Snapdragon 8 Elite Gen 6 and the 8 Elite Extreme Gen 6. Qualcomm’s thesis is that one part can’t serve a flagship tier that now spans slabs, Pro Max models, foldables, camera phones, and gaming devices. With memory pricing disrupting the status quo, offering a second option gives OEMs more room to manage cost while prioritizing higher efficiency or higher performance depending on the use case.

The Gen 6 is the mainstream flagship part, on a 2nm process with the Oryon CPU and LPDDR5. The Extreme is the more interesting of the pair. Qualcomm says its Oryon CPU is the first in mobile to break 5 GHz. It moves to LPDDR6, adds matrix cores inside the Adreno GPU, and puts a heat slug directly on the die, which Qualcomm says doubles how long the chip can hold peak performance.

Qualcomm says the Extreme’s Hexagon NPU can run 30-billion-parameter mixture-of-experts models directly from flash. Its redesigned Sensing Hub powers Personal Scribe, which builds a personal knowledge graph on the device. That’s the persistent context Amon said agents need.

Xiaomi’s 18 Pro uses the Gen 6, and the 18 Pro Max uses the Extreme, with Motorola, OnePlus, OPPO, vivo, and others in line. Qualcomm says the price gap between the tiers will likely land around $100.

Day two: Agents across more devices

Kedar Kondap, who runs Qualcomm’s compute business, boiled it all down to one question: Where should inference run? He said the cloud brings the largest models and the deepest reasoning, while devices bring low latency, privacy, and personal context. But at the scale Qualcomm is talking about, he argued, the economics of sending every token to the cloud simply don’t work. That’s a more defensible pitch than the AI PC claims of the last few years.

New PCs on Snapdragon X2

Kondap called the PC the anchor node, with the thermal headroom and memory to run larger models locally. More than 7,300 apps now run on Windows on Snapdragon, he said, across 170-plus designs.

Microsoft’s Brett Ostrum introduced the next Surface Pro 12-inch and Surface Laptop 13-inch, both on Snapdragon X2 Plus. Microsoft claims up to 95% faster on-device AI and more than 60% faster graphics gen-over-gen, plus 25% brighter displays. The Pro starts at $1,150 and adds an optional 5G model for business. The Laptop starts at $1,200 and is rated for up to 22.5 hours of video playback. Both ship October 13 in select markets, along with a new $80 haptic Surface Mouse.

Developers have wanted Linux on Snapdragon for years. Qualcomm said it’s now officially supported, starting with the X2 series. Debian arrives at the end of this year. Ubuntu certification from Canonical follows in the first half of next year, along with ASUS Zenbook A14 and A16 models running it. Kondap also pointed to HUMAIN’s Horizon Ultra, an X2 Elite PC unveiled at LEAP that runs a Linux-based agentic OS.

Googlebook: A new laptop category

Googlebook broke as its own announcement on Monday, and Google’s John Solomon filled in the details on stage. It’s a new laptop category that combines Chrome OS and Android, aimed at Android phone owners. Solomon pitched it as the end of the “toggle tax” between phone and laptop, with access to phone files, native desktop Android apps, full desktop Chrome, and Gemini features such as Magic Pointer.

Worth noting: Googlebook runs on the first-generation Snapdragon X Elite, not the new X2. Dell and HP are first, and pre-orders are open. You’ll find IDC’s full coverage on Googlebook here: New Player Has Entered: Googlebook Is a Modern, AI-Forward Take on Notebooks.

Glasses, watches, and pins

On wearables, Qualcomm said Snapdragon AR1 and AR1+ are the first personal AI platforms to support multimodal 1-bit models, built on work with PrismML that cuts memory needs enough to run larger models on glasses.

Google’s Shahram Izadi showed the Snapdragon-based Pixel Watch 5 and Android XR devices from Samsung and XREAL. Google is also working with Gentle Monster and Warby Parker on AR1 glasses due later this year.

Qualcomm’s Snapdragon START program packages an AR1 module, software, and white-label designs to get glasses to market faster. UK eyewear maker Inspecs showed camera-equipped enterprise safety glasses and said it went from kickoff to working device in four months. Awear showed an agentic OS built around eyewear. Qualcomm also previewed a 5G personal AI pin reference design on Snapdragon Wear Elite, and formed a Personal AI Health Alliance around uses such as fall-risk prediction and clinical notes.

Snapdragon Sound Elite Gen 2 introduced

For earbuds, Qualcomm introduced Snapdragon Sound Elite Gen 2, which it calls the first premium audio platform built for personal AI. Versus the S7 Gen 1 Sound platform, Qualcomm claims a 30% smaller footprint, twice the AI performance, and up to 40% lower power. It also adds integrated micro-power Wi-Fi, so earbuds can reach cloud services and agents without a phone.

That enables a Snapdragon Sound Apps and Agents ecosystem, starting with MicroEJ, Cinemo, and Roon. HP, Bose, and Cleer are developing products.

Software: Memory layers and developer tools

Liquid AI, whose models already run in Mercedes-Benz cars on Snapdragon silicon, announced Liquid Context, a memory layer that lets agents share context across devices. And Chris Lattner, who joined Qualcomm through its Modular acquisition, said Modular’s stack, including the open-source Mojo language, is coming to Snapdragon and Qualcomm’s Dragonfly datacenter chips. Mojo is intended to reduce friction for developers and challenge NVIDIA’s CUDA moat.

Across both days, Qualcomm made it clear that it wants to be the agent’s endpoint on every device a person carries, working alongside the cloud rather than against it. The open questions are which of these new form factors people will buy and which developers will prioritize. We’ll be watching.

Tom Mainelli

Tom Mainelli - Group Vice President, Device & Consumer Research

Tom Mainelli heads the Device & Consumer Research Group, overseeing a wide array of hardware and technology categories catering to both home and enterprise markets. His team's research spans PCs, tablets, smartphones, wearables, smart home devices, thin clients, displays, and…
Phil Solis

Phil Solis - Research Director, Computing Systems Platforms and Technologies, Enterprise Infrastructure

Phil Solis is Research Director within IDC's enterprise infrastructure global research domain, part of the computing systems, platforms, and technologies subdomain. He focuses on client computing and connectivity, with coverage spanning semiconductors for PCs, media tablets, and smartphones, as well…
Kiranjeet Kaur

Kiranjeet Kaur - Associate Research Director

Kiranjeet Kaur is an associate research director for IDC Asia/Pacific. She is involved in building IDC’s successful research tracker programs and producing core market data used across all of IDC's Asia/Pacific research reports and consulting projects. Based in Singapore, Kiranjeet…
Jitesh Ubrani

Jitesh Ubrani - Director, Consumer Devices Research

Jitesh Ubrani is a Director at IDC leading a team of analysts within the Worldwide Consumer Device Trackers group, covering wearables, augmented reality (AR), virtual reality (VR), tablets, phones, PCs, gaming, and smart home devices, with a focus on market…
Antonio Wang

Antonio Wang - Vice President, Client System Research, Imaging, Printing & Document Solution Research, and Research Operation Center, IDC China

Antonio Wang is currently the vice president for IDC China's Client System Research; Imaging, Printing & Document Solution Research; and the Research Operation Center. He leads the personal devices business and product team. With more than 18 years of experience…

A couple of months ago I was at Siemens Realize LIVE in Amsterdam, and over dinner at the media and analyst reception, a conversation with Tony Hemmelgarn, CEO of Siemens Digital Industries, took a turn I did not expect. From industrial AI to why a Delhi intersection, and a UK roundabout arrive at the same result through completely opposite means.

Somewhere in there the UK driving exam came up too (that thing is brutal!). Tony had lived in the UK for a while himself, so he knew exactly what that experience is like. Stay with me, it made more sense than it sounds.

I grew up in India and have lived in the UK for several years now, which means I have spent enough time driving in both places to appreciate just how different they really are. I also happened to be back home in Delhi this summer for the holidays.

If you have driven in Delhi, you know exactly what I mean. There is a kind of ‘beautiful chaos’ to it. Cars, bikes, buses, auto-rickshaws, pedestrians, delivery guys on scooters, all making their own decisions at the same time, somehow without a plan. There are traffic lights and lane markings, sure, but anyone who has actually spent time on those roads knows the rulebook is only half the story. You are reading the situation constantly. A gap opens, someone takes it. Someone else has already anticipated that. A scooter appears out of nowhere on your left. You adjust. And somehow, everyone just keeps moving.

Now put that next to driving in the UK. Almost comically different. Everything is explicit here. Lane discipline. Signalling. Right of way. Mirrors, always the mirrors. Anyone who has sat the UK driving test knows that knowing how to drive and proving you can drive by the book are two very different things. It took me more than one test (meh!) to pass the exam. Unlearning Delhi’s driving and then learning the UK’s is not a joke.

The UK runs on predictability and rules. Delhi has rules too, but what really keeps things moving is interpretation, anticipation, and a kind of collective improvisation. Being neck deep in all things ‘industrial AI’, I couldn’t stop seeing the parallel, because the real question underneath both is who’s in charge when the rulebook and the improvisation are happening at the same time. That’s basically the orchestration problem, just with more scooters!

What Happens When Rules Fail: The Real-World Orchestration Problem

My last blog was about why the context layer is becoming such a big battleground in industrial AI. The next part to this is what happens after a system actually has that context. Knowing what is going on is one thing. Deciding what to do about it, especially when nobody has seen the situation before, is a different problem entirely.

For decades, industrial automation has been really good at that second part, as long as we can write the rules down. If X happens, do Y. Temperature crosses a threshold, trigger the alert. Machine stops, stop the line. Inventory drops below a level, reorder. That is honestly why automation has worked so well for so long. We take messy physical processes and turn them into decisions that repeat the same way every time. Nothing wrong with that at all, when the world is predictable, rules are genuinely powerful.

But what happens when the world is not predictable?

A machine starts behaving differently because the material batch changed. A supplier is late. Energy prices spike overnight. A production order gets rewritten at the last minute. An operator does something slightly different because they know something the system was never told. Or, my personal favourite, three unrelated things go wrong at exactly the same moment, and now it is not one system’s problem to solve. It is a question of who is in charge, and who decides.

Usually, the problem is not that the system does not have enough data. It might be drowning in it. The problem is that nobody wrote this particular situation into the rulebook, because nobody thought to.

Scaling AI Adoption: Why Most Pilots Don’t Translate to Production

Scaling AI, automation, and digital transformation is now the single biggest business priority for manufacturers globally, ahead of cutting costs, ahead of innovation, ahead of supply chain resilience from IDC’s own Worldwide Manufacturing 2026 Survey. And yet only around one in five manufacturers have actually gotten AI to scaled production use across multiple workflows or sites. Most are still piloting, or stuck running it in a limited corner somewhere. That gap between what people say they want and what is actually running on the floor is, in my mind, that same problem playing out at scale.

Blending Rules and Improvisation: A New Model for Industrial AI

We tend to talk about AI in manufacturing through use cases. Predictive maintenance. Quality inspection. Scheduling. Forecasting. All fine, all useful, but underneath all of that sits a much simpler question: can a system make a good call when the situation was never anticipated in the first place? That, to me, is really the line between automation and autonomy. Automation runs a known response. Autonomy has to read a situation, weigh a few options, and decide what happens next.

This is where the traffic thing earns its place, I think. Delhi has not beaten the UK at this. The UK has not beaten Delhi either. Both systems genuinely work, they just carry complexity differently. The UK gives industrial AI something it cannot do without: rules, constraints, predictability. You do not want an agent treating a safety requirement as optional, or skipping a maintenance step because it spotted a quicker path. Delhi gives you the other half of the lesson, which is that the real world does not always follow the plan, and when it does not, something needs to be able to interpret and adapt in the moment.

Traditional automation asks what should happen when a condition occurs. AI lets us ask something harder: given everything happening right now, what should happen? The first question can be programmed in an afternoon. The second needs judgement, including knowing when to ask a human. It also needs an understanding that the locally smart move can quietly create a bottleneck or a maintenance headache somewhere else entirely.

The decision stops being about one machine or one KPI. It becomes about the whole system, which is probably why I think the next real fight after the context layer is orchestration, agents working across ERP, MES, EAM, supply chain and energy together, not just sitting inside one application doing their own thing.

So maybe the future of industrial AI needs a bit of both: the UK’s rules and Delhi’s improvisation, with AI doing the very unglamorous job of figuring out which one a moment is asking for, across every system involved, not just the one closest to the problem.

The Path Forward

Underneath all of this sits a broader challenge: how intelligence works across systems, processes, people, and increasingly other agents. That’s orchestration. And as industrial AI moves from individual use cases into day-to-day operations, it is becoming harder to ignore.

For manufacturers, I think this changes the discussion slightly. Most organisations already have plenty of AI ideas, pilots, and use cases. What feels less clear is what happens when those systems need to work together. How does an AI-driven decision in one part of the operation affect everything else that sits downstream? Who is coordinating across ERP, MES, EAM, supply chain, and energy systems? And when things do not go according to plan, how does the system decide whether to act on its own or pull a human into the loop?

Those feel like the more interesting questions because they are not really about AI models at all. They are about how intelligence, whether human or machine, is orchestrated across the operation. And that may be one reason why so many organisations still struggle to move from promising pilots to something that operates at scale.

Contact our experts to explore how industrial AI orchestration applies to your manufacturing strategy.

Gunjan Bassi

Gunjan Bassi - Research Manager

Gunjan Bassi has more than 14 years' experience working in the logistics and transportation sector. Before joining IDC, she worked with Transport Intelligence (Ti), a transportation and logistics research firm based in Bath, England, where she was responsible for vertical…

If part of your job is staying ahead of where the market is headed, the stack you built your 2026 plan around won’t survive to 2028. IDC predicts today’s enterprise stack won’t exist in its current form by 2030, and by 2028, 42% of the world’s largest companies will have already rationalized their entire agent and apps portfolio just to keep pace.

That’s not an IT problem. It’s a shift in where value gets created, how deals get won, and what your go-to-market story needs to say next, and it belongs at the leadership table, alongside your P&L and risk register.

That’s the premise behind IDC FutureScape 2027: The Great Rewrite, IDC’s annual technology predictions event for Asia/Pacific, taking place live on 13 November 2026, 8:00 AM–12:00 PM SGT, at the Conrad Singapore Orchard.

The theme: The Great Rewrite

Intelligence is leaving the cloud for the device. AI factories run large-scale inferencing, agents work the network edge, and hardware is being rebuilt to carry it. Robots have moved off the roadmap and onto the factory floor. Quantum computing is out of the lab and into five-year plans, with real use cases beyond cryptography. Connectivity is heading into orbit and under the ocean floor, redefining what “always-on” means. Work is being redivided between people and agents, creating job titles, like Agent Supervisor, that didn’t exist a year ago.

Even the money is being rewritten. Knowing matters less than acting fast, and that shift is forcing CFOs to build an entirely new financial discipline just to keep up with what AI actually costs.

Dozens of evidence-backed forecasts, across every layer of the stack and every industry, converge on one conclusion: AI isn’t being added on top of the technology industry. It’s rewriting it from underneath, from the compute layer to the org chart.

The morning is packed with evidence. Agenda highlights include:

  • Keynote: The Great Rewrite
  • AI’s Reckoning: The Race, the Bill and the Ground It Stands On
  • AI Infrastructure Race: The Compute, Power and Trust Stack Reshaping Asia Pacific
  • Industry Panel: From AI Deployment in Practice to Emerging Capabilities
  • Winning in the Reimagined Technology Leader Landscape
  • Live Q&A and closing remarks

The full agenda, with timings and speaker bios, is available to download from the event page.

As Sandra Ng, senior vice president, Research and Data Intelligence at IDC, puts it, “The technology stack isn’t just evolving, it’s being rewritten from the compute layer up. Every leader in this room has a stake in that rewrite, whether they’re setting strategy, building the roadmap, or telling the story to the market. FutureScape 2027 gives you the evidence to make your next move with confidence, instead of waiting to see how the rewrite plays out.”

Why business and technology leaders need to be in the room

Predictions are only useful if you can act on them. That’s what separates FutureScape from a typical trends report: IDC analysts turn these predictions into a diagnostic tool to help technology leaders develop their enterprise IT strategy and roadmap and navigate their decisions with confidence.  If you’re a technology provider, IDC FutureScape reports and analysts can help you identify where your brand needs to position to win in the agentic economy with third-party backed content.

Whichever seat you’re in, this belongs on your calendar for three reasons: early access to evidence-backed predictions for 2027–2031, ahead of your competition; the breakthrough trends defining 2027 and the evidence behind how they’ll reshape IT roadmaps, purchase decisions, and the stories you tell in-market; and the best network in the room, with 150+ industry peers, technology leaders, and IDC analysts all in one place.

This shows up differently depending on your seat:

  • C-suite: know which shifts change your capital allocation and risk exposure before your board asks.
  • Sales: get evidence that reframes client conversations around where the market is heading.
  • Marketing: get the narrative and proof points before your competitors publish theirs.
  • Product: see which categories, from Agent Supervisor tooling to sovereign AI infrastructure, are being created before they have established players.
  • MI/Strategy: build your next planning cycle on the same evidence the analysts used to produce it.

 “Every year, FutureScape asks the same question: what’s actually going to matter in the year ahead, and what’s just noise? This year the answer touches everything, from where AI infrastructure lives to who’s accountable for governing it. My hope is that leaders leave this room with a clearer view of where to place their bets for 2027,” said Dr. Chris Marshall, vice president, Financial Services Research, IDC.

IDC FutureScape has already earned the trust of business and IT leaders across the region:

“The AI buyer engagement playbook was a game-changer — exactly what service providers need to stay ahead.” – Analyst Relations, Software

“The insights at IDC FutureScape helped us understand not just where AI is going, but how to get there with confidence.” – VP Southeast Asia, Korea and Channels, Solutions Provider

“Appreciate IDC for the bold, global conversations around the future of tech — lots to take forward.” – Industry Analyst, Consulting

Your next move

The technology stack is being rewritten layer by layer, and the leaders who show up with evidence will navigate it with confidence. Seats across the region are limited.

Register now for IDC FutureScape 2027 and reserve your seat at the Conrad Singapore Orchard on 13 November 2026.

Vanessa Ong - Senior Marketing Specialist, Demand Generation - IDC Asia/Pacific

Vanessa Ong is Senior Marketing Specialist, Demand Generation, at IDC Asia/Pacific, where she develops and executes integrated marketing programs that generate qualified leads and support business growth across the APJ region. A seasoned marketing professional with strong expertise in social media marketing and content creation, she is known for turning creative ideas into high-impact campaigns. Vanessa has played a key role in flagship programs including the Future Enterprise Awards and FutureScape, and is recognised as a collaborative team player who brings energy and expertise to every initiative.

The worldwide enterprise application market is approaching $700 billion, and blended growth has held near 13% through the first half of 2026, the third consecutive year of double-digit expansion. Yet, none of that has stopped the industry from convincing itself that the ground beneath it is about to disappear.

Every few years, enterprise technology rediscovers its love of the apocalypse narrative. The cloud was going to destroy on-premises software. Platforms were going to exterminate point solutions. And now, apparently, AI agents are going to detonate the entire SaaS business model, leaving nothing behind but smoldering subscription contracts.

The SaaSpocalypse thesis has been circulating with real momentum this year, and its appeal is understandable. It has urgency. It has drama. It makes for an excellent conference keynote. The problem is that it is analytically wrong in almost every important way.

What the argument actually claims

The core SaaSpocalypse argument goes something like this: AI agents will soon be able to perform the tasks that SaaS applications perform today. If an AI agent can handle your expense reports, reconcile your accounts, manage your procurement workflow, and schedule your workforce, why do you need Workday, Coupa, SAP Concur, or ServiceNow? The SaaS layer becomes redundant. Vendors who built their businesses on automating discrete business processes get disintermediated by a general-purpose intelligence layer. Subscriptions collapse. The apocalypse arrives.

It is a neat narrative. It is also a fundamental misreading of what enterprise software actually is.

The data model is the product

Here is what the SaaSpocalypse narrative consistently misses.  Enterprise SaaS applications are not primarily task executors. They are systems of record, workflow orchestrators, and data governance frameworks. The value in a mature ERP or HCM deployment is not the UI through which a worker submits a purchase order. It is the underlying data model, the audit trail, the compliance configuration, the approval hierarchies, the integration fabric connecting dozens of upstream and downstream processes, and the years of organizational logic encoded in the system.

An AI agent needs clean, structured, governed data to act on. It needs a reliable process layer to execute against. Those things do not come from the AI agent itself. They come from the SaaS systems that the apocalypse is apparently supposed to replace. This is precisely why the most sophisticated AI deployments in enterprise software today are not replacing SaaS. They are running inside it.

SAP is embedding Joule across its entire portfolio. Workday has Illuminate. Salesforce has Agentforce woven into its CRM and service layers. Oracle is pushing AI throughout Fusion. These are not defensive maneuvers by vendors in denial. They are the logical architecture for how AI creates value in complex enterprise environments.  AI is augmenting the system of record, not circumventing it.

Data from IDC’s 2026 SaaS & Agent Path and CX Path studies show that 32.8% of companies say they will pay at least 10% more for AI agents embedded directly into their applications, and 18% will pay a premium of 30% or more. That is not simply goodwill toward AI in the abstract. That is a market rewarding vendors for doing the hard work of embedding intelligence into the system of record, which is precisely the incentive structure the SaaSpocalypse thesis assumes does not exist.

We have seen this movie before

In 1999, the conventional wisdom was that the Internet would destroy the enterprise software industry. Why would you pay for a licensed SAP installation when web-native applications could deliver the same functionality cheaper and faster? The web did transform enterprise software. But Oracle, SAP, and other leading software vendors did not disappear. Instead, the incumbents adapted, absorbed the new delivery model, and in many cases emerged stronger. The cloud era brought the same narrative. Salesforce’s “No Software” campaign was positioned by Marc Benioff as the end of not just its direct CRM competitors, but the entire software industry itself.  In the end, it did kill Siebel, which Oracle bought for $5.85 billion worth of scraps in 2005.  But the software industry lived on. The delivery model eventually changed, but the complexity, the sprawling product lines, the high switching costs, and the underlying need for integrated, process-driven enterprise systems (all of which defined the previous incumbents) did not. Some could also argue that many of today’s leading SaaS vendors increasingly bear a resemblance to the very incumbents they once disrupted, with significant total cost of ownership and complex enterprise implementations that can run six to twelve months.  Not to mention the acquisition-driven growth, sprawl, and switching costs that resemble the profile of prior legacy enterprise software vendors.

The Boundaries Are Dissolving. The Software Is Not.

IDC’s recently published report, The Agentic Evolution of Enterprise Applications Framework – September 2026 Update, details the expected phased progression of apps in 18 individual markets. Since last year’s release, the framework introduces a new phase called Cross-Application Agents, in which AI agents stop working inside a single application and instead dynamically assemble whatever combination of capability, workflow, and data a task requires, pulled from wherever it lives, in real time. Oracle’s Fusion Agentic Applications illustrate the point: they are not pre-built applications a user logs into, but agents created autonomously that draw simultaneously from ERP, financials, HCM, or any other relevant system. SAP’s Joule, extended by WalkMe across its application landscape, works the same way. A user states an outcome, and the system pulls from whatever underlying application estate is required, without the user ever needing to know which system was ultimately involved.

While the boundary between applications will increasingly dissolve over the next decade, applications themselves will not. A cross-application agent still has to pull from somewhere, such as the ERP ledger, the HCM record, the CRM pipeline, or the compliance configuration that took years to build. The system of record does not disappear just because the interface sitting on top of it becomes fluid. If anything, this phase raises the stakes on everything argued above. The vendor whose data model, APIs, and integration fabric are ready to be discovered and orchestrated by an agent gains ground, and the vendor who is not becomes invisible. Put simply, if an agent cannot find your capability, your capability does not exist. That is not an apocalypse for SaaS. That is a new, more demanding competitive bar for visibility, and vendors who fail to adapt will simply stop getting called.

The real disruption hides beneath the aggregate

The honest version of the AI-and-SaaS story is not apocalyptic. It is structural. AI is compressing the time-to-value for enterprise software implementations. It is reducing the manual labor embedded in processes that SaaS applications have long only partially automated. It is enabling smaller teams to manage more complex operations. It is shifting competitive advantage away from vendors who simply automate a process and toward vendors who can deliver measurable business outcomes through intelligent, adaptive workflows. 

Our pricing research further validates this trend. The share of enterprise application pricing tied to consumption or outcomes, just 13% today, is projected to reach 63% over the next ten years, as traditional licensing and straight per-seat subscription pricing give ground to models more closely tied to usage and results.

These shifts will create winners and losers, and this unevenness is already evident in recent market data. As I wrote in my last blog, blended growth across public application vendors during the first half of 2026 has so far landed at roughly 13%, a number that looks unremarkable until you split it apart. Collaboration software and finance automation are accelerating, with vendors like Atlassian and Monday.com posting revenue growth above 20%, while customer service platforms and HCM vendors sit closer to flat. The aggregate market number hides that divergence, which is exactly why the loudest apocalyptic claims are looking at the wrong number. 

The divergence runs deeper than revenue

If we look at the aggregate projection for apps reaching “Agent-Enhanced” and beyond across all 18 markets within IDC’s Agentic Evolution of Enterprise Applications Framework, the unevenness is impossible to miss. 

To show this clearly, the table below categorizes the trajectories of all 18 markets into four bands:

  • Agentic Leaders – that start ahead and stay ahead
  • Fast Accelerators – that start behind but catch up
  • Steady Movers – that climb at a more measured pace
  • Structural Laggards – that never fully close the gap
CategoryMarket202620302038
Agentic LeadersContact Center50%80%97%
HCM48%70%99%
Supply Chain40%75%100%
Procurement35%75%100%
Financial33%75%99%
Fast AcceleratorsCustomer Service25%80%100%
Advertising Technologies19%76%100%
Enterprise Collaboration8%63%94%
EAM/ALM6%53%100%
Engineering/R&D5%45%100%
Steady MoversTalent Acquisition27%57%90%
Customer Analytics17%45%95%
Marketing14%53%92%
Structural LaggardsERP22%39%80%
PSA22%42%84%
Sales Performance & Productivity15%45%87%
Employee Experience13%24%78%
Digital Commerce4%14%58%

Figures represent the share of each market projected to be Agent-Enhanced, Agent-Led, Agents as Apps, or Cross-Application Agents in the given year. Source: IDC’s Agentic Evolution of Enterprise Applications Framework, Sept 2026.

The dispersion here is the real story. Contact Center starts 2026 at 50% and Digital Commerce starts at 4%, a twelve-fold gap in the same year, on the same framework. By 2038, the gap is still more than forty points, from Digital Commerce’s 58%, up to a full 100% in several markets. That is not the shape of a single technology wave hitting every application market at once. It is a market that is fragmenting, with its underlying markets evolving on their own timelines.

ERP, a market at the center of the data-model argument above, falls in the Structural Laggards category and is only projected to sit at 80% Agent-Enhanced or beyond by 2038. Digital Commerce tops out at 58%. Even a decade from now, a meaningful share of enterprise application usage is projected to still be running on traditional or lightly AI-assisted interfaces.

The SaaSpocalypse makes for a great headline, but reality requires something a bit harder. You need to understand what enterprise software actually does, and why organizations need it to continue doing what they do best.  SaaS & Agent Path finds that AI-driven capability is now the single most important attribute in a vendor evaluation, at 34%, ahead of ease of integration at 29% and data security at 28%. Moreover, 24.7% of companies say they plan to switch vendors outright if the next release fails to ship adequate agentic features. That is real, usable leverage, and it belongs in the room at your next renewal negotiation. 

Rather than checking to see if the SaaSpocalypse sky is falling, my advice is to use that leverage on your next SaaS renewal, focusing on reviewing your contract to ensure your vendor’s pricing model is predictable, transparent, and constructed to deliver the best possible value for your investment.  It’s a far better use of your time.

Eric Newmark

Eric Newmark - Group Vice President & General Manager of IDC's SaaS, Enterprise Software, CX and Workplace Solutions Division

Eric Newmark is Group Vice President & General Manager of IDC’s SaaS, Enterprise Software, CX, and Workplace Solutions Division, which includes several teams of analysts covering SaaS, 18 enterprise application markets, software monetization, business platforms, marketplaces, and services firms focused…

Berlin once again became the European center of consumer technology from September 4 to 8, 2026. According to IFA 2026 organizers, the event brought together approximately 240,000 attendees from 155 countries, more than 2,000 exhibitors, and over 5,200 media representatives.

As IDC walked the show floor, one trend stood out above all others: artificial intelligence is rapidly becoming the defining feature of the modern device experience, shaping innovation across PCs, smartphones, wearables, smart home products, and connected consumer electronics.

AI development push closer to the endpoint

At AMD’s IFA 2026 keynote, Jack Huynh, Senior Vice President and General Manager of AMD’s Computing and Graphics Group, outlined a vision of “Personal AI,” where capable AI development and inference workloads can run directly on endpoint devices rather than relying exclusively on cloud infrastructure. Key announcements included support for Project Zenith, Microsoft’s ready-to-code Windows developer experience for Ryzen AI Halo systems, as well as next-generation high-memory AI platforms designed to run larger models locally. Together, these announcements show how AI PCs are evolving from productivity devices into serious AI development platforms, combining powerful silicon with a streamlined software stack.

Jack Huynh introduces the AMD Threadripper Halo Station during his keynote at IFA 2026, showcasing AMD’s next-generation AI workstation platform. Source: AMD YouTube, 2026.

AMD also emphasized the importance of ecosystem integration beyond the device itself. The new Threadripper Halo Station brings workstation-class AI compute into a local development environment, while the company’s collaboration with SUSE aims to help developers move AI applications from Ryzen AI Halo systems into enterprise-scale production deployments. New Ryzen AI Max-powered systems announced by Acer and Lenovo across notebooks, mini PCs, and desktop form factors backed up that strategy. These announcements point to a broader industry trend: AI innovation is shifting toward high-performance endpoint devices that combine compute, memory capacity, developer tooling, and ecosystem support to bring AI development and inference closer to the user.

A multi-platform strategy takes shape at IFA 2026

Acer presented a broad portfolio spanning consumer, gaming, and commercial segments, making clear that the vendors on the show floor were tailoring device designs to specific workloads rather than relying on a single platform strategy. The announcements included the AMD Ryzen AI-powered Aspire G 3D 16, the Intel-based Predator Atlas 7 gaming handheld, the Veriton RI110 AI Mini Workstation for professional AI and productivity workloads, and a compact system concept based on NVIDIA’s RTX Spark platform. Together, these devices reflected the expanding range of AI, gaming, content creation, and workstation use cases emerging across the PC market.

The Acer Predator Atlas series handheld gaming PC, powered by Intel’s Arc G3 processor platform and designed for high-performance portable PC gaming. Source: IDC, 2026.

Acer’s announcements pointed to a pragmatic approach to device design: instead of committing to a single processor roadmap, the company matched AMD, Intel, and NVIDIA platforms to distinct customer requirements and form factors. As AI PCs continue to evolve and new categories such as AI workstations and gaming handhelds gain traction, the vendors IDC saw at IFA are competing on their ability to address a wider range of usage scenarios. Acer’s IFA portfolio confirmed how this flexibility can help position the company across several of the fastest-growing segments of the client device market.

Broadening reach across consumer and gaming segments

At IFA 2026, Dell Technologies focused on two areas of the device market that continue to show strong demand: affordable premium consumer laptops and high-performance gaming peripherals. The new Dell 14S extends design elements often associated with premium PCs into a more accessible segment, combining a lightweight aluminum chassis, multiple color options, long battery life, and a portable form factor. According to Dell, the device is aimed at students and young adults who prioritize mobility, battery life, design, and value in equal measure. The Dell 14S doesn’t compete solely on specifications. It reflects how personal devices are becoming both productivity tools and expressions of personal preference.

Dell 14S consumer laptop in Velvet Green finish, displayed as part of the Dell 14S lineup shown at a product showcase. Source: IDC, 2026.

Dell also used IFA to expand the Alienware display portfolio with new OLED gaming monitors targeting different player segments. The Alienware 32 4K OLED focuses on visual fidelity across PC and console gaming, while the Alienware 25 560Hz QD-OLED targets competitive esports players and was developed in collaboration with Team Liquid. That partnership shows how gaming hardware vendors are pulling direct feedback from professional players into product development. As competitive gaming continues to mature, factors such as refresh rate, motion clarity, ergonomics, and setup consistency are becoming important differentiators alongside raw performance specifications.

Lenovo brought new products and new concepts

Lenovo brought two RTX Spark machines to its Innovation World event before IFA. The Yoga 9n 2-in-1 is likely the first Spark-based 360-degree convertible and includes dual-surface pen input designed for creators. It will ship in 11- and 16-inch sizes. The Pro 9n is the more conventional creator laptop, but also includes pen input, and will ship in 11 and 15-inch sizes. No pricing yet, but as with all RTX systems, we expect them to be premium.

The IdeaPad Vibe is Lenovo’s answer to Apple’s MacBook Neo. Seven Pantone colors, swappable keycaps, and a starting price of $699. Snapdragon X or Ryzen AI inside, Lenovo’s work-in-progress AI assistant Qira on board, Copilot+ across the line, shipping in October. The IdeaCentre AIO i desktop also comes in a variety of colors and will sell for $939 in November.

Lenovo IdeaPad Vibe series, unveiled at Lenovo Innovation World 2026. Source: Lenovo, 2026.

Lenovo also showed two Intel-powered concept notebooks. Project AeroBlade is a 14-inch fanless notebook under 10mm thick and 830g, using Frore’s AirJet solid-state cooling in place of a traditional fan. And Project Swan’s 14-inch display rolls out to 17 inches.

Over on the Motorola side, Lenovo announced two smartphones and a smart watch. The special-edition razr, done in a PANTONE Meteorite finish with 35 hand-set Swarovski crystals and a faceted crystal on the hinge. It includes a 3.6-inch cover screen, a 6.9-inch folding AMOLED interior screen, and dual 50MP cameras. No pricing yet.

Over on the Motorola side, Lenovo announced two smartphones and a smart watch. The special-edition razr, done in a PANTONE Meteorite finish with 35 hand-set Swarovski crystals and a faceted crystal on the hinge. It includes a 3.6-inch cover screen, a 6.9-inch folding AMOLED interior screen, and dual 50MP cameras. No pricing yet.

The moto watch ultra is Motorola’s first LTE smartwatch, runs Wear OS, leans on Polar for health tracking, and is the first wearable with Qira built in. Stainless steel, 46mm, and it offers up to two days of battery life. It began shipping in the US on September 10 at $350.

Local AI infrastructure takes shape in the NAS category

MINISFORUM focused on an emerging area of the AI market: local AI infrastructure. The company’s new AI Agent NAS N5 MAX-P495 and AI Mini Workstation MS-S1 MAX-P495, both powered by the AMD Ryzen AI MAX+ PRO 495, combine high memory capacity, local AI compute, and large-scale storage in compact form factors. The AI Agent NAS supports up to 192GB of memory and up to 200TB of local storage, positioning it as a platform for AI agents, retrieval-augmented generation (RAG), knowledge bases, and other data-intensive AI workloads that benefit from keeping both compute and data local.

The MINISFORUM N5 MAX AI NAS. Source: IDC, 2026

The AI Agent NAS stood out as an example of how local AI infrastructure is evolving beyond standalone AI PCs and workstations. By combining AI compute, large memory configurations, and substantial local storage in a single system, MINISFORUM is addressing two common constraints in enterprise AI deployments: memory capacity and data locality. As AI agents and RAG-based applications become more prevalent, bringing compute and data together within the same platform could become as important as increasing AI performance itself.

The Mini PC reimagined as a distributed AI platform

GEEKOM showcased a range of AI-focused devices, including the new GeekBook X16 Pro laptop, IT13 Mega mini PC, and A5 2027 Edition, but the most notable demonstration was a distributed AI cluster built from four A9 Mega mini PCs connected over USB4 and operating as a single system. Powered by AMD Ryzen AI Max+ 395 processors, the cluster showed how compact AI PCs can be combined to run large AI models locally without traditional server infrastructure. Alongside the cluster, GEEKOM positioned the GeekBook X16 Pro as a portable AI-capable laptop and introduced new mini PC options addressing both high-performance and mainstream productivity use cases.

GEEKOM’s setup skipped the dedicated workstation or datacenter-class system entirely. It showed how multiple compact AI PCs could function as a small AI cluster located next to a user’s desk. As organizations evaluate privacy, data sovereignty, and local AI deployment models, these distributed architectures may emerge as an alternative path for increasing AI capacity while retaining the flexibility and form factor advantages of mini PCs.

Ecosystem strategy expands across PCs and intelligent edge devices

Qualcomm pointed to both the expanding scale of its Snapdragon-powered PC ecosystem and its broader ambitions in edge computing. Across the show floor, systems from ASUS, Acer, HP, Lenovo, and Samsung showed how OEMs are applying Snapdragon platforms across multiple form factors and user segments. The expansion of the Snapdragon X Plus 8-core portfolio continues to extend Copilot+ PCs into broader market tiers, and the range of devices on display backed up Qualcomm’s ecosystem-driven approach, where design differentiation is becoming important alongside platform performance.

Qualcomm used its booth to showcase the breadth of the Snapdragon-powered Copilot+ PC ecosystem. Source: IDC, 2026.

Beyond PCs, Qualcomm introduced the Dragonwing Q-2390 and Dragonwing IQ-2390 processors, expanding its reach into consumer, commercial, and industrial intelligent edge devices. The new platforms integrate AI, vision, connectivity, and edge processing capabilities in a single design, targeting applications ranging from smart home products and retail systems to industrial automation and building management solutions.

Smart home robotics extends beyond the vacuum

Away from computing devices, one of the most dynamic categories at IFA 2026 was smart home robotics. IDC data shows that vendors shipped just over 30 million household cleaning robots worldwide in 2025, spanning robot vacuums, robotic pool cleaners, and robotic lawn mowers. That represented 17% growth year over year and a third consecutive year of double-digit expansion, and the pace has since accelerated, with IDC figures showing the market up more than 25% year over year in the first half of 2026. Robot vacuums still account for close to 80% of units, but the fastest movement is happening at the edges of the category. Robotic lawn mowers grew 64% in 2025 and a further 76% in the first half of 2026, while premium vacuums priced above $600, equipped with multi-sensor AI navigation, now make up roughly a third of all units shipped. The category is shifting from single-purpose indoor cleaning toward multi-scenario automation that spans both indoor and outdoor tasks, and IDC expects shipments to reach 54 million units by 2030.

Roborock exhibit at IFA 2026, showcasing the company’s robotic cleaning and lawn care product ecosystem under the theme “Rocking Life with You.” Source: IDC, 2026.

Roborock used the show to make that shift concrete. The company has been the number-one robot vacuum vendor globally by unit share every quarter since the second half of 2023, reaching an all-time high of 27% in the second half of 2025, and IDC ranks it first across the entire household cleaning robot category with close to 18% share in 2025. At IFA, Roborock extended well beyond the vacuum. Alongside new flagship models led by the Saros 20 Flow Complete and Saros 20 Neo, and the wheel-legged Saros Rover designed to climb stairs and move between floors, the company introduced its first robotic pool cleaner, the RockAqua P1, and a wire-free lawn mower, the RockNeo Q2, which uses LiDAR-based navigation to operate without boundary wires. The lineup signals a deliberate push into adjacent outdoor categories: Roborock is applying the navigation, mapping, and obstacle-avoidance technology it refined indoors to the two fastest-growing outdoor segments, where navigation performance is likely to be a key competitive differentiator among vendors.

ECOVACS was also on the show floor showing off new products including the DEEBOT X12S OmniCyclone which offered 97% removal of dust mite allergens and 95% removal of dust mites from carpet, claims that were verified by third-party SGS. The X12S also introduced a physical privacy shield that covers the onboard camera, giving users a hardware-level privacy control in addition to the software toggle. These are the type of features that ECOVACS typically introduces first on its X series vacuums before they trickle down to the more mainstream-priced T series.

ECOVACS DEEBOT lineup showcased at IFA 2026. Source: IDC, 2026.

Beyond the vacuum, ECOVACS also used IFA to reinforce its multi-category ambitions, emphasizing that cleaning will be one of many in-home categories where the company will offer robotic solutions. The company already holds the top rank in robotic window cleaners with its WINBOT line, according to IDC. The company also offers pool cleaning and lawn mowing robots with the ULTRAMARINE and GOAT products, respectively.

The IDC take

IFA 2026 showed how AI is shaping device design decisions across the industry. Rather than being treated as a standalone feature, AI is influencing hardware architectures, memory configurations, software environments, and deployment models across PCs, workstations, and edge systems. A recurring theme across the show was the emphasis on local AI execution, enabled by advances in compute performance, expanded memory capacity, and more capable endpoint devices.

From an IDC perspective, differentiation is gradually shifting beyond processor performance alone. The vendors IDC tracked at IFA are competing on their ability to deliver integrated AI platforms that combine compute, memory, storage, developer tools, and ecosystem partnerships. Whether through high-memory AI PCs, workstation-class systems, AI-enabled storage platforms, or clustered mini-PC architectures, the focus is on creating practical environments for deploying and running AI workloads closer to users and data.

Smart home robotics confirmed the same pattern from a different angle. In cleaning robots, differentiation is moving decisively toward AI and navigation, and the category leaders are now extending those capabilities from indoor floors into outdoor tasks such as lawn care and pool cleaning. IDC expects household cleaning robots to remain one of the fastest-growing consumer device categories through 2030. Roborock and ECOVACS are already running that playbook, extending their indoor navigation stacks into lawn care and pool cleaning ahead of most competitors.

Francisco Jeronimo

Francisco Jeronimo - VP, Data and Analytics, Devices, IDC EMEA

Francisco Jeronimo is VP for Data and Analytics at IDC EMEA. Based in London, he leads the research that covers mobile devices, personal computing devices, emerging technologies and the circular economy trends across EMEA. His team delivers data on personal…
Jitesh Ubrani

Jitesh Ubrani - Director, Consumer Devices Research

Jitesh Ubrani is a Director at IDC leading a team of analysts within the Worldwide Consumer Device Trackers group, covering wearables, augmented reality (AR), virtual reality (VR), tablets, phones, PCs, gaming, and smart home devices, with a focus on market…
Tom Mainelli

Tom Mainelli - Group Vice President, Device & Consumer Research

Tom Mainelli heads the Device & Consumer Research Group, overseeing a wide array of hardware and technology categories catering to both home and enterprise markets. His team's research spans PCs, tablets, smartphones, wearables, smart home devices, thin clients, displays, and…
Mohamed Hefny

Mohamed Hefny - Senior Program Manager, Data and Analytics

Mohamed Hefny leads market research in EMEA on professional workstation PCs and solutions. He also reports on professional computing semiconductors, processors, and accelerators (CPUs and GPUs), as well as breakthroughs and trends related to the market. In addition, Mohamed is…