内容来源：https://aibusiness.com/generative-ai/ai-demand-outpacing-the-scaffolding-support-it

内容总结：

全球AI应用加速扩张，基础设施与治理短板凸显

随着生成式人工智能（AI）需求持续飙升，从数据中心容量到企业治理，支撑AI运转的系统和基础设施正面临空前压力。谷歌云赞助的研究显示，AI需求增速已远超现有体系、基础设施及组织的承载能力。

市场数据印证爆发式增长：Synergy Research Group报告指出，全球云业务收入同比激增35%，预计2026年第一季度将达1290亿美元，较去年同期增加350亿美元。其中，亚马逊云科技（AWS）、谷歌云和微软合计占据全球市场63%份额。英特尔CEO Lip-Bu Tan在4月23日的财报电话会上表示，AI需求“极其旺盛”，驱动首季营收达136亿美元，且需求持续超过供给，其至强（Xeon）服务器CPU仍是AI计算的核心支撑。Meta本周亦宣布将增加支出，以满足未来数据中心扩张及供应链合作的基建需求。此外，AI应用正加速渗透至客户体验（CX）、零售及供应链领域。

基础设施瓶颈日益突出：微软AI业务的爆发已暴露数据中心容量缺口，电力、冷却及土地资源约束催生了对浮动式、太空数据中心及新型能源的探索。企业层面，AI采用速度远超治理体系建设，导致成本失控、安全风险加剧及工具泛滥。缺乏有效监管下，企业既面临AI投入浪费，又难以掌握实际使用情况。

行业布局与劳动力影响凸显：本周行业动态显示，AI正从技术可行性转向系统支撑力的比拼，并将深刻影响企业决策、运营及劳动力结构：

• AI原生营销平台：Hightouch获1.5亿美元融资（估值27.5亿美元），推动企业营销流程智能化；
• 裁员预警：Bed Bath & Beyond CEO警告，AI将导致核心业务部门“显著减员”；
• 人机协作培训：Stellantis与微软扩大AI培训合作，分析师称此举凸显技能转型的战略价值；
• 供应链智能化：亚马逊AWS推出智能供应链工具，整合多系统实现数据与决策中枢化；
• 机器人实测场景：日本机场启动人形机器人试点，标志现实场景应用迈出关键一步；
• 消费解码新工具：企业CFO正借助AI与“合成受众”实时模拟消费者行为，重塑零售分析模式。

核心结论：AI发展的下一阶段挑战与机遇，将不再局限于技术本身，而在于能否突破基础设施瓶颈、完善治理体系，并助力组织实现系统性能力跃迁。

中文翻译：

由谷歌云赞助
选择您的首批生成式AI应用场景

要从生成式AI入手，首先应聚焦于那些能够改善人类与信息交互体验的领域。

从数据中心容量到企业治理，围绕AI的各类体系正疲于应对需求的增长。

AI需求的增速已超过支撑其发展的系统、基础设施及组织架构。

需求正在全面激增。

云业务收入同比增长35%。根据Synergy Research Group的数据，2026年第一季度全球云业务收入预计将较去年同期增加350亿美元，达到1290亿美元。

这包括AWS、谷歌云和微软，三者合计占据全球市场份额的63%。

英特尔指出，AI需求“极其旺盛”，推动其第一季度营收达136亿美元，且各业务线的需求持续超过供给。CEO陈立武在4月23日的财报电话会议上表示，CPU架构仍然是AI计算在生产中的支柱，尤其是该公司的至强服务器CPU。

Meta也在确保获得所需资源，并于本周表示将增加支出，以满足未来数据中心扩建和供应链合作所需的技术基础设施要求。

AI的应用也在客户体验、零售和供应链领域不断扩大。

但基础设施却未能跟上步伐。

微软的AI激增暴露了数据中心容量的缺口，同时电力、冷却和土地资源方面的限制日益显现，随着工作负载变得愈发复杂，实际和预测需求持续超过现有容量，这推动了对浮动式、太空式数据中心以及替代能源等新模式兴趣的增长。

而在企业内部，随着AI应用持续加速超越治理能力，局面正变得混乱，带来了新的成本、安全风险以及工具泛滥问题。缺乏更严格的监管，企业可能面临AI支出过度，同时对使用情况失去可见性。

综合来看，这一模式难以忽视。AI的局限已不再主要取决于技术本身的能力，而在于围绕技术的结构——从运行所需的设施，到管理所需的系统，再到试图跟上节奏的组织。

这正是下一波挑战和机遇的源头。

除了上述制约因素，本周的报道还突显出AI如何开始重塑企业层面的决策、运营和劳动力结构。

企业级智能营销平台估值达27.5亿美元：Hightouch已融资1.5亿美元，这凸显出原生AI厂商重塑企业营销工作流的势头。

Bed Bath & Beyond CEO：AI将导致“显著裁员”：这家家居用品零售商的警告表明，AI如何影响核心业务职能岗位仍存在不确定性。

分析师盛赞Stellantis与微软的AI培训合作：Stellantis与微软扩大合作，表明随着AI应用规模化，劳动力培训的重要性日益凸显。

AWS推出智能体式AI供应链工具：该公司正将智能体AI引入运营，通过整合多种工具，实现供应链数据的集中化和决策的协同。

人形机器人将在日本机场开展试点：人形机器人早期部署于机场运营，既展示了实际应用机器人技术的进展，也暴露了面临的挑战。

CFO借助AI和“合成受众”解读消费者行为：企业正越来越多地利用AI实时模拟和审视客户行为，这预示着客户体验与零售分析领域的新思路。

英文来源：

Sponsored by Google Cloud
Choosing Your First Generative AI Use Cases
To get started with generative AI, first focus on areas that can improve human experiences with information.
From data center capacity to enterprise governance, the systems surrounding AI are struggling to keep pace with demand.
AI demand is accelerating faster than the systems, infrastructure and organizations needed to sustain it.
Demand is surging everywhere.
Cloud revenue is up 35% year over year. Worldwide cloud revenue for Q1 2026 is expected to be up $35 billion from this time last year, reaching $129 billion, according to Synergy Research Group.
That includes AWS, Google Cloud and Microsoft, which together hold 63% of the world market.
Intel cited “tremendous” AI demand for driving $13.6 billion in first-quarter revenue, with demand continuing to exceed supply across its business. CEO Lip-Bu Tan noted on an April 23 earnings call that CPU-based architectures remain the backbone of AI computing in production, particularly for the company’s Xeon server CPUs.
Meta, too, is making sure it’s getting what it needs, saying this week that it’s increasing its spending to meet its technology infrastructure requirements for future data center expansion and supply chain deals.
AI adoption is also expanding across CX, retail and the supply chain.
But infrastructure is falling behind.
Microsoft's AI surge is exposing data center capacity gaps, while constraints on power, cooling and land are emerging, driving interest in new models like floating and space-based data centers and alternative energy sources as workloads grow more complex and actual and forecast demand continues to outpace available capacity.
And inside the enterprise, it’s getting messy as AI adoption continues to accelerate faster than governance, creating new layers of cost, security risk and tool sprawl. Without stronger oversight, companies risk overspending on AI while losing visibility into how it’s being used.
Taken together, the pattern is hard to ignore. AI is no longer limited as much by what the technology can do, but by the structures around it, from the infrastructure required to run it to the systems needed to manage it and the organizations trying to keep up.
That’s where the next set of challenges and opportunities will come from.
Beyond those constraints, this week’s coverage highlights how AI is starting to reshape decision-making, operations and the workforce across the enterprise.
Agentic Marketing Platform for Enterprises Valued at $2.75B: Hightouch has raised $150 million, highlighting growing momentum for AI-native vendors reshaping enterprise marketing workflows.
Bed Bath & Beyond CEO: AI Will Lead to ‘Significant Reduction in Headcount’: The home goods retailer’s warning signals ongoing uncertainty around how AI will impact jobs across core business functions.
Analysts Hail AI Training Partnership Between Stellantis and Microsoft: Stellantis’ expanded partnership with Microsoft points to the growing importance of workforce training as AI adoption scales.
AWS Unveils Agentic AI Supply Chain Tool: The company is pushing agentic AI into operations, combining multiple tools to centralize data and decision making across supply chains.
Humanoid Bots to Start Airport Pilot in Japan: Early deployments of humanoid robots in airport operations highlight both progress and challenges of real-world robotics adoption.
CFOs Lean on AI, ‘Synthetic Audiences’ to Decode Consumer Behavior: Organizations are increasingly using AI to simulate and examine customer behavior in real time, signaling new approaches to CX and retail analytics.