全球人工智能支出预计将在2027年增长47.5%至4.05万亿美元，其中基础设施占总额近五分之三，研究和咨询公司Gartner周四表示。
Global artificial intelligence spending is projected to rise 47.5% to $4.05 trillion in 2027, with infrastructure accounting for nearly three-fifths of the total, research and consulting firm Gartner said Thursday.
总AI支出预计将从2026年的约2.75万亿美元攀升，因为组织扩大采用，技术提供商构建额外的计算能力。
Total AI spending is forecast to climb from about $2.75 trillion in 2026 as organizations expand adoption and technology providers build additional computing capacity.
Gartner研究主管John-David Lovelock表示，用户需求增长快于预期，正在增加对运行AI应用和训练日益计算密集型模型所需基础设施的需求。
John-David Lovelock, Gartner's chief of research, said faster-than-anticipated growth in user requirements is increasing demand for infrastructure to run AI applications and train increasingly computing-intensive models.
根据该公司在巴塞罗那发布的声明，AI基础设施支出预计明年将增长53%至约2.4万亿美元，占全球AI支出的59%。
Spending on AI infrastructure is expected to increase 53% to approximately $2.4 trillion next year, representing 59% of worldwide AI expenditure, according to the company's release issued in Barcelona.
AI优化服务器支出作为最大的基础设施细分市场，预计将从今年的5276亿美元激增至2027年的9572亿美元。
Spending on AI-optimized servers, the largest infrastructure segment, is projected to surge from $527.6 billion this year to $957.2 billion in 2027.
Gartner将这一增长归因于大型云服务提供商和专业AI云公司持续扩张基础设施，以及由内存成本推动的服务器价格上涨。
Gartner attributed the increase to continued infrastructure expansion by large cloud providers and specialized AI cloud companies, alongside rising server prices driven by memory costs.
AI服务支出预计将增长29.5%至7460亿美元，而AI软件支出预计将增长42%至6490亿美元。
AI services spending is expected to grow 29.5% to $746 billion, while spending on AI software is forecast to increase 42% to $649 billion.
Lovelock表示，服务提供商正通过战略、数据准备、集成和部署帮助组织扩大AI计划。随着供应商将AI功能融入现有产品并推出管理和治理其使用的平台，企业也在增加软件预算。
Lovelock said service providers are helping organizations expand AI initiatives through strategy, data preparation, integration and deployment. Businesses are also increasing software budgets as vendors incorporate AI capabilities into existing products and introduce platforms to manage and govern their use.
AI网络安全支出预计明年将达到860亿美元，高于2026年的510亿美元，而AI代理和助手支出预计将从300亿美元增至670亿美元，翻倍以上。
Spending on AI cybersecurity is projected to reach $86 billion next year, up from $51 billion in 2026, while spending on AI agents and assistants is expected to more than double from $30 billion to $67 billion.
与此同时，生成式 AI 模型支出预计将增长 82% 至 510 亿美元，这得益于对自主 AI 智能体、融合不同类型内容的应用以及行业特定能力的需求。
Meanwhile, spending on generative AI models is forecast to grow 82% to $51 billion, supported by demand for autonomous AI agents, applications combining different types of content and industry-specific capabilities.
尽管专用和领域特定模型的支出预计将增长 99%，但基础模型预计将在 2027 年继续占据生成式 AI 模型支出的最大份额，达到 81%。
Although spending on specialized and domain-specific models is projected to increase 99%, foundation models are expected to retain the largest share of generative AI model spending at 81% in 2027.