Logo
Become a Client

Global disruption is not caused by an isolated event. It is continuous, multidimensional, and increasingly interconnected. Economic volatility, geopolitical fragmentation, regulatory expansion, workforce transformation, and rising customer expectations are converging to reshape how organizations operate and compete.

IDC’s FutureScape 2026 identifies this convergence of forces as Navigating the Crosscurrents of Disruption. This defines how organizations can respond to these overlapping pressures with deliberate strategy rather than reactive adjustment. At the core of that response is agentic AI, not as an isolated capability, but as a governed, strategy-aligned force that converts disruption into momentum.

Without deliberate navigation, leaders risk being dragged sideways by disruption, leaving them unprepared to capture the benefits of the agentic economy.

A framework for navigating compounding disruption

Crosscurrents are not independent variables that can be managed in sequence. They have a cascading effect.

A regulatory shift in one market affects supply chain strategy. Geopolitical instability reshapes technology sourcing decisions. Workforce disruption intersects with AI adoption. Leaders acting on one pressure are already absorbing consequences from several others.

This is the decision environment mapped by FutureScape. Not as a catalog of threats, but as an analytical lens for understanding how simultaneous forces compound and where deliberate action creates leverage.

Leaders must balance competing priorities while maintaining forward momentum. Success depends less on predicting disruption and more on navigating it effectively.

AI sits at the intersection of these crosscurrents. It is subject to regulatory and sovereignty constraints and is often limited by fragmented implementation. Yet it can also serve as a mechanism for coordination, enabling organizations to integrate data, align decisions, and respond more dynamically to economic and operational complexity at scale.

When governed and aligned with enterprise strategy, AI transforms the crosscurrents into momentum. Deployed without that alignment, it can amplify the complexity leaders are already facing.

The structural impact of disruption

The crosscurrents are no longer hypothetical pressures on the horizon. They are already reshaping infrastructure decisions, cost structures, and leadership accountability across technology suppliers and enterprise buyers.

Disconnected systems, duplicated investments, and uneven execution are introducing new layers of cost and operational burden that organizations must actively manage.

Architecture is being reshaped by sovereignty and regulation

This shift is most visible in AI architecture. Sovereignty laws and evolving regulatory requirements are forcing organizations to make deliberate decisions about where data resides, how models are deployed, and which systems can operate across jurisdictions.

Compliance fragmentation is becoming a defining constraint. Organizations must navigate inconsistent regulatory frameworks across regions, often requiring localized architectures, governance models, and data environments. This limits standardization and increases structural complexity.

IDC predicts that 60% of global firms will split their AI stacks across sovereign zones by 2028. Enterprise architecture is no longer designed for global uniformity. It must accommodate regulatory divergence while maintaining cohesion across the organization.

Implementation complexity is slowing outcomes

As AI adoption scales, the focus is shifting from experimentation to execution. Disconnected investments across business units, regions, and use cases are creating duplication, inefficiencies, and inconsistent results.

This increases cost and slows progress. Organizations must fund parallel initiatives, manage overlapping capabilities, and invest in integration to connect systems separated by region and function.

According to IDC research, 40% of organizations will miss their AI goals due to implementation complexity in 2026. The issue is rarely the technology itself. It is the gap between AI deployment and the enterprise structures, governance models, and integration required to make it work at scale.

Complexity becomes the primary constraint

As these forces converge, complexity becomes the defining constraint on progress.

For leaders, this translates into rising costs, slower execution, and reduced strategic flexibility. Organizations must coordinate across architectural boundaries and operational silos while managing regulatory demands and investment trade-offs.

Those that build alignment across architecture, investment, and execution will be better positioned to sustain momentum. Those that do not risk being constrained at every layer of the enterprise.

Maintaining direction in a shifting environment

Navigating the crosscurrents requires getting three foundational areas right. These areas are interdependent, and gaps in any one limit what is possible in the others.

To move forward, leaders must:

1. Define ownership and strategic direction
Crosscurrents do not respect organizational boundaries. Economic risk, AI governance, regulatory compliance, and technology investment are converging into the same course of decision making. Leaders must ensure the right stakeholders own these decisions and that CIO and CFO mandates are aligned around shared outcomes rather than managed in parallel.

2. Evolve workforce models for new operating realities
Workforce fragmentation is increasing across regions, regulatory environments, and technology stacks, creating inconsistencies in execution and decision-making. Navigating at the pace of change requires unifying this environment through human–AI collaboration as a core operational capability.

3. Strengthen foundations for integration and adaptability
Infrastructure decisions made today define future strategic options. Data foundations, integration architecture, and sovereignty-ready systems are not simply IT priorities. They are decisions about where the organization can operate and how quickly it can adapt.

Deliberate navigation across these three fronts separates organizations positioned to capture the agentic economy’s upside from those that lose direction in the crosscurrents.

The stakes of standing still

The crosscurrents shaping the agentic economy are not temporary conditions. Economic and geopolitical volatility, regulatory fragmentation, and workforce disruption are structural features of the environment in which leaders now operate.

Leaders who build strategic alignment, workforce capability, and infrastructure readiness will not only maintain direction within this changing environment but also convert external pressures into coordinated progress. That preparation cannot happen at the point of disruption. It must be in place before challenges arise.

Navigating the crosscurrents is where that work begins. Understanding the crosscurrents, mapping their interactions, and identifying where deliberate action creates the most leverage are the first steps to success in the agentic economy.

Explore navigating the crosscurrents of disruption in depth

Explore navigating the crosscurrents of disruption in depth

FutureScape 2026 includes detailed research, analyst perspectives, and webinars that expand on the themes within navigating the crosscurrents.

Core research

Analyst perspectives

On-demand webinars

eBooks

Explore the full FutureScape 2026 predictions.

IDC - -

LinkedIn
International Data Corporation (IDC) is the premier global market intelligence, data, and events provider for the information technology, telecommunications, and consumer technology markets. With more than 1,300 analysts worldwide, IDC offers global, regional, and local expertise on technology and industry opportunities and trends in over 110 countries. IDC’s analysis and insight help IT professionals, business executives, and the investment community make fact-based technology decisions and achieve their key business objectives.

IDCの最新レポートでは、今回の中東での戦争が2026年に向けてクラウドのレジリエンス、サイバーリスク、サプライチェーン、IT計画にどのような影響を及ぼすかを考察しています。
地政学的危機は、明確な予兆なく発生することがほとんどです。そして一度発生すれば、デジタルインフラ、サプライチェーン、テクノロジー運用に即座に大きな負荷がかかります。
今回の中東での戦争は、現代のデジタル経済、そして混乱下でも事業継続を担うCIOにとって、構造的なストレステストとなっています。
過去の紛争と異なり、現在の企業IT環境はクラウドインフラ、サブスクリプション型サービス、そしてグローバルに相互接続されたサプライチェーンに大きく依存しています。そのため、影響は局所的にとどまらず、地域・システム・パートナーを横断して急速に拡大します。
CIOにとっての課題は、単なるリスク管理ではありません。変化する状況に適応しながら事業運営を維持し続けることです。

IDC の見解：From Conflict to Continuity: How CIOs Can Respond to Disruption from the Middle East War.

なぜ今回の危機はCIOにとってこれまでと異なるのか

企業ITは、従来の内部完結型環境から、高度に分散されたエコシステムへと移行しています。

現在の企業は以下に依存しています：

  • クラウドプロバイダーおよびプラットフォームサービス
  • 分散型インフラおよび運用
  • テクノロジー供給および提供におけるグローバルサプライチェーン

この依存構造は、新たなリスクを生み出しています。

地域の不安定化は、以下に影響を与えます：

  • アプリケーションの可用性およびパフォーマンス
  • ハードウェア導入スケジュール
  • サイバー攻撃の増加
  • インフラおよびエネルギーコスト

これらの圧力は、既存のデジタル戦略の前提をすでに揺るがしています。

CIOにとっての優先事項は、自社のリスク露出（エクスポージャー）を把握し、それが業務にどう影響するかを理解することです。

エクスポージャーマッピングとシナリオプランニングから始める

最初のステップは、どこにリスクが集中しているかを特定することです。

CIOは以下の4つの観点で依存関係を整理すべきです：

  • 影響地域に所在する従業員・契約社員
  • 影響市場に関連する顧客および収益源
  • 混乱が発生している物流ルートやサプライヤー
  • 地域インフラに依存するアプリケーション、データ、運用

このマッピングが意思決定の基盤となります。

その上で、シナリオプランニングにより複数の展開に備えることが可能になります。

IDCは、CIOが検討すべき2つのシナリオを提示しています：

  • 地域の不安定状態が長期化するケース
  • エネルギーやサイバー領域に波及する広範なエスカレーション

シナリオごとに、レジリエンス、セキュリティ、投資の優先順位は変化します。

CIOが今優先すべきこと

CIOは以下の5つを直ちに見直す必要があります：

1. クラウドとインフラのレジリエンス再評価

単一リージョンや特定プロバイダーへの依存度を確認し、フェイルオーバー体制のギャップを特定する。

2. サイバーセキュリティの強化

脅威の増加を前提に、検知・対応・復旧能力を強化する。

3. テクノロジーサプライチェーンの多様化

供給のボトルネックを特定し、単一供給源への依存を低減する。

4. データ主権とコンプライアンスの見直し

地政学的緊張はデータローカライゼーションや規制強化を加速させる。

5. 人材・業務継続計画の整備

リモートワークや代替コミュニケーション手段を含め、業務継続を確保する。

これらは新しい課題ではありませんが、「同時に、かつ迅速に」対応する必要性が高まっています。

IDC アジアウェビナー（英語）：Asia Pacific IT Spending Outlook 2026: Where to Win Amid Market Volatility

混乱下でのリーダーシップ

レジリエンスは技術課題であると同時に、リーダーシップの課題でもあります。

CIOは不確実性の中で明確な方向性を示す必要があります。成功する組織は、チームが目的を理解し、迅速に行動できる組織です。

有効なリーダーシップ行動には以下が含まれます：

  • 重要システム保護への明確なフォーカス
  • 意思決定の迅速化
  • 課題の分解と優先順位付け
  • 環境の簡素化と強化の機会特定

こうした局面では、技術的負債や運用の非効率、レジリエンスの欠如が顕在化します。

優れた組織は、混乱を「停止」ではなく「行動の契機」として捉えます。

混乱からオペレーショナル・レディネスへ

現在の状況は、地政学とデジタル運用の関係が変化していることを示しています。

CIOはもはや単発のインシデントに備えるのではなく、複数領域に同時影響が及ぶ前提で対応を進めなくてはなりません。

そのためには、継続的なレジリエンス強化が必要です：

  • 依存関係とリスクの可視化の継続
  • 複数シナリオを前提とした計画
  • 日常業務へのレジリエンスの組み込み

この能力を構築できた組織は、不確実性の中でもパフォーマンスを維持できます。

シナリオフレームワークの活用

リスクの把握は出発点に過ぎません。重要なのは、それを意思決定に落とし込むことです。

IDCのレポートでは以下について詳述しています：

  • IT支出やAI投資への影響
  • インフラ、サイバーセキュリティ、人材継続性への考慮点
  • 状況変化を把握するためのリスク指標

執筆者（Authors）

Rick Villars – Group VP, Worldwide Research – IDC

原文：2026年3月23日公開（英語）｜日本語版監修：寄藤 幸治

出典：CIO readiness in the face of the Middle East war: From disruption to resilience

関連する調査やご相談について

より詳細なインサイトや市場動向については、当社アナリストへお気軽にご相談ください。

BEIJING, April 15, 2026 – China’s smartphone market declined 3.3% year over year in the first quarter of 2026, with shipments reaching approximately 69.0 million units. Despite a slight contraction, performance exceeded initial expectations due to strong demand for premium products from Huawei and Apple. Rising memory and component costs constrained supply and forced vendors to prioritize high-end models, signaling a from a volume-driven recovery to a margin-protection phase, making quality growth and operational efficiency the primary indicators for the year.

China smartphone market shipments Q1 2026 IDC data

“China’s smartphone market is entering a phase where profitability matters more than shipment growth,” said Will Wong, senior research manager, Devices Research, IDC Asia/Pacific. “Vendors are making deliberate trade-offs by reducing low-end exposure and focusing on premium segments to offset rising costs and protect margins,” Wong ends.

What happened in the China smartphone market in Q1 2026?

The market declined 3.3% year over year to 69.0 million units, but outperformed expectations due to strong premium demand led by Huawei and Apple. Growth was constrained by rising component costs and supply shortages, prompting vendors to reduce low-end exposure and prioritize profitability over shipment volume.

China Smartphone Market at a Glance — Q1 2026

  • Total shipments: 69.0 million units (–3.3% YoY)
  • Primary growth driver: Premium demand from Huawei and Apple
  • Key constraint: Rising memory and bill-of-materials costs
  • Supply challenge: Component shortages limiting full demand realization
  • Market shift: Transition from volume growth to margin protection

Why did the market change?

Premium demand remained resilient, particularly for Huawei’s Mate 80 series and foldable Pura X, as well as Apple’s iPhone 17 lineup. However, rising memory costs increased overall device production expenses, forcing vendors to reduce exposure to low-margin segments. Supply constraints further limited shipment potential, especially for Apple, where growth could have been higher without shortages.

IDC outlook

The first quarter is expected to be the strongest period of 2026. Vendors are revising annual targets downward and maintaining tight control over low-end inventory. Market performance for the rest of the year will depend on how effectively vendors balance innovation, cost management, and supply chain resilience amid sustained pricing pressure.

Vendor Highlights

Huawei maintained market leadership, supported by improved supply of its flagship and foldable devices, with the Pura X exceeding 1.5 million units in shipments. Apple recorded the fastest growth among the top five vendors, with shipments increasing 33.3% year over year, although overall volume was constrained by supply limitations.

FAQs

Why did the market decline despite strong premium demand?

Rising component costs and supply constraints offset gains from premium devices. Vendors reduced low-end production to protect margins, resulting in an overall shipment decline.

Which vendors benefited the most?

Huawei and Apple led market resilience. Huawei benefited from strong flagship and foldable demand, while Apple recorded the highest growth rate among leading vendors.

What risks could impact the market in 2026?

Persistent component cost inflation, supply chain disruptions, and reduced low-end demand could limit recovery. Vendor discipline on inventory and pricing will remain critical.

-Ends-

About IDC Trackers

IDC Tracker products provide accurate and timely market size, company share, and forecasts for hundreds of technology markets from more than 100 countries around the globe. Using proprietary tools and research processes, IDC’s Trackers are updated on a semi-annual, quarterly, and monthly basis. Tracker results are delivered to clients in user-friendly excel deliverables and on-line query tools. The IDC Tracker Charts app allows users to view data charts from the most recent IDC Tracker products on their iPhone and iPad.

About IDC

International Data Corporation (IDC) is the premier global provider of market intelligence, advisory services, and events for the information technology, telecommunications, and consumer technology markets. With more than 1,300 analysts worldwide, IDC offers global, regional, and local expertise on technology and industry opportunities and trends in over 110 countries. IDC’s analysis and insight helps IT professionals, business executives, and the investment community to make fact-based technology decisions and to achieve their key business objectives. Founded in 1964, IDC is a wholly-owned subsidiary of International Data Group (IDG), the world’s leading tech media, data and marketing services company. To learn more about IDC, please visit www.idc.com/ap. Follow IDC on Twitter at @IDCAP and LinkedIn. Subscribe to the IDC Blog for industry news and insights.

Will Wong - Senior Research Manager - IDC

Analyst Profile LinkedIn
Will Wong is a Senior Research Manager with IDC’s Asia/Pacific Client Devices Group. Based in Singapore, he covers the mobile phone market and is responsible for formulating valuable insights to help clients stay competitive and successful.

当生成式 AI 从“技术试验”走向“业务核心”，企业真正面临的挑战已不再是模型能力，而是如何高效、可控地规模化落地。训推一体化优化，正在成为企业构建竞争壁垒的关键抓手。

生成式 AI 正在加速向企业核心业务渗透，其价值重心也从“功能创新”转向“效率重构”。IDC研究显示，到2026年，近半数中国企业将部署超过10个生成式 AI 应用场景，这意味着AI已从探索阶段进入规模化应用阶段。在这一过程中，企业逐渐意识到，大模型的真正挑战并不在于“是否拥有”，而在于“是否能够稳定、高效、低成本地运行并持续优化”。

一、生成式 AI 进入规模化落地阶段

从行业演进来看，生成式 AI 正在经历从“点状应用”向“系统性重构”的关键跃迁。早期应用主要集中在内容生成、电商营销等互联网场景，而当前则加速渗透至金融、制造、医疗等核心行业，并逐步深入到采购、销售、财务以及IT运维等企业关键流程之中。这一变化意味着AI不再只是提升局部效率的工具，而正在成为重构企业运营模式的重要基础设施。

与此同时，用户侧的使用习惯也在发生根本性变化。生成式 AI 的使用频率不断提升，越来越多用户已经形成日常依赖，这从侧面推动了企业对AI系统稳定性和响应能力提出更高要求。在此背景下，智能体（Agent）逐渐成为新的应用形态，其“认知—决策—执行”的闭环能力，使AI能够直接参与业务流程，而非仅提供辅助支持。IDC认为，智能体将成为企业数字化转型的核心驱动力之一。

二、Scaling Law 之下的工程复杂性挑战

虽然大模型能力仍然遵循Scaling Law，即通过增加数据规模、模型参数和算力投入来持续提升效果，但在实际落地过程中，这一规律正面临越来越明显的工程约束。随着模型规模扩大，企业不仅需要应对训练成本的指数级增长，还必须解决推理延迟、系统吞吐以及服务稳定性等问题。

更重要的是，系统瓶颈正在发生转移。在早期阶段，算力是主要限制因素，而在当前阶段，跨节点通信、数据传输以及系统调度能力逐渐成为新的瓶颈。特别是在多节点、多GPU环境下，网络带宽和通信效率直接影响整体性能，使得“等数据”而非“等算力”成为常态。

因此，大模型优化已经从单一算法问题，演变为涵盖计算、存储、网络和软件栈的复杂系统工程问题。这一转变也意味着，企业竞争的焦点正在从模型能力本身，转向整体工程能力和架构设计水平。

三、训推一体化成为主流优化路径

在上述背景下，企业逐渐倾向于采用端到端的训推一体化框架，以降低系统复杂度并提升整体效率。这类框架能够贯穿模型生命周期，从数据处理、模型训练到推理部署，实现统一管理与持续优化，从而显著缩短模型迭代周期。

在训练阶段，优化重点已经从“增加算力投入”转向“提升算力利用效率”。通过多维分布式并行策略以及混合精度训练技术，企业可以在有限资源条件下显著提升训练效率，并降低硬件成本。同时，显存优化和通信优化技术的应用，使得大规模模型训练逐渐具备可扩展性和可持续性。

进入后训练阶段，模型优化的重点转向业务适配能力。通过参数高效微调、模型蒸馏以及强化学习等技术，企业能够在控制成本的同时提升模型在特定场景中的表现。尤其是在智能体应用中，强化学习成为提升复杂推理能力的关键手段，但其高计算成本和系统复杂度也对企业提出了更高要求。

在推理部署阶段，优化的核心目标则是实现性能与成本之间的动态平衡。随着模型规模和上下文长度不断增加，推理系统需要同时满足低延迟、高吞吐以及高并发需求。在这一过程中，KV Cache优化、动态批处理以及低精度量化等技术成为关键手段，而PD分离架构（Prefill与Decode分离）及其配套的缓存管理机制，已逐渐成为行业共识。

在当前主流的大模型开发框架中，通常会提供覆盖模型训练与推理全流程的多种优化模型与工具。下图展示了在一个典型的大模型应用流程中，这类框架所包含的核心优化工具及其对应的主要优化方案。

四、基础设施成为新一轮竞争焦点

随着大模型应用规模的扩大，底层基础设施的重要性显著提升。IDC数据显示，中国生成式 AI 基础设施市场正处于高速增长阶段，预计未来几年将保持超过60%的年复合增长率。这一趋势表明，企业对算力资源、存储能力以及网络架构的需求正在快速提升。

更深层次来看，AI竞争正在从“模型竞争”转向“基础设施与系统能力竞争”。高性能GPU、低精度计算能力、多级缓存体系以及高速互联网络，正在共同构成新一代AI基础设施的核心。这些能力不仅决定模型训练效率，也直接影响推理成本和服务质量，从而成为企业构建长期竞争优势的重要基础。

五、对技术决策者的关键建议

大模型正在快速迭代和扩展，在训练和推理阶段都面临算力成本和数据质量的挑战。随着模型规模的增长和新技术的涌现，如何在提高训练效果和推理效率的同时，确保模型的稳定性和可控性，是技术提供方需要解决的重要问题。

模型训练阶段：采用多维分布式并行（如数据并行、张量并行、流水线并行）和混合精度训练（BF16/FP16），可大幅提升训练效率，缩短开发周期，降低硬件资源消耗。利用高效的数据管道和动态负载均衡，确保算力利用最大化，减少资源闲置。

后训练（微调/蒸馏）阶段：应用参数高效微调技术（如LoRA、Prefix Tuning）和模型蒸馏，可在保持模型性能的同时显著降低部署成本，提升模型适应性。结合自动化超参搜索和增量训练，提升模型在特定业务场景下的表现。

推理环节：采用低精度量化（FP8/INT4）、内核融合、KV Cache优化和动态批处理等技术，能有效提升推理吞吐量和响应速度，降低显存和算力需求。部署高效的服务架构（如PD分离、异步调度），保障高并发场景下的稳定性和可扩展性。

结论

大模型技术正在快速走向成熟，但真正拉开企业差距的，未来将不仅仅是模型本身，更是围绕训练、后训练与推理的系统化优化能力，这也决定了基础设施是否能输出高质量、有效Token。实践表明，通过构建统一的优化框架并持续迭代技术栈，企业不仅能够加速AI应用落地，还能够显著降低创新成本，提升整体业务价值。

本文IDC相关报告：

  • IDC《大模型训练推理优化部署的最佳实践》

基于上述分析，IDC在大模型、生成式AI以及智能体等领域已形成系统化的研究体系。围绕中国AI与GenAI市场、智能体与自动化应用、以及Data+AI与Data Agent等方向，IDC持续发布涵盖市场规模与预测、技术趋势洞察、厂商竞争格局评估（如MarketScape）、产品与能力评测（Tech Assessment / ProductScape），以及最佳实践与行业案例等多类型研究成果。同时，IDC还可为企业提供定制化咨询服务，包括技术选型与架构规划、市场进入与竞争分析、产品策略与生态评估，以及行业应用落地路径设计等。

请点击此处与我们联系。

IDC在2025年下半年实施的一系列消费者研究显示，伴随着“国补”政策的逐渐退坡，日益增长的价格使得消费者对终端商品的选择更加理性。端侧AI与智能体能力的不断增强一方面推动AI渗透率与使用率的提升，另一方面也促使智能体验成为影响用户NPS的关键因素。

通过IDC对消费终端市场最为核心的笔记本、平板以及手机市场主要厂商NPS排名与市场动态的解析，可以看到智能体验等因素如何影响消费者对终端厂商的评价，以及在2026年智能终端消费市场，厂商应该聚焦哪些方面来达成用户NPS的提升。以下为本次研究的主要发现：

笔记本市场

笔记本市场消费者在2025年下半年仍处于价格敏感期，普遍升高的到手价格对大部分品牌都产生了一定的负面影响。产品体验对于用户对品牌的推荐意愿影响力提升，笔记本产品能否提供最佳的日常使用体验以及超出预期的智能体验成为影响用户口碑的关键

AI PC细分市场

由于用户对笔记本智能体验关注的的持续提升，IDC针对AI PC市场开展了独立的调研。根据研究，AI PC消费者重视设备是否搭载有功能丰富的AI助手。本地知识库与大模型，AI创作与研究能力等功能的可用性及使用体验均对用户NPS有直接影响。是否与系统深度融合，能否跨生态智能互联或将成为AI PC竞争的新高地。

平板市场

平板市场厂商通过不断推出新品拉动消费者热情，2025年下半年密集上市的新品也提升了用户对于产品体验的关注度。差异化的使用场景使得不同平板群体的关注因素存在差异，但性能与高性价比是平板消费者共同关注的重点。能否通过AI功能进一步提升使用体验将是影响用户选择的关键因素。

手机市场：

手机市场延续数年的产品“内卷” 出现缓和迹象，不断上升的成本促使厂商对新品迭代更加精打细算。但消费者对于手机产品体验的预期也在不断提升，任何因素导致的负向体验都会对品牌NPS产生较大影响。消费者对手机端AI与智能体体验的关注度快速提升，并且显著高于其他终端市场。

市场洞察与建议

洞察一：消费逐渐回归理性，务实主义有望回归

2026年，关键元器件成本的持续上涨将推动终端产品价格持续上行，用户的购买决策将趋向理性与谨慎。产品功能，使用体验，品牌口碑以及价格因素将成为用户未来购买消费终端产品的首要考虑因素

洞察二：智能需求升级，端侧AI将成为关键要素

AI智能体认知率与使用率不断提升，“AI PC”与“AI手机”等概念被更多消费者熟知。用户对于智能体验的需求也将快速迭代。端侧搭载AI能力将逐渐成为市场标配，而能否为用户提供跨场景，无缝的智能体验将成为厂商成功的关键

洞察三：体验决定价值，场景痛点更受关注

用户的价值感受中枢将回归到场景与体验，用户的高频痛点更加具象化，且与场景深度绑定。如果厂商提供的产品与服务能够帮助用户解决高频场景下的核心体验问题，将能为用户带来最强的价值感受。同时，更多的用户开始期待AI智能体在特定场景下的表现与体验，在垂类场景深度优化的智能体验有望快速形成正向口碑传播。

分析师观点：

IDC中国研究经理王楷表示，中国智能终端消费市场处于持续变化阶段，涨价与AI体验升级预计将成为影响2026年市场的核心要素。更高的购买成本将进一步提升用户对于“买的值”的期待，能否在保证产品的基础体验过硬的同时，通过AI与智能体为用户带来实质性的体验升级，将成为厂商能否获得更多用户推荐的关键。

如需了解更多IDC相关研究或进一步与我们沟通，欢迎识别二维码与 IDC 联系，我们将安排专人与您对接，为您提供定制化的市场洞察与咨询服务。


如需了解IDC在智能终端领域的最新研究报告、数据产品及行业分析，请扫描二维码，在线获取完整研究目录与内容简介，助您精准把握市场脉动。

请点击此处与我们联系。

AI promises to transform IT service management (ITSM): faster resolutions, automated service desks, and the ability to catch problems before they happen. Yet many AI initiatives fall short.

In my work with end-user clients, the difference between success and frustration almost always comes down to one thing: the configuration management database (CMDB).

A CMDB is simply a system that tracks all your IT assets (servers, laptops, software) and how they connect. When your CMDB is accurate, complete, and includes cost information, AI becomes a powerful tool. It can route problems to the right people, fulfill requests automatically, and support better contract negotiations with external suppliers.

When your CMDB is a mess, AI just makes that mess happen faster.

Two ways organizations manage their CMDB

Across hundreds of client engagements, we see a clear divide.

At one end are “black box” environments where the CMDB is a relic, populated once, rarely maintained, and lacking cost or relationship data.

At the other end are “cost-enriched” environments where the CMDB is a live, trusted source of information, continuously updated and directly linked to service management processes.

The wrong way: AI amplifies bad CMDB data

A global manufacturer learned this the hard way. It deployed an AI virtual agent to automate service desk requests. The tool was designed to check asset availability and resolve common issues without human intervention.

But the CMDB was years out of date. When employees requested laptop refreshes, the AI encountered duplicate and obsolete records. Unable to determine which devices were actually in use, it escalated nearly every request to a human agent.

There was no reduction in ticket volume. In fact, the service desk spent more time correcting AI outputs than resolving issues.

Predictive incident management also failed. The CMDB lacked accurate relationships between applications and infrastructure, so the AI could not prioritize incidents by business impact. Average resolution times remained unchanged.

The right way: a cost-enriched CMDB unlocks AI value

Now contrast that with a mid-size financial services organization that invested 18 months in CMDB hygiene and enrichment.

Every configuration item (laptops, virtual machines, storage, and more) included unit cost, supplier, lease end date, and relationship data. Automated discovery ran continuously, and the CMDB was integrated with finance, procurement, and virtualization platforms. It became the single source of truth for IT service management.

When this organization deployed an AI-powered service management platform, the value was immediate.

A department head requested 50 new high-spec laptops. Instead of automatically generating a purchase order, the AI queried the CMDB and identified 20 unassigned units already in inventory. It reserved those devices and routed a budget approval request for the remaining 30, including total cost (hardware plus licensing).

Fulfillment time dropped from five days to less than 24 hours. The service desk was freed from a high-volume request category, and real-time cost visibility helped prevent budget overruns.

The biggest impact came during renewal of the organization’s managed services contract. The managed service provider (MSP) priced support per virtual machine and per terabyte of storage.

With an integrated CMDB, AI analytics enabled a pre-renewal audit. It identified:

  • 50 virtual machines that had been powered off for more than 90 days, with no activity
  • 15 terabytes of allocated storage with no active data

The CMDB confirmed no business owner or application dependency, making these safe to remove.

Eliminating this unused infrastructure reduced MSP costs by €4,000 per month, or €48,000 annually. It also created a more transparent and accountable partnership with the provider.

What the data shows: CMDB maturity drives AI outcomes

IDC research shows that organizations with a mature, cost-enriched CMDB achieve significantly better AI outcomes.

They:

  • Deliver 2.5x higher return on AI investments
  • Reduce service request times by 30–50%
  • Improve average resolution times by 15–25%
  • Lower total service management costs by 10–15%

Other factors that influence AI success in ITSM

A strong CMDB is the foundation, but it is not the only requirement.

Successful AI initiatives also depend on:

  • Clear, well-documented processes
  • Teams that understand how to work with AI tools
  • A culture that supports change

Without these, even high-quality data will not deliver full value. But without accurate, cost-enriched data, even the best processes and teams will struggle to make AI effective.

What you can do now

Treat your CMDB as a strategic foundation for IT service management, not just a compliance requirement.

  • Use automated tools to maintain accuracy
  • Assign clear ownership for CMDB quality
  • Add cost and relationship data to every asset
  • Integrate the CMDB with finance, procurement, and IT systems
  • Clean your data before deploying AI

When negotiating with suppliers, use CMDB-driven insights to validate usage and challenge invoices.

Final thought: AI is only as good as your data

In the rush to adopt AI in ITSM, success will not come from buying the most advanced tools. It will come from investing in the data those tools depend on.

Organizations that win will build strong data foundations, supported by the right processes, skills, and culture.

An accurate, cost-enriched CMDB is no longer just an operational necessity. It is a competitive advantage for driving efficiency, improving service quality, and reducing costs with AI.

Tom Collins - Senior Consultant, IT Sourcing & Benchmarketing - IDC

LinkedIn
Tom Collins is a Senior Consultant in IDC’s IT Sourcing and Benchmarking practice, advising organizations on IT cost management, sourcing strategy, and technology procurement.

Wholesale has traditionally been a scale-driven business focused on connectivity and volume. That model is now evolving. 

Across the telecom industry, wholesale providers are rethinking how they deliver value, moving toward more flexible, platform-based approaches that go beyond traditional network services. 

This shift is being driven by changing customer expectations, new technologies, and increasing pressure to create sustainable growth. 

Wholesale telecommunications is shifting toward platform-based models 

IDC highlights 2026 as a key moment in the transition toward more automated, API-driven, and AI-enabled wholesale models. 

Rather than offering static products, wholesale providers are increasingly expected to deliver services that are more flexible, on-demand, and easier to integrate into customer environments. 

This includes: 

  • Greater use of APIs to expose network capabilities  
  • Increased automation across ordering, provisioning, and operations  
  • More dynamic and usage-based pricing models  

As a result, wholesale telecommunications is gradually adopting characteristics typically associated with cloud platforms. 

Customer expectations in wholesale telecommunications are changing 

Wholesale customers are no longer only looking for access to infrastructure. They expect solutions that can adapt to their specific requirements and business models. 

Flexibility, scalability, and ease of integration are becoming key decision factors. 

This is particularly relevant as enterprise and service provider customers operate in more complex, multi-vendor environments and require greater control over how services are consumed and managed. 

Wholesale providers are responding by offering more configurable services and by simplifying how customers interact with their networks. 

Ecosystems are becoming more important in wholesale telecom 

As wholesale models evolve, the role of ecosystems is expanding. 

Providers are increasingly working with partners to extend coverage, enhance capabilities, and co-develop new services. This includes collaboration across technology vendors, platform providers, and other telecom operators. 

At the same time, there is a growing focus on standardization, particularly around APIs and emerging technologies, to enable interoperability and scale across ecosystems. 

Managing these ecosystems effectively is becoming a key capability for wholesale providers. 

Vendor strategies in telecommunications are evolving 

This shift is happening alongside changes in how telcos approach their vendor landscape. 

Operators are becoming more selective and are reducing the number of partners they work with. There is a clear move toward strategic partnerships with vendors that can deliver end-to-end capabilities and take on greater accountability. 

This reflects the increasing complexity of telecom environments, where fragmented ownership across multiple vendors can slow down transformation and increase operational challenges. 

Fewer, more integrated partners can help simplify execution and align outcomes more closely with business objectives. 

What this means for wholesale telecom providers 

For wholesale telcos, the transition to platform-based models requires both technology and organizational change. 

This includes: 

  • Modernizing legacy systems to support API-driven services  
  • Investing in automation and AI capabilities  
  • Rethinking product design toward more modular, flexible offerings  
  • Building and managing partner ecosystems more actively  

At the same time, providers need to balance innovation with the realities of existing infrastructure and customer commitments. 

Wholesale telecommunications is becoming a strategic growth lever 

Wholesale is no longer just a supporting function within telecom organizations. It is increasingly seen as an area for differentiation and new revenue generation. 
Platform-based models, ecosystem collaboration, and more flexible service delivery approaches are opening up new opportunities to monetize infrastructure and reach new customer segments. 

As these models mature, the ability to execute effectively will determine which providers can translate this shift into sustainable growth. 

Download the full analysis 

Wholesale transformation is one of several trends reshaping the telecom market. In the IDC eBook State of the Telco Market 2026, you’ll find detailed data, forecasts, and analysis on platform-based models, API strategies, and evolving telco business models. 

Download the eBook to explore how wholesale is evolving and what it means for telecom providers. 

If you’re currently evaluating how platform models or ecosystem strategies could impact your wholesale business, our experts are happy to exchange perspectives. Whether you’re just starting or already transforming your model, we welcome the conversation. Get in touch with our team to continue the discussion. 

Jan Hein Bakkers - Senior Research Director, European Infrastructure and Telecoms - IDC

Analyst Profile
Jan Hein Bakkers is responsible for IDC's research efforts in the European enterprise and wholesale communications domain. His personal areas of expertise include internet access and WAN services, as well as wholesale connectivity markets. His research has a particular focus on the evolution of wholesale models, WAN transformation and the role of key growth segments, such as SD-WAN, cloud connectivity and very high bandwidth services within that. His work is published in IDC's EMEA Wholesale Telecoms Strategies and European Enterprise Communications Services programs, as well as the Worldwide Telecom Services Tracker. In addition, he provides his insights, opinions, and advice to a broad base of clients via custom engagements. He is a regular speaker at industry, client, and IDC events, and is frequently quoted in the press. Since joining IDC in 2001, he has analyzed a range of telecommunications and networking areas, including broadband equipment, TV services, and consumer multiplay strategies. He is based in the Netherlands and has degrees in international marketing and technical business administration.

For the past few years, large language models (LLMs) have dominated the AI conversation, and for good reason. They have transformed how we interact with software, accelerated content creation, and unlocked new forms of productivity.

But here is the reality enterprises need to internalize in 2026: the future of AI is no longer about a single model architecture.

A new AI model ecosystem is rapidly taking shape, one that is more diverse, more specialized, and far more powerful when orchestrated correctly.

The shift is subtle, but its implications are massive.

What is replacing LLM-only AI strategies? The shift to “any-to-any” models

LLMs, largely built on transformer architectures, still play a central role. But they are no longer sufficient on their own. Across industries, we are seeing a more diverse model landscape emerge that includes:

  • Deep reasoning models with adaptive thinking
  • Multimodal models (text, image, video, audio)
  • Small, efficient models for edge and latency- or cost-sensitive use cases
  • Domain-specific models tailored to industries like healthcare, finance, and manufacturing
  • Dedicated language models that excel in underrepresented languages
  • Quantitative and physics-based models for scientific and simulation-heavy workloads and real-world problems, ultimately enabling physical AI
  • Models with novel architectures beyond transformers, such as structured state space models (SSMs), mixture of experts (MoE), liquid neural networks (LLNs), and world models that solve problems differently, including time series and spatial tasks
  • Vision-language models built on novel architectures
  • Vision-action and world models that interact with real environments
  • Autonomous AI agents that interact with the real world, search the web, manage knowledge work, and use tools

This expansion is happening because real-world problems are not purely linguistic, and many business requirements are not easily solved by transformer-based models. Enterprises are discovering that while LLMs “think in language,” many business problems require reasoning in numbers, space, time, and physics.

To truly accomplish work and enable action-oriented outcomes, models need different skill sets and must be trained on a more diverse corpus of data. Solving a business problem may require a combination of models optimized for both reasoning and execution.

Why multi-model AI matters for enterprise strategy

This is not just a technical evolution. It is a strategic one. Organizations that continue to treat AI as a single-model problem will quickly hit limitations in:

  • Accuracy
  • Cost efficiency
  • Scalability
  • Use case coverage

Meanwhile, those adopting a multi-model, multimodal, and multi-agent approach will unlock entirely new capabilities, including action-oriented AI.

How to build a multi-model AI strategy: Three actions to take now

1. Treat model selection as a core capability
Most enterprises still underinvest in model selection. That is a mistake. Choosing the right model is no longer a one-time decision. It is an ongoing enterprise capability involving:

  • Matching model types to use cases
  • Evaluating trade-offs (cost versus performance versus latency)
  • Routing tasks dynamically across models

Action: Build internal processes or platforms for continuous model evaluation and benchmarking. Think of this like cloud cost optimization, but for AI.

2. Design for a “model portfolio,” not a single stack
Your future AI architecture will look less like a stack and more like a portfolio, with different models playing distinct roles. This includes general-purpose LLMs, models for large quantitative tasks, and vision-language models for generating video and other content. The possibilities are extensive.

This “constellation of models” is becoming the new normal.

Action: Start mapping your top 10 AI use cases and identify where different model types could outperform a single LLM approach.

3. Invest in AI-ready data infrastructure
As models diversify, data becomes the unifying layer. Without the right data foundation, even the best models will underperform. The shift toward a multi-model world is driving a major evolution in data platforms, including:

  • Converged data architectures
  • Real-time pipelines
  • Vector and multimodal databases

Action: Prioritize data readiness, including governance, pipelines, and accessibility across model types.

Final thought: AI success will depend on model orchestration

The organizations that win will be those that:

  • Embrace model diversity
  • Build flexible AI architectures that can quickly incorporate new model types and innovations
  • Invest in AI-ready data architecture and orchestration

Because in the new AI landscape, it is not about having the best model. It is about having the right combination of models.

Learn more

For a deeper dive into how the model ecosystem is evolving and what it means for enterprise strategy, explore IDC’s latest research on AI model landscapes and emerging architectures.

Tim Law - Research Director, AI & Automation - IDC

Analyst Profile LinkedIn
Timothy Law is a Research Director for AI & Automation, responsible for the generative AI lifecycle tools and technologies research practice. Mr. Law’s core research coverage includes the evolution of generative AI infrastructure and platforms, foundation models, developer tools, observability solutions, agentic systems, and generative AI services. This research analyzes the trends and developments in the AI software markets, including the costs, benefits, and impact of generative AI technologies.

Zhenshan Zhong - Vice President, Emerging Technology Research - IDC

Analyst Profile LinkedIn
Zhenshan Zhong is the vice president of Emerging Technology Research, leading the research teams focusing on emerging technologies (the four pillars and innovation accelerators). Zhenshan and his team are responsible for the overall success of these research domains, including participation and program management of key client engagements and the daily management of research team members.

The gap between execution and strategy often widens as hype cycles gain steam. This is evident in the growing disconnect in retail and restaurants between AI investment and foundational readiness. Executives are placing big bets on AI to transform operations and CX, but at a basic level, AI is only as good as the data underpinning it, and many organizations are still operating on data foundations that were never designed for real-time decision making or intelligent automation.

This gap is becoming one of the biggest risks to competitiveness. No matter how advanced the AI, it cannot overcome fragmented systems, unsynchronized data, and limited visibility across the business.

Before AI can deliver value, the data must work.

Retailers and restaurant operators are simultaneously trying to modernize their businesses while continuing to deliver consistency, efficiency, and growth in an environment that remains highly unpredictable.

Economic volatility, shifting tariffs, supply chain disruption, labor constraints, margin pressure, rising customer expectations, and intensifying competition are all converging at once. At the same time, business leaders are being told that AI can help address many of these challenges, from improving forecasting and pricing to strengthening personalization and automating decision making.

In many ways, this is like trying to build the plane while in the air, and there is an important reality that is becoming clearer: AI is only as effective as the data foundation beneath it.

That is one of the central themes explored in the IDC MarketScape: Worldwide Data Modernization Services Providers for Retail and Restaurants 2026 Vendor Assessment.

From modernization as IT project to business necessity

As we look across the market, it is evident that retailers and restaurant operators are no longer thinking about modernization as a narrow IT project. They are increasingly focused on modernizing data flows, processes, and systems as a business necessity, one that can enable greater operational efficiency now and more advanced AI capabilities over time.

For most organizations, the issue is not, nor has it ever been, a lack of data. Rather, restaurateurs and retailers often bemoan not knowing where data is, who owns it, and what to do with it. The common refrain is that retailers and restaurants already generate enormous volumes of data across POS systems, ecommerce platforms, loyalty programs, ERP environments, supply chain applications, labor systems, partner networks, and digital customer touchpoints.

The challenge is that much of this data remains fragmented, delayed, inconsistent, or difficult to govern. As a result, organizations often struggle to turn raw information into timely, trusted, and actionable insight.

Enter data modernization

A modern data environment is not just about migrating workloads to the cloud. It is about creating an architecture that can unify distributed data, improve visibility, support real-time decision making, and establish the governance needed to scale analytics and AI responsibly.

In retail and restaurants especially, where business conditions can change by the hour and differ significantly by location or region, synchronized and accessible data is foundational to better execution.

IDC research reinforces how significant this issue has become. According to IDC’s Global Retail Technologies and Business Processes Trends Survey, 2025, retail and restaurant organizations cite data visibility and accessibility, data synchronization, and data unification among the most important challenges to staying competitive. The same research also shows that poor data synchronization and integration, along with lack of access to real-time data and analytics, remain persistent barriers to successfully executing AI strategies.

Why AI efforts stall without modern data

Many organizations are eager to scale AI, but their data is still trapped in disconnected environments, governed inconsistently, or refreshed too slowly to support intelligent action. In that context, AI initiatives can easily stall, underperform, or produce outcomes that business users do not fully trust.

This is why so many retailers and restaurant operators are now prioritizing modernization efforts to improve how data moves across the enterprise and unlock efficiency and innovation. They want to reduce friction between systems and teams. They want better governance, better quality, and better observability. Increasingly, they want flexible architecture to support future use cases they may not have fully defined yet.

What this means for technology buyers and providers

Technology buyers eager to adopt the latest AI capabilities must think about data modernization and how they engage services partners to support this work. This goes beyond migrating legacy data estates. That is only the beginning.

Services providers must help brands design more adaptable, composable, and cloud-enabled foundations that support retail and restaurant-specific outcomes. This may include real-time inventory visibility, improved demand forecasting, dynamic pricing, customer intelligence, shrink reduction, faster fulfillment, or more efficient store and restaurant operations.

Retail and restaurant businesses do not need generic modernization strategies disconnected from business realities. They need partners that understand frontline complexity, omnichannel operations, distributed environments, and the increasing pressure to improve both efficiency and customer experience at the same time.

Data modernization as a strategic lever

For retailers and restaurant operators, data modernization is far more than a back-end initiative. It is becoming a strategic lever for resilience, innovation, and competitive differentiation. AI may be driving urgency, but modernization is what makes progress possible.

To better understand how services providers are helping retailers and restaurants modernize data foundations for AI, the IDC MarketScape: Data Modernization Services Providers for Retail and Restaurants, 2026 provides in-depth analysis and guidance.

Dorothy Creamer - Sr. Research Manager - IDC

Analyst Profile
Dorothy Creamer is Senior Research Manager for IDC Research, Hospitality & Travel Digital Transformation Strategies, providing research and advisory services for hotels, casinos, restaurants and travel organizations. Ms. Creamer's research will focus on how these business segments are transforming and leveraging technology to increase efficiencies, deliver operational benefits and identify new revenue streams. Ms. Creamer's research will report on effective digital strategies to empower both guests and employees and analysis of areas of opportunity in a fast-evolving and highly competitive segment.

Margot Juros - Research Director, Worldwide Retail AI, Platforms, and Technologies - IDC

Analyst Profile LinkedIn
Margot Juros is a Research Director for IDC Retail Insights responsible for the Worldwide Retail Platforms & Technologies research program. Ms. Juros’s core research focuses on examining best practices, discerning emerging trends/critical business concerns, evaluating market conditions and vendor offerings to provide authoritative advice on IT investment strategies and optimal use of technologies for modern retail IT infrastructure. Her research covers key technologies for retail transformation, including cloud/edge, AI/GenAI/agents, cybersecurity/security, data management, payments, unified platforms, mobile, and networking/5G/Wi-Fi.

2026年4月8日，Anthropic正式官宣推出前沿大模型Claude Mythos Preview，并同步启动网络安全合作计划Project Glasswing，这场技术发布不仅改写了全球网络攻防的力量格局，更给中国网络安全行业带来了深刻的冲击与思考。

作为当下最具颠覆性的AI模型，Mythos在网络安全领域展现出的能力已超越多数顶尖人类安全专家，而Project Glasswing计划的核心参与方中，却未见任何一家中国网络安全厂商的身影。在全球AI安全竞争日趋激烈、技术壁垒逐渐形成的背景下，Mythos带来的不仅是行业机遇，更有不容忽视的挑战，中国传统网络安全产品的生存空间被重新拷问，国内厂商如何破局突围，成为当下最紧迫的命题。

Anthropic近期惊艳表现：Mythos与Project Glasswing重构攻防规则

作为Anthropic迄今为止最强大的模型，Mythos并非专门为网络安全场景训练的专用模型，但其通用智能的自然溢出，使其在漏洞挖掘、代码推理、漏洞复现与利用等核心安全场景中实现了跨代级提升。根据Anthropic官方披露，Mythos已在所有主流操作系统和浏览器中发现数千个零日漏洞，其中多个被定级为高危，其能力甚至超越了除最顶尖安全专家之外的所有人类，且全程无需人工引导即可自主完成相关操作。

为应对Mythos能力带来的潜在安全风险，同时推动防御方抢占先机，Anthropic同步启动了Project Glasswing计划，该计划以“防御先行”为核心定位，旨在让防御方在AI攻击能力向更广泛行为者扩散前，获得足够的防御支撑。据悉，Project Glasswing的创始合作伙伴包括AWS、苹果、谷歌、微软、思科、Palo Alto Networks等12家科技与安全巨头，此外还有超过40家构建或维护关键软件基础设施的组织获得扩展访问权限，用于扫描和加固自身及所依赖的开源系统。为支撑该计划落地，Anthropic投入了最高1亿美元的API使用额度，同时向开源安全组织捐赠400万美元，推动漏洞披露、供应链安全等领域的行业协作，并承诺在90天内披露阶段性研究成果。但中国网络安全厂商均未进入这一合作体系，无法直接共享Mythos的模型能力与相关安全资源。

Mythos对全球及中国网络安全的重大冲击

共性挑战：AI重构攻防逻辑，传统安全体系面临失效

无论对于全球还是中国，Mythos带来的核心挑战，本质上是“AI驱动的攻防速度差”与“传统安全产品的能力断层”，这种冲击正在重构整个行业的游戏规则。

首先，攻防速度差急剧压缩，被动防御模式彻底失灵。传统网络安全依赖人工分析与静态规则匹配，漏洞挖掘、攻击响应的周期以小时或天为单位，而Mythos驱动的攻击可实现分钟级全链路渗透——从自主挖掘零日漏洞，到生成攻击代码、构造攻击链路，全程无需人工干预，这让传统“告警-分析-处置”的流程完全无法适配，防御方陷入“防不住、响应慢”的困境。

其次，攻击门槛大幅降低，黑产工具呈现“平民化”趋势。Mythos的能力若被恶意利用，将彻底打破网络攻击的技术壁垒，非专业攻击者也能借助模型快速生成恶意代码、挖掘零日漏洞、构造高仿真钓鱼内容，甚至自主完成漏洞利用。

最后，传统网络安全产品的核心能力全面失效。传统漏扫产品、防火墙、IPS/IDS等产品，核心依赖静态特征库与人工规则，无法识别AI生成的无特征零日攻击与隐蔽攻击。例如，传统漏扫工具无法自主挖掘代码深层逻辑漏洞，而Mythos可快速定位隐藏数十年的安全隐患；

中国特有的额外挑战：技术壁垒与不对称竞争

相较于参与Project Glasswing计划的海外厂商，中国网络安全领域面临着更为特殊且严峻的挑战：

核心挑战之一是技术获取不对称，与海外同行形成明显技术代差。Project Glasswing计划的参与方可优先借助Mythos的强大能力，开展漏洞挖掘、威胁检测与防御体系优化，同步共享相关安全研究成果与开源资源，实现防御能力的快速迭代。而中国厂商被完全排除在该合作体系之外，无法直接获取Mythos的模型能力与相关安全资源，只能依靠自身力量研发相关技术，这导致在AI安全技术的迭代速度上，中国与海外同行存在天然差距，尤其是在零日漏洞挖掘、AI对抗等高端领域，这种代差可能进一步扩大，短期内难以实现追赶。

核心挑战之二是网络安全威胁大幅升级，关键基础设施防御压力剧增。中国金融、能源、政务、医疗等关键基础设施领域，广泛使用各类开源软件与通用操作系统，而Mythos已在这些系统中发现大量高危漏洞。参与Project Glasswing计划的海外厂商可借助模型快速获取漏洞信息、生成修复方案，及时完成系统加固；而中国厂商无法获得相关漏洞信息与修复指引，只能依靠自主排查、自主修复，不仅大幅增加了防御成本，更延长了漏洞暴露时间，显著提升了关键基础设施被攻击的风险。同时，随着Mythos能力的扩散，中国面临的国家级APT攻击、黑产攻击将更加隐蔽、高效，攻击手段也将更加多样，进一步加剧了网络安全防御的难度，对国家网络安全底线构成严重挑战。

此外，中国网络安全行业还面临着AI安全人才短缺、自主技术研发投入压力大等衍生问题。海外依托Project Glasswing计划形成了“技术共享+人才协同”的良性生态，而中国在AI安全领域的高端人才储备不足，自主研发缺乏成熟的技术参考与生态支撑，进一步制约了防御能力的提升，使得中国在应对Mythos带来的AI攻击时，处于更加被动的地位。

中国网络安全行业未来方向与市场机会

结合前文Mythos带来的攻防变革、中国面临的技术代差与安全威胁，IDC立足中国网络安全行业现状，为中国网络安全行业未来发展方向和市场机会提出以下几点建议：

  • 一是正视挑战，摒弃侥幸心理。国内厂商需清醒认识Mythos带来的技术冲击，正视无法接入海外先进模型的技术代差、关键基础设施防御压力加剧、人才短缺等现实难题，摒弃“固守传统、被动防御”的惯性思维，主动将挑战转化为转型动力，明确AI时代的发展短板，精准发力破局。
  • 二是勇于变革，突破发展瓶颈。面对行业重构机遇，厂商需主动打破传统产品与技术壁垒，以变革思维推动自身升级，不局限于现有产品体系，主动适配AI主导的攻防新格局，将“AI赋能”作为转型核心，摒弃低效、落后的防御模式，实现从“被动应对”到“主动突围”的转变。
  • 三是加快AI能力迭代，筑牢核心竞争力。AI是未来网络安全产品的核心支撑，也是弥补技术代差的关键。厂商需加大AI安全技术研发投入，优先推进AI在漏洞识别、攻击研判、威胁防御等核心场景的落地，快速迭代AI算法与产品模块，适配本土合规与场景需求，打造AI原生安全产品，杜绝技术空转，切实提升防御实效，缩小与海外同行的差距。
  • 四是重视生态协作，凝聚众家合力。单打独斗难以应对行业变革，厂商需主动联动AI企业、科研机构、行业协会，构建自主可控的AI安全生态。通过技术合作借力标准化AI资源，降低研发成本；联合攻关核心技术，补齐人才与技术短板；参与行业标准与开源生态建设，共享资源、联动防御，挖掘关键基础设施、中小企业等细分场景市场机会，弥补无法接入海外生态的资源缺口。

目前，国内网络安全技术提供商持续加大对AI技术的资源投入，并已经在网络安全运营、数据发现与分类分级、威胁检测与响应、钓鱼邮件检测等多个场景取得显著效果。IDC的预测，中国网络安全相关AI Agent应用收入的市场规模在未来五年的复合年均增长率将高达106.5%，并在2030年达到593.5亿元人民币（如下图所示）。

同时，大模型安全评估、大模型安全围栏、智能体威胁检测等专注于AI安全防护领域的安全产品也正在快速涌入市场。根据IDC的预测，中国人工智能安全收入的市场规模在未来五年的复合年均增长率将达到50.5%，并在2030年达到340.3亿元人民币（如下图所示）。

IDC在2026年将聚焦AI自身安全与AI赋能安全启动多项市场研究，洞察网络安全技术发展趋势、展现安全厂商最新技术能力、推动中国网络安全市场持续变革与快速发展，欢迎广大技术提供商积极关注。

IDC更多相关研究

IDC已于2026年启动AI安全技术系列研究，围绕AI原生安全架构、安全智能体成熟度评估、AI驱动DevSecOps实践路径及企业级AI治理框架展开深入分析。

AI攻防规则已变。IDC深耕大模型安全与智能体治理研究，助您量化风险、重构防线。欢迎联系IDC中国，获取最新洞察与定制化咨询，共探破局之道。请点击此处与我们联系。