Agent-based AI detonates enterprise demand for local and edge infrastructure without worrying about overbuilding (Goldman Sachs)
Agentic AI is driving enterprises to invest heavily in on-premise and edge infrastructure due to token cost, latency and governance advantages.
Agentic AI is driving enterprises to invest heavily in on-premise and edge infrastructure due to token cost, latency and governance advantages. This creates significant tailwinds for enterprise computing vendors. Experts point out that the huge demand for AI eliminates the risk of over-construction of infrastructure. Enterprises save operating expenses by replacing old servers with new high-performance platforms and laying off employees to raise hardware investment budgets. This trend is good for data center hardware, edge computing equipment and enterprise-level IT service providers, and Goldman Sachs’s view suggests that the market’s previous concerns about an AI capital expenditure bubble may be exaggerated. The current market enthusiasm for investment in AI has been partially priced in, but the shift to enterprise localization deployment may not yet be fully priced in. One sentence conclusion: The wave of agent AI is changing the investment paradigm of AI infrastructure. The enterprise local and edge computing market will usher in deterministic growth, which provides a broad space for related hardware and solution providers. Positive/negative: Positive for enterprise computing hardware manufacturers (such as Dell, HPE), edge computing and network equipment manufacturers. Some of the market's enthusiasm for AI investment has been reflected, but the structural increment brought about by the localization deployment trend may not yet be priced in. Catalysts: 1) changes in capital expenditure guidance of major CSPs (cloud service providers); 2) macro data of enterprise-level IT expenditures; 3) implementation cases and penetration data of agent-based AI applications in enterprises.